Dissociating Semantic and Phonological Verbal Fluency
Inaugural - Dissertation
zur Erlangung der Doktorwürde
der Wirtschafts - und Verhaltenswissenschaftlichen Fakultät
der Albert - Ludwigs - Universität Freiburg i. Br.
vorgelegt von
Charlotte Sophie Schmidt
geboren in Dortmund
Sommersemester 2018
Dekan: Prof. Dr. Alexander Renkl , Albert - Ludwigs - Universität Freiburg
Erstgutachter: Prof. Dr. Markus Heinrichs, Albert - Ludwigs - Universität Freiburg
Zweitgutachter: PD Dr. Joseph Krummenacher, Albert - Ludwigs - Universität Freiburg
Betreuer: Dr. Christoph P. Kaller, Dipl. - Psych., Universitätsklinikum Freiburg
Datum der Disputat ion: 16. November 2018
Acknowledgments
Several people have supported me during the completion of this doctoral thesis and I would like to thank everyone who contributed .
First of all , my special thanks go to Dr. Christoph Kaller for his close supervision . I am grateful for the amount of time he spent discussing new ideas and open questions with me and for his patience with my w ork.
Furthermore, I would like to thank Prof. Heinrichs and Dr. Krummenacher for being my doctorate supervisor s .
Many thanks to my colleagues Lena Schumacher, Konrad Schumac her, Kai Nitschke , Sandra
Loosli , and Lora Minkova and all people from the Freiburg Brain Imaging Center who helped me with valuable comments and who cre ated a good working atmosphere.
In addition, I would like to thank my friends from the University Sports - Club Freiburg who helped me refuel my energy after a long day at work .
A big thank you to all the participants, patients , and their families who volunteered to take part in my studies. Without their support it would not have been possible to presen t my results as I am doing now.
Furthermore, I would like to thank my family and friends who al ways gave me courage in the comple tion of my thesis .
Last but not least my special thanks go to my boyfriend Florian, who supported me the most during the last years and always had a bite of chocolate in strenuous times .
Thank you all !
iii
Abstract
Semantic and phonological verbal fluency tasks, that require the examinee to produce words to a given semantic (i.e. category) or phonological (i.e. letter) cue, are often used in clinical and experimental neuropsychology to assess language abilities and executive functioning. As in the two task variants word generation is believed to involve different types of search processes, it is also implicitly assumed that semantic and phonological verbal fluency differ both in their u nderlying cognitive processes and neural correlates involved for successful retrieval. However, in previously published factor analyses, semantic and phonological verbal fluency loaded on only one common factor. Moreover, the often assumed double dissociat ion between lesions in temporal vs. frontal brain regions and impairments in semantic vs. phonological verbal fluency was neither explicitly tested nor shown.
By using exploratory factor analyses in healthy participants and further confirmatory factor anal yses in chronic stroke patients with measures of semantic and phonological verbal fluency only, the first study of this thesis showed that inter - individual differences in semantic and phonological verbal fluency are represented by two factors. These two fa ctors show both common and distinct shares of variance and can therefore be interpreted as evidence for the involvement of both distinct and common cognitive processes.
The second study of this thesis used voxel - based lesion - behavior mapping analyses in ch ronic stroke patients and results revealed that a specific impairment in semantic vs. phonological verbal fluency resulted from lesions in temporal vs. frontal brain regions, respectively. Further analyses showed that a specific impairment depends on the e xact localization, especially for lesions in frontal brain areas and that semantic and phonological verbal fluency may be impaired to a similar extent.
In order to confirm the clinical utility of the German version of the verbal fluency task, applied in th e previous studies, the third study was concerned with its psychometric properties
v
by taking particularly the impact of item difficulty into account. Results indicate good test - score reliability, criterion - related concurrent validity, and test - retest relia bility, and reliability clearly benefits from aggregation across multiple items.
Overall results demonstrate that semantic and phonological fluency are based on clearly distinct but also common sets of cognitive processes and that they are differentially m ediated by temporal and frontal brain areas, respectively. Furthermore, for a reliable assessment of verbal fluency, item difficulty should be considered. Taken together, this thesis contributes to a more comprehensive understanding of verbal fluency and i ts underlying cognitive and neural correlates.
vi
Zusammenfassung
Aufgaben zur semantischen und phonologischen Wortflüssigkeit, bei denen Wörter zu vorgegebenen semantischen (i.e. Kategorien) und phonologischen (i.e. Buchstaben) Aufgaben gesagt werden müssen, werden in der klinischen und experimentellen Neuropsychologie häufig zur Erfassung sprachlicher und exekutiver Leistungen eingesetzt. Da man davon ausgeht, dass die Wortgenerierung bei den bei den Aufgabenvarianten von unterschiedlichen
Suchprozessen abhängt, wird auch implizit angenommen, dass sich die semantische und phonologische Wortflüssigkeit sowohl in ihren zugrundeliegenden kognitiven Prozessen als auch beteiligten neuralen Korrelaten un terscheiden. Bisher veröffentlichte Faktoranalysen zeigen jedoch, dass Aufgaben zur semantischen und phonologischen Wortflüssigkeit auf nur einem gemeinsamen Faktor laden. Die oft angenommene doppelte Dissoziation zwi schen
Läsionen in temporalen vs. fronta len Hirnarealen und semantischer vs. phonologischer
Wortflüssigkeit wurde bislang weder explicit getestet noch aufgezeigt.
Unter Verwendung explorativer Faktoranalysen in gesunden Probanden und weiterführenden konfirmatorischen Faktoranalysen in chronische n Schlaganfallpatienten, zeigt die erste Studie dieser Arbeit, dass inter - individuelle Unterschiede in semantischer und phonologischer
Wortflüssigkeit durch zwei Faktoren abgebildet werden. Die beiden Faktoren zeigen zudem getrennte als auch gemeinsame Var ianzanteile, was als Hinweis auf die Beteiligung unterschiedlicher und auch gemeinsamer kognitiver Prozesse interpretiert werden kann.
Die zweite Studie dieser Arbeit verwendet voxel - basierte Läsions - Verhalten Analysen mit
Daten chronischer Schlaganfallpa tienten und die Ergebnisse zeigen, dass eine spezifische
Beeinträchtigung in semantischer vs. phonologischer Wortflüssigkeit mit Läsionen in links temporalen beziehungsweise links frontalen Hirnarealen einhergeht. Weitere Analysen zeigen, dass die Spezifit ät der Beeinträchtigung, insbesondere bei Läsionen in frontalen
vii
Hirnarealen, lokalisationsabhängig ist und die semantische und phonologische Wortflüssigkeit auch gleichermaßen beeinträchtigt sein kann.
Um den klinischen Nutzen dieser deutschen Version zur Wortflüssigkeit, die in den vorherigen Studien verwendet wurde, zu bestärken, beschäftigt sich die dritte Studie mit seinen psychometrischen Eigenschaften, wobei der Einfluss der Aufgabenschwierigkeit besonders berücksichtigt wurde. Die Ergebnisse zeigen e ine gute Testergebnis - Reliabilität,
Konstruktvalidität und Test - Retest Reliabilität. Zudem profitiert die Reliabilität eindeutig von der Aggregation über mehrere Aufgaben hinweg.
Insgesamt zeigen die Ergebnisse, dass semantische und phonologische Fluenz auf klar unterschiedlichen aber auch gemeinsamen kognitiven Prozessen basieren und dass sie durch unterschiedliche Regionen im Temporal - beziehungsweise Frontallappen vermittelt werden.
Für eine zuverlässige Beurteilung der Reliabilität sollte zudem die Au fgabenschwierigkeit mit berücksichtigt werden. Zusammenfassend trägt diese Arbeit zu einem umfangreicheren
Verständnis der Wortflüssigkeit und ihren zugrunde liegenden kognitiven und neuralen
Korrelaten bei.
viii
Content
Acknowledgments ...... iii Abstract ...... v Zusammenfassung ...... vii Content ...... ix List of Figures ...... xi List of Tables ...... xiii List of Abbrevia tions ...... xiv Chapter 1 ...... 17 Introduction ...... 17 Chapter 2 ...... 21 Theoretical Background ...... 21 2.1 Language Abilities and Impairments ...... 21 2.2 Speech Production ...... 24 2.2.1 Neural Correlates of Word Production ...... 28 2.3 Verbal Fluency ...... 32 2.3.1 Semantic and Phonological Verbal Fluency ...... 32 2.3.2 Cognitive Processes of Verbal Fluency ...... 34 2.3.3 Neural Correlates of Verbal Fluency ...... 37 2.4 Stroke ...... 54 2.5 German Version of the Verbal Fluency Task ...... 55 2.6 Psychometric Properties of Verbal Fluency Tasks ...... 58 2.7 Present Studies and Aims of the Thesis ...... 60 Chapter 3 ...... 63 First Study – Factor analysis of the verbal fluency task ...... 63 3.1 Abstract ...... 64 3.2 Introduction ...... 65 3.3 Methods ...... 68 3.4 Results ...... 73 3.5 Discussion ...... 78 3.6 Conclusion ...... 83 3.7 Acknowledgement ...... 84 3.8 References ...... 85
ix
Chapter 4 ...... 89 Second Study – Dissoci ating neural correlates of verbal fluency ...... 89 4.1 Abstract ...... 90 4.2 Introduction ...... 92 4.3 Materials an d Methods ...... 98 4.4 Results ...... 107 4.5 Discussion ...... 117 4.6 Conclusion ...... 127 4.7 Acknowledgements ...... 128 4.8 Refere nces ...... 129 Chapter 5 ...... 135 Third Study – Psychometric properties of the verbal fluency task ...... 135 5.1 Abstract ...... 136 5.2 Introduction ...... 137 5.3 Methods ...... 147 5.4 Results ...... 153 5.5 Discussion ...... 164 5.6 Acknowledgement ...... 171 5.7 References ...... 172 Chapter 6 ...... 177 General Discussion ...... 177 6.1 Summary of key findings ...... 177 6.2 Item Difficulty ...... 179 6.3 Clinical application ...... 181 6.4 Limitations and future directions ...... 182 6.5 Conclusion ...... 184 Appendix ...... 187 Supplementary Material Chapter 3 ...... 187 Supplementary Material Chapter 4 ...... 191 Supplementary Material Chapter 5 ...... 195 References ...... 197 Curriculum Vitae ...... 207 List of Publications ...... 209
x
List of Figures
Figure 2.1. Localization of Broca’ and Wernicke’s area ...... 24
Figure 2.2. Model of serial processing in speech production ...... 26
Figure 2.3. Serial two - system architecture of lexical selection and from encoding ...... 28
Figure 2.4. Anatom ical and cytoarchitectonic details of the left hemisphere...... 30
Figure 2.5. Fluency tasks tapping different stages of the serial model of word production ... 31
Figure 2.6. Effects of cue type (phono logical vs. semantic) and task demand (high vs. low) . 57
Figure 3.1. Bibliometric overview of 4136 published journal articles on verb al fluency ...... 65
Figure 3.2. Factor solutions of the exploratory factor analysis in healthy young adults ...... 73
Figure 3.3. Venn diagram depicting the two factors’ unique and common shares of explained
variance for the variation in inter - individual differences in the ve rbal fluency task...... 74
Figure 3.4. Venn diagram depicting the two factors’ unique and common shares of explained
variance in stroke patients ...... 77
Figure 4.1. Lesion overlays ...... 100
Figure 4.2. Illustrations of the double dissociation based on the significant two - way
interaction cue type × lesion location ...... 107
Figure 4.3. Overview of the voxel - wise results in the voxel - based lesion - behavior mapping
analysis...... 110
Figure 5.1. Effect of cue type (semantic vs. phonological) and item difficulty (easy vs. hard)
...... 160
Figu re 5.2. Pairwise correlation for the matching variable (A) age and (B) education...... 162
Figure 5.3. Bivariate product - moment correlati ons ...... 166
Supplementary Figure S3.1. Scree plot of the exploratory factor analysis in healthy young
adults ...... 189
Supplementary Figure S3.2. Output of the two - factor model for the CFA ...... 189
xi
Supplementary Figure S3.3. Factor solutions of the exploratory factor analysis in stroke
patients ...... 190
Supplementary Figure S4.1. Lesion overlays for the subsample of patients ...... 194
Supplementary Figure S4.2. Illustrations of the double dissociation based on the significant
two - way interaction cue type × lesion location ...... 194
xii
List of Tables
Table 2.1. Overview of lesion studies investigating verbal fluency ...... 39
Table 2.2. Overview of functional neuroimaging studies investigating verbal fluency ...... 51
Table 3.1. Results of the confirmatory factor analysis (CFA) for the clinical sample of 174
stroke patients ...... 75
Table 4.1. Overview of performance scores for semantic and phonological fluency reported in
previous lesion studies ...... 93
Table 4.2. Overview of clusters with significant voxels for the main effect of lesion ...... 111
Table 4.3. Overview of clusters with significant voxels for the interaction cue type × lesion
...... 115
Table 5.1. Chronological overview of studies reporting o n the test - score and test - retest
reliability of the verbal fluency task ...... 138
Table 5.2. Descriptive Statistics of Performance on the Verbal Fluency Task for Session 1 and
Session 2 in Normal Young Adults ...... 154
Table 5.3. Descriptive Statistics of Performance on the Verbal Fluency Task for Normal Old
Adults and Stroke Patients ...... 154
Table 5.4. Reliabaility Indices and Bootstrapping Results ...... 156
Supplementary Table S3.1. Items of the verbal fluency task ...... 188
Supplementary Table S3.2. Results of the confirmatory factor analysis (CFA) for the clinical
sample of stroke patients without aphasia ...... 188
Supplementary Table S5.1. Descriptive statistics and test - retest reliability for individual items
in healthy young adults ...... 195
xiii
List of Abbreviations
AGFI Adjusted Goodness of F it Index
ATR Anterior Thalamic R adiation
BA Brodmann Areal
BDI - II Beck Depression Inventory - II
CFA Confirmatory Factor A nalysis
COWAT Controlled Oral Word Association Test df Degrees of F reedom
DLPFC Dorsolateral Prefrontal C ortex
EFA Exploratory Factor A nalysis fMRI Functional Magnetic Resonance I maging
GFI Goodness of F it I ndex
HC Healthy C ontrol
IFG Inferior Frontal G yrus
IFOF Inferior Fronto - Occipital F asciculus
KMO Kaiser - Meyer - Olkin measure
LH Left H emisphere
LIFG Left Inferior Frontal G yrus
MAP Velicer’s minimum partial test
MFG Middle Frontal G yrus
MoCA Montreal Cognitive Assessment
MRI Magnetic Resonance I maging
xiv
MTG Middle Temporal G yrus
NIX Non - paramentric interaction effects
PET Positron Emission T omography
PFC Prefrontal C ortex phon Phonological v erbal fluency pSTG Posterior part of the Superior Temporal G yrus
RH Right H emisphere
RMSE Root - Mean - Square E rror of approximation
RS Resting S tate
RWT Regensburger Wortflüssigkeitstest sem Semantic verbal fluency
SEM Standard Error of the M ean
SFG Superior Frontal G yrus
SMA Supplementary Motor A rea
SPSS Statistical Package for Social Sciences
STG Superior Temporal G yrus
TBI Traumatic Brain I njury tDCS Transcranial Direct Current S timulation
TIA Transient Ischemic A ttack
TOL - F Tower of London - Freiburg version
VLBM Voxel - wise Lesion - B ehavior M apping
V LSM Voxel - wise Lesion - Symptom M apping
xv
Chapter 1 Introduction 17
“We speak not only to tell other people what we think, but to tell ourselves what we think. Speech is a part of thought.” Oliver Sacks, Seeing Voices
Chapter 1
Introduction
The communication between two people is largely based on speech. Speech processes, which are not only based on speech perception but also on speech production, are therefore very important in everyday life. In addition, speech makes up a large part of our own identity and is one of the most important forms of social interaction as well as the most distinctive human trait (Damasio & Geschwind, 1984; Friederici, 2011; Levelt, Praamstra, Meyer, Helenius, &
Salmelin, 1998; Levelt, Roelofs, & Meyer, 1999). Cognitive researchers and neuroscientists have been concerned with the investigation of speech perception and speech production for decades (Damasio & Geschwind, 1984; Levelt et al., 1989, 1999; Levelt, 2001). Both Speech perception and production are very complex processes that can be divided into semantic (i.e. meaning of words and sentences), syntactic (i.e. structure of words within a sentence), and phonological (i.e. pronunciation of words) parallel operations (Poldrack et al., 1999; Small &
Burton, 2002). In addition to speech itself executive processes are also involved in every day communication. For example, in a conversation between two people the proper content of the conversational topic needs to be recalled from memory and rapid switches between topics and conversational partners need to be controlled in order to properly execute and follow the conversation (Heim, 2005; Ye & Zhou, 2009). Traditionally, lesion data and more recently modern brain imaging procedures have shown that speech is generated in a complex neural network of several brain areas that are situated in the temporal and frontal lobes (Friederici,
2011; Hirshorn & Thompson-Schill, 2006; Price, 2010). Brain injuries as well as brain
18 Chapter 1 Introduction diseases that involve a lesion in this complex network or parts of it reduce a person's ability to either follow a conversation or lead to a reduction in the manner of expression and therefore, communication is aggravated.
Language abilities and executive functioning, which are necessary for adequate speech production, are often examined using a verbal fluency paradigm. The most common test to assess verbal fluency is the verbal fluency task in which the amount of verbal information that is recalled to a given topic is measured (Lezak, Howieson, Bigler, & Tranel, 2012). Most typically verbal fluency is assessed by requiring the examinee to produce words to a given semantic (e.g. words that belong to the category animals) or phonological (e.g. words starting with the given initial letter S) cue in a pre-set timer interval (e.g. 60s; Lezak et al., 2012). In the current literature different versions and variants of the verbal fluency task exist and it is still debated whether semantic and phonological fluency measure common or distinct sets of cognitive processes (see Section 2.3.3; Ardila, Rosselli, & Bateman, 1994; Bizzozero et al.,
2013; Unsworth, Spillers, & Brewer, 2011; Whiteside et al., 2016). On the neural level it has been suggested that semantic and phonological fluency are mediated by distinct brain regions
(see Section 2.3.4; Baldo, Schwartz, Wilkins, & Dronkers, 2006, 2010; Borkowski, Benton, &
Spreen, 1967; Jurado, Mataro, Verger, Bartumeus, & Junque, 2000; Szatkowska, Grabowska,
& Szymanska, 2000; Thompson-Schill, et al., 1998; Troyer, Moscovitch, Winocur,
Alexander, & Stuss, 1998; see Henry & Crawford, 2004, for a meta-analytic review).
However, findings are primarily based on qualitative analyses and this suggested dissociation has not been confirmed in statistical terms.
The present thesis consists of three studies that investigated the verbal fluency task in both healthy adults and stroke patients to resolve open questions found in the literature. The outline of the thesis is as follows:
Chapter 2 will give an introduction into the theoretical background of language abilities and their impairments and a detailed description for speech production and its underlying neural
Chapter 1 Introduction 19 correlates and association to verbal fluency. Findings from the literature of the different verbal fluency task versions and variants on the underlying cognitive processes and neural correlates will be summarized and an overview of the causes and symptoms of a stroke will be given. An introduction to a German version of the verbal fluency task proposed by Katzev and colleagues (2013) that was used as experimental paradigm in the three studies of this thesis will be given as well.
Chapter 3 will present the first study of this thesis that investigated whether semantic and phonological fluency measure distinct or common sets of cognitive processes with the aid of factor analyses.
Chapter 4 represents the findings from the second study that investigated the frequently implied double dissociation between semantic and phonological fluency and temporal and frontal brain areas by assessing verbal fluency of chronic left hemisphere stroke patients.
Chapter 5 will present the third study of this thesis that investigated the psychometric properties of the German version of the verbal fluency task that has been applied in the two previous studies.
Finally, Chapter 6 will summarize the key findings of the previous studies and both integrate and discuss them with the actual state of the current literature. In addition, limitations and directions for future studies will be addressed.
Chapter 2 Theoretical Background 21
“As you know, in most areas of science, there are long periods of beginning before we really make progress.” Eric Kandel
Chapter 2
Theoretical Background
2.1 Language Abilities and Impairments
Language is a complex cognitive function that is uniquely developed in the human species
(Indefrey & Levelt, 2000; Friederici, 2011). Neuropsychological observations have demonstrated that in the majority of individuals language is mainly controlled by the left hemisphere of the brain (Roby-Brami, Hermsdorfer, Roy, & Jacobs, 2012) and therefore, often impaired after left hemisphere brain damage such as stroke. Language processes and the question of which brain areas are involved in these processes have been of great interest for several decades (Gazzaniga, 1989; Demonet et al., 1992; Levelt, 1999; Ullmann, 2001; Small
& Burton, 2002; Heim, 2005; Friederici, 2011). During the last centuries numerous different methods have been used to investigate and understand the neural underpinnings of speech production and speech perception (Hughlings Jacksons, 1878; Wood, Goff, & Day, 1971;
Netsell & Daniel, 1974; Weiller et al., 1995; Indefrey & Levelt, 2000; Levelt, 2001; Small &
Burton, 2002; Dronkers et al., 2004; Heim, 2005;). Most of our knowledge about language and speech stems from the investigation of patients with aphasia, which is an acquired neurological language disorder that is most often caused by stroke. The earliest studies investigating language abilities and its impairments have been based on the combination of the investigation of speech behavior during lifetime and anatomical examinations of the brain post-mortem (Broca, 1861; Wernicke, 1874; Friederici, 2011). In the 19th century the two neurologists Pierre Paul Broca and Carl Wernicke first described the relation between
22 Chapter 2 Theoretical Background function of speech and neural anatomy. They found that speech can be subdivided into two different components, namely speech production and speech perception, which were differentially impaired by distinct lesion locations in the brain (Broca, 1861; Wernicke, 1874;
Heim, 2005; Friederici, 2011). Speech production is the process to convert thoughts into words and sentences, which includes the appropriate selection of words, use of correct grammatical rules as well as proper articulation of the sounds for each word. Speech perception involves the understanding and interpretation of sounds that are either read or heard (Broca, 1861; Heim, 2005; Price, 2010).
In his report Broca (1861) presented the brain of a deceased patient who had a documented speech production deficiency during his lifetime. The behavioral pattern of this patient, while still alive, showed that he was capable to understand written and spoken language, but speech production was heavily impaired. Examining his brain post-mortem, Broca found a circumscribed lesion in the frontal lobe and more specifically in the posterior part of the left inferior frontal gyrus (LIFG), which was later termed Broca’s areal. From this he concluded that the LIFG was involved in the component of speech production, which corresponds to the
Brodmann Areal (BA) 44 and 45 (Fig. 2.1). Patients with Broca aphasia show a slow and non- fluent speech production that is characterized by agrammatical sentences and the omission of words (Goodglass, 1993; Acharya & Dulebohn, 2017). Several years later and independent of the findings of Broca, Carl Wernicke described a patient with no difficulties in speech production, but remarkable impairments in the component of speech perception and comprehension. Again, examining the brain of this patient post-mortem, Wernicke found a circumscribed lesion in the posterior part of the left superior temporal gyrus (pSTG), which was later termed Wernicke’s areal and corresponds to BA 22 (Fig. 2.1; Wernicke, 1874;
Heim, 2005). Although patients with Wernicke aphasia show fluent speech as compared to those with Broca aphasia, this speech is meaningless because words cannot be unambiguously recalled from the mental lexicon which leads to semantic, phonological, or morphological
Chapter 2 Theoretical Background 23 approximation of the target word and hence neologisms. Since the late 20th century, these early studies that investigated the brain anatomy post-mortem were complemented by functional imaging studies that started to emerge with the invention of methods such as positron emission tomography (PET) or functional magnetic resonance imaging (fMRI). One of the early studies using PET could clearly distinguish the prevalent loci of Broca and
Wernicke aphasia (Damasio & Geschwind, 1984; Small & Burton, 2002). These have shaped our understanding of the neural substrates and functional interplay of brain regions involved in the network of speech processes (Demonet et al., 1992; Small & Burton, 2002). Using new methods recent studies could further show that the two areas described by Broca and
Wernicke are connected with each other through the extreme capsule (Saur et al., 2008;
Weiller et al., 2011) and that not only lesions in these circumscribed brain areas but also lesions in their connecting fibers, such as the fasciculus arcuatus (Geschwind et al., 1968;
Demonet et al., 1992), were found to cause speech deficits (Saur et al., 2008; Weiller et al.,
2011; Li et al., 2017). Although the classical model of speech as suggested by Broca and
Wernicke still serves as a basis for the understanding of speech perception and speech production, recent studies showed that the two components are not solely processed in the two described brain areas but also involve the activation of other brain areas which form larger and partly overlapping neural networks (Price, 2010).
The main focus of this thesis will be on speech or word production which began to be studied systematically in the late 1960s (Cohen, 1966; Levelt, 1999). Speech production is most commonly assessed using picture naming tasks, such as the Boston Naming Test (Ferraro &
Lowell, 2010). But also other tasks, such as the verbal fluency task are increasingly used lately. Speech production tasks have been the most frequently applied language tasks in neurocognitive research for decades (Indefrey & Levelt, 2000). A detailed description of speech production and its neural correlates will be given in the following section.
24 Chapter 2 Theoretical Background
Figure 2.1. Localization of Broca’ and Wernicke’s area (http://thebrain.mcgill.ca).
2.2 Speech Production
Speech production is the process to convert thoughts into words and sentences, which includes the appropriate selection of words, use of correct grammatical rules as well as proper articulation of the sounds for each word (Indefrey & Levelt, 2000; Heim, 2005; Price, 2010).
Although most studies have investigated speech production by asking the examinee to produce single words, one should keep in mind that also larger speech production processes are required for an adequate communication (Indefrey & Levelt, 2000). The different steps involved in speech production will be explained with reference to the serial model proposed by Levelt (1999) (Fig. 2.2; taken from Indefrey & Levelt, 2000). In this model Levelt suggested that there are two underlying systems of language production (Levelt, 1999;
Indefrey & Levelt, 2000). The first system is called the rhetorical/semantic/syntactic system, which includes the conceptual preparation of a message as well as grammatical encoding of this message into a linguistically expressible and lexicosyntactic structure (Indefrey & Levelt,
2000). In order to say something in a conversation, a preparation of a preverbal representation
(e.g. message) is commenced first (Levelt, 1999; Indefrey & Levelt, 2000; Indefrey, 2011).
Chapter 2 Theoretical Background 25
Second, knowledge of the internal and external world is taken into consideration and the semantic related lexical entries are activated from the mental lexicon (i.e. a lemma is chosen).
After the process of selecting the appropriate lemma from the mental lexicon that is linked to the conceptual preparation of a message, the second system, which is called the phonological/phonetic system steps into action. The appropriate articulation of the lemma and the corresponding sound properties are selected and are phonologically encoded. For the retrieval of a single word, phonemes are combined to syllables and an articulatory representation is formed. This articulatory representation is then used as a blueprint for the activation of the appropriate speech muscles (Levelt, 1999; Indefrey & Levelt, 2000; Indefrey,
2011). Hence, semantics (i.e. concepts), syntactics (i.e. lemmas), and phonology of a word are retrieved during different stages of the production process to form the correct verbal output
(Levelt, 1999; Indefrey & Levelt, 2000; Heim et al., 2008). Moreover, during speech production a control mechanism (self-monitoring) is activated which compares the articulatory representation as well as the final utterance with the initial conceptual preparation of the message (Levelt, 1999). The same brain areas are activated whether we listen to ourselves or to another person while talking (McGuire, Silbersweig, & Frith, 1996; Indefrey
& Levelt, 2004). Corrections can take place even before a word is said overtly (Levelt, 1999;
Indefrey & Levelt, 2000; Heim et al., 2008).
26 Chapter 2 Theoretical Background
Figure 2.2. Model of serial processing in speech production as proposed by Levelt (1999) (from Indefrey & Levelt, 2000).
The preceding paragraph demonstrated the procedure of speech production in a spontaneous conversation. However, speech production varies if a person is requested to say a word in accordance to a specific criterion, as for example in verbal fluency tasks , where the examinee is required to produce words to a given category or starting with an initial letter (Lezak et al.,
2012). This engages specific search processes and spontaneous utterances of own thoughts need to be suppressed. This type of speech production is most frequently studied using
Chapter 2 Theoretical Background 27 picture-naming tasks, which will be used in the following to explain the process of task-based word production (Levelt, 2001). As described in Levelt (2001) naming of a presented picture
(e.g. horse) requires the selection of the appropriate word from the mental lexicon (i.e.
‘lexical selection’) followed by the generation of the correct articulatory shape (i.e. ‘form encoding’) (Fig. 2.3; Levelt, 2001). During the process of lexical selection coactivations of different related concepts to the target word take place, which are also dependent on the subject’s interpretation of the task (Levelt, 2001). A correct answer to the picture depicting a horse might either be simply horse, but also stallion or animal (Levelt, 2001). The subject needs to focus on a concept that seems to be the most appropriate in this case, which is called
‘perspective thinking’ (Levelt, 2001). During this step related concepts are also coactivated, which further spread activation to the corresponding lexical item (i.e. lemma) in the mental lexicon. Once the correct lemma is chosen the form encoding system is activated and the correct phonological code is selected (Levelt, 2001). Regarding the verbal fluency task, which is the main neuropsychological measure used in this thesis and also requires task-based word production, similar processes are needed analogous to picture-naming. However, verbal fluency, which is characterized by a rapid intrinsic word generation, also requires the examinee to initiate effective retrieval strategies in order to organize thinking and aid word generation (see Section 2.3; Troyer et al., 1997; Abrahams et al., 2003). Hence, verbal fluency not only engages processes of word production and retrieval but also strongly engages executive processes (Troyer et al., 1997; Abrahams et al., 2003).
The neural correlates underlying the different core processes of word production, described above will be explained in the following section.
28 Chapter 2 Theoretical Background
Figure 2.3. Serial two-system architecture of lexical selection and from encoding (Levelt, 2001).
2.2.1 Neural Correlates of Word Production
The neural correlates related to speech production have been studied with various tasks, such as picture-naming or word generation, which are thought to include all underlying components of word production. In their meta-analyses, Indefrey & Levelt (2000, 2004) concluded that the word production network involves several brain areas of the left and right hemisphere. Brain areas of the right hemisphere involved in speech production include the mid superior temporal gyrus, the medial and lateral cerebellum as well as supplementary motor area (SMA), whereas brain areas of the left hemisphere include the posterior inferior frontal gyrus, the ventral precentral gyrus, SMA, mid and posterior superior and middle temporal gyri, posterior temporal fusiform gyrus, anterior insula, thalamus, and medial cerebellum (Indefrey & Levlt, 2004). Considering the neural correlates of the core processes
Chapter 2 Theoretical Background 29 of word production as have been proposed by Indefrey & Levlt (2000, 2004), the following association of brain areas has been found (see also Indefrey, 2011). The first process, lexical selection, involves the activation of the mid-section of the left middle temporal gyrus. For the phonological code retrieval the following areas are involved: the right SMA, left anterior insula, and the left posterior superior and middle temporal gyri (Wernicke’s area). For the process of syllabification the left posterior inferior frontal gyrus (Broca’s area) has been found to be primarily activated. Considering the comparison of covert and overt speech, in overt speech additional areas such as left ventral precentral gyrus, bilateral mid superior temporal gyri, left posterior fusiform gyrus, left thalamus, as well as the right medial cerebellum, have been found to be activated (Indefrey & Levelt, 2000, 2004; Indefrey, 2011). The process of articulation requires mainly the activation of brain areas that are part of the central nervous motor system (bilateral ventral motor and sensory regions, right dorsal motor region, right
SMA, left and medial right cerebellum, bilateral thalami, and right midbrain), but also other areas such as, right posterior inferior frontal gyrus, left orbital gyrus, bilateral posterior lingual gyrus, and right posterior medial fusiform gyrus have been associated with the process of articulation. The control process of self-monitoring is found to be associated with activation in bilateral superior temporal gyri (Indefrey & Levelt, 2000, 2004; Indefrey, 2011).
The brain areas associated with the different core processes of speech production are depicted in Figure 2.4 (Figure taken from Friederici, 2011).
30 Chapter 2 Theoretical Background
Figure 2.4. Anatomical and cytoarchitectonic details of the left hemisphere. The major language relevant gyri (IFG, STG, MTG) are color coded. Numbers indicate language- relevant Brodmann Areas (BA) which Brodmann (1909) defined on the basis of cytoarchitectonic characteristics. Broca’s area consists of the pars opercularis (BA 44) and the pars triangularis (BA 45). Wernicke’s area is defined as BA 42 and BA 22 (Friederici, 2011).
As the verbal fluency task also involves word production similar processes and associated brain areas are required for an adequate retrieval of words in this task. However, verbal fluency comprises both a semantic and phonological variant. They tap different stages during the process of word production and are thought to require distinct brain areas (Heim et al.,
2008). The different variants of the fluency task tapping different stages of the serial model of speech production as proposed by Levelt (1999) can be seen in Figure 2.5 (Heim et al., 2008).
Chapter 2 Theoretical Background 31
Figure 2.5. Fluency tasks tapping different stages of the serial model of word production (Heim et al., 2008).
Even though the verbal fluency task is one of the most prominent and often applied neuropsychological measures of word production (Indefrey & Levelt, 2000; Strauss, Sherman,
& Spreen, 2006; Lezak et al., 2012) there exist still some discrepancies in the literature regarding its underlying cognitive processes and neural correlates especially with regard to dissociating semantic and phonological verbal fluency. Therefore, in the following section the theoretical underpinnings of the verbal fluency task and insights from the current literature regarding cognitive processes and neural correlates will be illustrated first. Second, open questions addressed in this thesis will be highlighted. This will be followed by a more specific
32 Chapter 2 Theoretical Background description of a German version of the verbal fluency task applied in the studies included in this thesis.
2.3 Verbal Fluency
2.3.1 Semantic and Phonological Verbal Fluency
Verbal fluency (e.g., Milner, 1964; Benton, 1968; Borkowski, Benton, & Spreen, 1967) is one of the most frequently used neuropsychological measures of language abilities and executive functioning (Moscovitch, 1994; Strauss, Sherman, & Spreen, 2006; Lezak et al., 2012; Shao,
Janse, Visser, & Meyer, 2014; Chouiter et al., 2016) and further measures the efficiency of lexical access (Jurado & Rosselli, 2007). Verbal fluency is defined as a cognitive function that facilitates information retrieval from memory and measures an individual’s ability to retrieve specific information within restricted search parameters (Patterson, 2011; Lezak et al., 2012).
Therefore, executive control is required over several cognitive functions such as selective attention and inhibition, mental set shifting, internal response generation, and self-monitoring to exhibit successful retrieval (Patterson, 2011). Because of the executive component in verbal fluency tasks, its extensive use is based on the belief that verbal fluency tasks are the most sensitive to dysfunctions of the frontal lobes (Stuss & Benson, 1986; Jurado & Rosselli,
2007; Chapados & Petrides, 2013).
There are two common variants of the verbal fluency task The first variant in which participants are required to name as many words as possible to a given category (e.g. animals) is called semantic verbal fluency. The second variant in which participants are required to name as many words as possible starting with a given letter (e.g. S) is called phonological verbal fluency. To perform well in these tasks an intact phonological and semantic knowledge as well as the ability for an adequate and coordinated retrieval is indispensable (Shao et al.,
2014; Bose, Wood, & Kiran, 2017). For both fluency variants the most typical outcome measure is the amount of words produced in a preset time interval (e.g. 60 sec; Strauss,
Chapter 2 Theoretical Background 33
Sherman, & Spreen, 2006; Lezak et al., 2012). Other qualitative measures such as the amount of clusters (e.g. the production of words within semantic or phonemic subcategories) and switches (e.g. the ability to switch between clusters) as well as the number of intrusions or perseverations have been assessed to learn more about the applied search strategies or to investigate more subtle deficits for example in clinical samples (Troyer et al., 1997; Reverberi et al., 2006). The first verbal fluency task described in the literature was developed by Louise
Thurstone in 1938. Participants had to write as many words as possible beginning with the letter S within a time interval of 5 minutes followed by writing as many words as possible beginning with the letter C within a time interval of 4 minutes (Thurstone, 1938; Kolb &
Whishaw, 2008). In the English-speaking world, a related version of the verbal fluency task is the Controlled Oral Word Association Test (COWAT) in which participants have to name words beginning with the letters F, A, and S. In German-speaking countries the Regensburger
Wortflüssigkeitstest (RWT; Aschenbrenner, Tucha, & Lange, 2000) which includes 14 subtests of semantic and phonological fluency (Aschenbrenner, Tucha, & Lange, 2000) is commonly used. In contrast to the Thurstone Word Fluency Test (Thurstone, 1938), the two latter variants of the verbal fluency task require the examinee to orally present the correct answers. The time limit for word production has been reduced from 4 and 5 minutes
(Thurstone, 1938) to commonly 60s and in rare cases to 2 minutes (Strauss, Sherman, &
Spreen, 2006; Kolb & Whishaw, 2008). The most common variant used in previous neuropsychological assessments and research studies is that of the oral form (e.g. Phelps et al., 1997; Lezak et al., 2012), although also covert word retrieval studies have been conducted
(Curtis et al., 1998; Perani et al., 2003). Especially in the functional neuroimaging literature the covert variant has been preferred as even small head movements may lead to excessive scanner artifacts (Paulesu et al., 1997; Birn, Bandettini, Cox, Jesmanowicz, & Shaker, 1998;
Huang, Carr, & Cao, 2002; Hirshorn & Thompson-Schill, 2006; Golestanirad, Das,
Schweizer, & Graham, 2015). In overt speech words are said aloud, in covert speech words
34 Chapter 2 Theoretical Background are only imagined. Although scanner artifacts might be kept at a minimum with covert speech production, there are several limitations. First, the behavioral output cannot be recorded or monitored. Second, processes of phonological encoding or articulation cannot be investigated
(Levelt, 2001). In addition, to the differentiation between covert and overt speech verbal fluency condition, there are also two different modes for the generation of words (Basho et al., 2007). These are called paced and unpaced/self-paced. In the paced mode an acoustic signal is given upon which the examinee has to generate a word, in the self-paced mode the examinee can freely generate as many words as possible in a given time interval (Basho et al.,
2007). Since for the current thesis the behavioral output with the total amount of words said was the most important determinant of verbal fluency performance and the task was not performed in a scanner, an overt and self-paced paradigm was chosen (see Section 2.5).
There are several different cognitive processes required for the generation of words in semantic and phonological fluency. Although generation of words to a semantic or phonological cue is partially based on common cognitive processes, also differences between these two variants exist, especially with regard to the search processes required for successful retrieval (Henry & Crawford, 2004). In the following section an overview of the cognitive processes involved in semantic and phonological fluency will be given. Inconsistencies found in the literature regarding this matter will be highlighted and the aims of the first study (see
Chapter 3) will be given.
2.3.2 Cognitive Processes of Verbal Fluency
Although both semantic and phonological fluency require common cognitive abilities such as processing speed and attention (van der Elst et al., 2006; Biesbroek et al., 2016), working memory (Baldo et al., 2006; Robinson et al., 2012), language functioning (Ruff et al., 1997;
Li et al., 2017), and mental flexibility (Troyer et al., 1998), several different abilities are also required for each variant (Parks et al., 1992; Rosen & Engle, 1997; Schwartz et al., 2003;
Chapter 2 Theoretical Background 35
Shao et al., 2014). Since both types of verbal fluency tasks require the examinee to select one correct answer out of potentially several correct answers and suppress incorrect answers, they might require common cognitive processes. For both types of verbal fluency self-monitoring of oral output or suppression of previously retrieved responses is important as repeated naming is prohibited (Weiss et al., 2003; Azuma, 2004; Katzev et al., 2013). A fluent word production requires an effective search strategy. Findings from previous studies suggest that these search strategies differ with respect to the type of verbal fluency (Li et al., 2017) and further corroborate the idea of different cognitive processes to be required for the two variants of verbal fluency (Ruff et al., 1997; Unsworth et al., 2011; Biersbroek et al., 2016). Semantic fluency is most likely driven by association chains and spreading activation within cue-related subcategories (Gruenewald & Lockhead, 1980; Katzev et al., 2013), whereas phonological fluency involves the systematic syllabification of initial letters (Mummery et al., 1996; Rende et al., 2002). It is thought that semantic fluency is usually easier than phonological fluency.
These differences in difficulty level between semantic and phonological fluency have been explained by the fact that retrieving words according to a semantic rule relies on the natural semantic organization of our conceptual knowledge and therefore, this strategy is easier to apply and can be executed more automatically (Simanova et al., 2014; Chouiter et al., 2016).
In semantic verbal fluency tasks words can be easily retrieved by its meaning, whereas this natural habit needs to be suppressed in phonological verbal fluency tasks. In this task variant, proper syllabification is necessary for good performance (Mummery et al., 1996; Rende et al.,
2002). Regarding the serial model of word production (Levelt, 1999), the two different types of verbal fluency are thought to tap different stages within the model and this corroborates the idea of distinct cognitive processes for semantic and phonological verbal fluency (Fig. 5;
Levelt, 1999; Heim et al., 2008).
Not only differences between semantic and phonological fluency in general can be observed, but also differences in difficulty levels between two semantic categories or between two
36 Chapter 2 Theoretical Background letters can be found (Lezak et al., 2012; Katzev et al., 2013). For example, depending on the amount of words to be said (i.e. size of the search space) for a given category or letter, retrieval might be easier for one letter but harder for the other letter (Katzev et al., 2013).
Findings from lesion and neuroimaging studies, which will be discussed in more detail in the following section (see Section 2.3.3), confirm this notion of distinct cognitive processes.
These studies found that semantic and phonological fluency differentially rely on temporal and frontal brain areas, respectively and it has been suggested that probably also different cognitive processes are involved in the two task variants (Baldo & Shimamura, 1998;
Szatkowska, Grabowska, & Szymanska, 2000; Baldo et al., 2006; Chapados & Petrides, 2013; see Henry & Crawford, 2004 for a meta-analysis).
Findings from factor analytic approaches challenge these findings of dissociable cognitive processes for semantic and phonological fluency (Ardila, Rosselli, & Bateman, 1994;
Unsworth, Spillers, & Brewer, 2011; Bizzozero et al., 2013; Whiteside et al., 2016). Factor analytic approaches found that both semantic and phonological fluency loaded on only one factor, suggesting that they measure the same cognitive processes. Potential limitations of these previous studies using factor analytical approaches exist, because a very limited and partly disparate number of items have been applied for assessing semantic and phonological fluency. These previous studies not only used a limited number of items, but also compared measures of semantic and phonological fluency with other cognitive constructs, such as tests of executive functions, language, working memory capacity, or processing speed (Ardila,
Rosselli, & Bateman, 1994; Unsworth, Spillers, & Brewer, 2011; Bizzozero et al., 2013;
Whiteside et al., 2016). As has been stated above, semantic and phonological fluency can be expected to share common cognitive processes at least to some extent. By comparing semantic and phonological fluency with a variety of other cognitive constructs, a direct and unbiased comparison between the two fluency sub-tasks is lacking. Drawing firm conclusions about whether semantic and phonological fluency measure distinct or common cognitive
Chapter 2 Theoretical Background 37 processes may not be allowed (Schmidt et al., 2017). By using factor analytical approaches with measures of semantic and phonological fluency only, the first study of this thesis approaches these inconsistencies as well as methodological limitations of previous studies.
The question of whether semantic and phonological fluency measure distinct or common cognitive processes (see Chapter 3) will be addressed in a sample of healthy adults and chronic stroke patients (Schmidt et al., 2017).
In the following, findings from the lesion and neuroimaging literature regarding the underlying neural correlates of verbal fluency will be discussed. Furthermore, the relevance and aims of the second study included in this thesis (see Chapter 4) will be presented.
2.3.3 Neural Correlates of Verbal Fluency
Several different brain areas are involved in the processes of speech production and also during verbal fluency tasks. As can be seen from the previous section behavioral studies investigating the cognitive components suggest dissociable cognitive processes for semantic and phonological fluency, which further lead to the suggestion that also dissociable brain areas are required for the two types of verbal fluency. These differences have further been attributed to the different search strategies required for the two task variants (see Section
2.3.2; Mummery et al., 1996; Rende et al., 2002; Katzev et al., 2013; Simanova et al., 2014;
Chouiter et al., 2016). In general, it is thought that the temporal lobe of the brain is involved in semantic fluency, whereas phonological fluency relies more on the activation of the frontal lobe (Baldo et al., 2006). Our knowledge and memories are stored in semantic clusters and the brain areas involved for memory consolidation and retrieval are situated in the temporal lobe
(Levy, Beyley, & Squire, 2004). Although in phonological fluency words are also retrieved from memory, the search strategy used for this type of fluency is different as compared to semantic fluency (see Section 2.3.2). Here, successful syllabification and inhibition of retrieving words by its meaning are indispensable for appropriate retrieval. The brain areas
38 Chapter 2 Theoretical Background involved in inhibition such as the prefrontal cortex are situated in the frontal lobe and phonological fluency has been suggested to rely on frontal lobe functioning (Milner, 1964).
These general assumptions have further been postulated from both lesion as well as functional imaging studies which will be reviewed in this section. Lesion studies found that decreased performance in phonological fluency was associated with lesions in more anterior regions, including the left frontal cortex, whereas decreased performance in semantic fluency was associated with lesions in more posterior regions, primarily the left temporal cortex (Baldo &
Shimamura, 1998; Szatkowska, Grabowska, & Szymanska, 2000; Baldo et al., 2006;
Chapados & Petrides, 2013; see Henry & Crawford, 2004 for a meta-analysis; see Table 2.1 for an overview of lesion studies). For example, one of the earliest studies conducted by
Milner in this realm, which used a phonological verbal fluency paradigm, found that patients with left frontal lobe lesions were significantly impaired when compared to patients with left temporal lesions (Milner, 1964). Furthermore, Jurado and colleagues (2000) who investigated traumatic brain injured (TBI) patients with focal frontal lesions found that TBI patients produced fewer words for phonological fluency but not for semantic fluency when compared to healthy controls (Jurado et al., 2000). A recent voxel-based lesion-symptom mapping
(VLSM) analysis in left-hemisphere stroke patients found that semantic and phonological fluency deficits correlated with lesions in temporal and frontal cortices, respectively (Baldo et al., 2006). Based on the findings and in agreement with the prevailing hypothesis, the authors suggested that the temporal cortex primarily subserves word retrieval in semantic fluency whereas the frontal cortex is primarily involved in phonological fluency (Baldo et al., 2006,
2010).
Chapter 2 Theoretical Background 39
Table 2.1. Overview of lesion studies investigating verbal fluency
Study Author Year Sample Description Verbal Fluency Task Results
Semantic Fluency Phonological Fluency
1 Milner 1964 LH Frontal (n = 7) N.A. Thurstone’s word LH Frontal < LH Temporal LH Temporal (n = 7) fluency test LH Frontal < RH Frontal RH Frontal (n = 4) 2 Borkowski 1967 LH Brain-damaged (n = 10) N.A. 1. J, U Brain-damaged < HC et al. RH Brain-damaged (n = 9) 2. N, G Bilateral Brain-damaged (n 3. P, T = 11) 4. F, S HC (n = 30) 3 Benton 1968 LH Frontal (n = 10) N.A. F, A, S LH Frontal < RH Frontal RH Frontal (n = 8) Bilateral Frontal < RH Frontal Bilateral Frontal (n = 7) No difference between LH and Bilateral Frontal 4 Perret 1974 LH Frontal (n = 23) N.A. S, B LH < RH < HC LH Temporal (n = 15) LH Frontal < LH Temporal and Posterior LH Posterior (n = 18) LH Frontal < Posterior RH Frontal (n = 27) No difference between Temporal and Posterior RH Temporal (n = 17) groups RH Posterior (n = 18) HC (n = 26) 5 Miceli et al. 1981 LH Brain-damaged N.A. F, A, S LH < RH RH Brain-damaged 6 Pendleton 1982 LH Frontal (n = 20) N.A. Thurstone’s word All Brain-damaged < HC et al. RH Frontal (n = 23) fluency test Frontal < HC LH Nonfrontal (n = 22) Mixed < HC RH Nonfrontal (n = 17) Nonfrontal < HC LH Mixed (n = 16) RH Mixed (n = 16) HC (n = 134) 7 Miller et al. 1984 LH Frontal Lesion (n = 15) N.A. F, A, S LH Frontal < HC LH Posterior Lesion (n = 15) RH Frontal < HC RH Frontal Lesion (n = 15) Patients with Dementia < HC
40 Chapter 2 Theoretical Background
RH Posterior Lesion (n = 15) Patients with Dementia (n = 20) HC (n = 30) 8 Joanette & 1986 RH Vascular Lesion (n = 35) Animal, Furniture B, R Patients < HC (sem) Goulet HC (n = 20) 9 Wertz et al. 1986 LH Brain-damage (n = 40) N.A. S, T, P, C LH < RH < HC RH Brain-damage (n = 40) No difference between RH and Bilateral Bilateral Brain-damage (n = 40) HC (n = 40) 10 Janowsky 1989 LH Frontal (n = 2) N.A. F, A, S LH Frontal < HC et al. RH Frontal (n = 3) Bilateral Frontal < HC Bilateral Frontal (n = 2) No difference between RH Frontal and HC HC (n = 11) 11 Butler et al. 1993 LH Frontal (n = 4) N.A. F, A, S All Tumor Patients (n = 17) < HC LH Fronto-Temporal (n = 2) RH Frontal (n = 6) RH Fronto-Temporal (n = 2) Bilateral Frontal (n = 2) Bilateral Fronto-Temporal (n = 1) HC (n = 17) 12 Martin et 1990 LH TLE (n = 15) Animals, C, F, L, P, R, W LH TLE < RH TLE (both tasks) al. RH TLE (n = 17) Fruits/Vegetables LH TLE < HC (both tasks) HC (n = 25) RH TLE < HC (both tasks) 13 Owen et al. 1990 Uni- and Bilateral Frontal Animals (90 sec) F, A, S Patients < HC (both tasks) Lesion (n = 19) HC (n = 19) 14 Loring et 1994 LH Anterior Temporal Animals F, A, S COWAT and semantic fluency decrease after al. Lobectomy (n = 12) both left and right temporal lobectomy in the RH Anterior Temporal acute post-operative stage Lobectomy (n = 11) 1 year post-surgery patients returned to normal baseline assessment before surgery as control
Chapter 2 Theoretical Background 41
15 Stuss et al. 1994 LH Frontal (n = 10) N.A. F, A, S LH Frontal < HC RH Frontal < HC RH Frontal (n = 9) Bilateral Frontal < HC Bilateral Frontal (n = 13) HC (n = 15) 16 Vilkki et al. 1994 LH Anterior Lesion (n = 19) Animals (20) 1 S (10) 1 LH Anterior and Posterior Lesion < RH LH Posterior Lesion (n = 16) Anterior and Posterior Lesion (both tasks) RH Anterior Lesion (n = 10) RH Posterior Lesion (n = 15) 17 Gershberg 1995 LH and RH Frontal Lesion N.A. F, A, S LH Frontal < HC & (n = 7) LH Frontal < RH Frontal Shimamura HC (n = 12) RH Frontal < HC 18 Varley 1995 RH damage (n = 20) Animals, Colors, N.A. RH < HC (total number of responses across all HC (n = 20) Furniture, Fruit, categories) Time Units (90 sec) RH < HC (animals) RH < HC (furniture) 19 De Vreese 1996 LH Brain-damaged (n = 8) Animals, Fruits, P, F, L LH < LH (phon) et al. RH Brain-damaged (n = 8) Car makes LH < HC (both tasks) HC (n = 28) RH < HC (sem) 20 Stuss et al. 1996 LH Frontal (n = 10) Animals F, A, S LH Frontal < Young Aged HC (both tasks) RH Frontal (n = 9) RH Frontal < Young Aged HC (both tasks) Bilateral Frontal (n = 13) Bilateral Frontal < Young Aged HC (both tasks) Young Aged HC (n = 20) LH Frontal < Middle Aged HC (both tasks) Middle Aged HC (n = 20) Bilateral Frontal < Middle Aged HC (both Old Aged HC ( n = 20) tasks) LH Frontal < Old Aged HC (both tasks) Bilateral Frontal < Old Aged HC (both tasks) 21 Goulet et 1997 RH Stroke Patients (n = 15) Animals, Clothes, P, M, T, V, L, N RH Stroke Patients < HC (both tasks) al. Sports, Vegetables, HC (n = 15) Tools, Weapons 22 Baldo et al. 1998 LH Frontal Lesion (n = 6) Animals, Fruits, F, A, S LH Frontal < HC (both tasks) RH Frontal Lesion (n = 6) Occupations RH Frontal < HC (both tasks) HC (n = 12) 23 Baldo & 1998 All Frontal Lesion (n = 12) Animals, Fruits, F, A, S LH Frontal < HC (both tasks) Shimamura LH Frontal Lesion (n = 6) Occupations RH Frontal < HC (both tasks)
42 Chapter 2 Theoretical Background
RH Frontal Lesion (n = 6) HC (n = 12) 24 Helmstadte 1998 FLE (n = 33) N.A. F, R, K Pre-operatively FLE < TLE r et al. LH Frontal Lobe Surgery (n = 17) RH Frontal Lobe Surgery ( n = 16) TLE (n = 45) LH Temporal Surgery (n = 21) RH Temporal Surgery (n = 24) HC (n = 22) 25 Rogers et 1998 All Frontal Lesion (n = 12) Animals (90 sec) F, A, S Frontal Lesions < HC (both tasks, tendency for al. LH Frontal Lesion (n = 6) larger reduction in phon than sem) RH Frontal Lesion (n = 6) LH Frontal < RH Frontal (both tasks) HC matched with Frontal LH Frontal < HC (both tasks) Lesion (n = 14) PD Patients (n = 12) HC matched for PD (n = 12) 26 Stanhope et 1998 Diencephalic Lesions (n = F, A, S Frontal Lesions < HC al. 15) Temporal Lesions (n = 14) Frontal Lesions (n = 15) HC (n = 12) 27 Stuss et al. 1998 LH Frontal (n = 20) Animals F, A, S LH Frontal < HC (both tasks) LH Non-Frontal (n = 11) LH Non-Frontal < HC (both tasks) RH Frontal < HC (more pronounced for sem) RH Frontal (n = 19) RH Non-Frontal = HC (both tasks) RH Non-Frontal (n = 9) Bilateral Lesions (n = 15) HC (n = 37) 28 Thomson et 1998 RH Tumor (n = 33) N.A. A, B no control group, no statistical analysis al. 29 Troyer et 1998 LDLF (n = 14) Animals F, A, S LDLF < HC (both tasks) al. RDLF (n = 11) SMF < HC (both tasks) SMF (n = 17) RDLF < HC (sem)
Chapter 2 Theoretical Background 43
IMF (n = 11) IMF < HC (sem) LTL (n = 9) LTL < HC (sem) RTL (n = 14) HC matched with FL patients (n = 37) HC matched with TL patients (n = 18) 30 Cohen et al. 1999 Bilateral Frontal (n = 18) Animals N.A. Bilateral Frontal < Controls (on descriptive Patients with Chronic Pain level) as Controls (n = 20) 40 Stuss et al. 1999 LH Frontal (n = 6) Animals F LH Frontal < HC (both tasks, on descriptive RH Frontal (n = 10) level) Bilateral Frontal (n = 6) RH Frontal < HC (both tasks, on descriptive LH Posterior (n = 7) level) RH Posterior (n = 7) LH Frontal < RH Frontal (both tasks, on HC (n = 19) descriptive level) 41 Tucha et al. 1999 LH Frontal (n = 45) N.A. S LH Frontal < HC RH Frontal (n = 50) HC (n = 25) 42 Channon & 2000 LH Frontal (n = 6) N.A. S (5 min) LH Frontal < HC Crawford RH Frontal (n = 13) LH Posterior (n = 4) RH Posterior (n = 8) HC (n = 60) 43 Jurado et al. 2000 TBI all (n = 13) Animals, F, A, S (90 sec) TBI Frontal < HC (phon) RH TBI Lesion (n = 3) Supermarket Items TBI Frontal = HC (sem) LH TBI Lesion (n = 3) (90 sec) Bilateral TBI Lesion (n = 7) HC (n = 26) 44 Leclercq et 2000 RH and LH TBI (n = 16) Animals (2 min) P (2 min) RH and LH TBI < HC (both tasks) al. Bilateral AACA (n = 9) Bilateral AACA < HC (both tasks) HC (n = 25) 45 Szatkowska 2000 LH DLPFC (n = 6) Animals K LH DLPFC < other groups (phon) et al. RH DLPFC (n = 6) LH DLPFC < HC (sem) LH Ventromedial PFC (n = RH DLPFC < HC (sem) 6) RH Ventromedial PFC < HC (sem) RH Ventromedial PFC (n =
44 Chapter 2 Theoretical Background
6) HC (n = 10) 46 Baldo et al. 2001 LH Frontal (n = 6) N.A. F, A, S LH < RH < HC RH Frontal (n = 5) HC (n = 11) 47 Luckhurst 2001 LH Temporal Lobectomy (n Animals, Fruits, N.A. LH Temporal Lobectomy < HC & Lloyd- = 5) Vegetables, RH Temporal Lobectomy < HC Jones RH Temporal Lobectomy (n Clothing, Houshold = 4) items, Furniture (2 HC (n = 8) min) 48 Schwartz & 2001 LH Frontal Lesion (n = 6) Animals N.A. Hierarchical Clustering and Correspondence Baldo RH Frontal Lesion (n = 5) analysis of the 25 most frequently generated Bilateral Frontal Lesion (n = words. 2) Both groups generated the 25 words with the HC (n = 11) same frequency. 49 Stefanova 2002 ACoA Operated Patients (n Animals S, K, L Patients < HC (both tasks) et al. = 30) HC (n = 31) 50 Sylvester & 2002 LH Frontal Lesion (n = 7) Animals N.A. LH Frontal Lesion < HC Shimamura RH Frontal Lesion < HC RH Frontal Lesion (n = 4) HC (n = 10) 51 Goldstein et 2004 Anterior Lesion (LH and RH Animals F, A, S Anterior Lesion < HC (both tasks) al. combined) (n = 26) Posterior Lesion < HC (both tasks) Posterior Lesion (LH and RH combined) (n = 20) HC (n = 57) 52 Baldo et 2006 LH Temporal and Frontal Fruits, Animals, F, A, S Phon associated with LH Frontal Lesion (BA 4, al. 3 Lesion (n =48) Supermarket Items 6, 44) (90 sec) Sem associated with LH Temporal Lesion (BA 22, 37, 41, 42) 53 Davidson et 2008 RH Frontal Lesion (n = 20) Animals F, A, S Patients < HC (both tasks) al. HC (n = 32) 54 Baldo et al. 2010 LH Temporal Lesion (n = 1) Fruits, Animals, F, A, S LH Temporal: phon > sem Supermarket Items
Chapter 2 Theoretical Background 45
LH Frontal Lesion (n = 1) Fruits, Animals, F, A, S LH Frontal: sem > phon Supermarket Items
55 Peterburs et 2010 Cerebellar Lesion (n = 14) Names of Countries B Patients < HC (phon) al. HC (n = 14) 56 Schweizer 2010 LH Cerebellar Lesion (n = Animals F, A, S RH Cerebellar Lesion < LH Cerebellar Lesion < et al. 12) HC (phon) RH Cerebellar Lesion (n = RH Cerebellar Lesion < HC (sem) 10) HC (n = 30) 57 Robinson et 2012 LH and RH Posterior Fruits, Vegetables S LH Frontal < HC (both tasks) al. Lesions (n = 20) RH Frontal < HC (both tasks) LH Frontal Lesion (n = 20) LH Posterior < HC (sem) RH Posterior < HC (sem) RH Frontal Lesion (n = 27) HC (n = 35) 58 Almairac et 2015 LH DLGG surgery (n = 31) Animals (120 sec) P (120 sec) decrease of performance after surgery (both al. tasks) 59 Biesbroek 2016 Ischemic Stroke Patients All Animals A, N LH Lesion < RH Lesion (both tasks) et al. 3 (n = 93) Infratentorial Lesion < RH Lesion (both tasks) LH Lesion (n = 34) Multiple Lesion Locations < RH Lesion (both RH Lesion (n = 40) tasks) Infratentorial Lesion (n = 12) Multiple Lesion Locations (n = 7) 60 Chouiter et 2016 LH Lesion (n=108) Animals M (132 patients) or S phon: associated with lesions in anterior middle al. 3 (59 patients) temporal and superior temporal areas RH Lesion (n=83) sem: associated with lesions of posterior middle temporal gyrus both: left subcortical areas (putamen, caudate nucleus, pallidum) and left cortical areas (superior and middle temporal gyri, angular gyri, insula, parts of supramarginal gyri) 61 Li et al. 2017 Stroke Patients (n=51) Animals, Fruits and /Da4/ 2 Patients < HC (both tasks) Vegetables, Tools /Bu4/ 2 HC (n=39)
46 Chapter 2 Theoretical Background
Note. HC, healthy control; LH, left hemisphere; RH, right hemisphere; TBI, traumatic brain injury; N.A., not available; sem, semantic verbal fluency; phon, phonological verbal fluency; DLPFC, dorsolateral prefrontal cortex; LDLF, left dorsolateral frontal; RDLF, right dorsolateral frontal; SMF, superior medial frontal; IMF, inferior medial frontal; LTL, left temporal lobe; RTL, right temporal lobe; AACA, aneurysma of the anterior communicating artery; ACoA, Anterior Communicating Artery. Note that the number of applied items for phonological and semantic fluency differ in some of the studies. 1 In this study performance was assessed as the time subjects needed to generate 20 and 10 semantically (animals) and phonologically (S) cued words, respectively. 2 In this study Chinese syllables were used as phonological cues. 3 These studies used the voxel-based lesion-symptom mapping analysis approach.
Chapter 2 Theoretical Background 47
Besides the investigation of the involvement of grey-matter regions in verbal fluency the integrity of white-matter networks contributing to the processing in semantic and phonological fluency have also been investigated recently (Li et al., 2017). In their study Li and colleagues (2017) found five left-lateralized white-matter tracts to be significantly correlated with the scores of both semantic and phonological fluency. These are the left anterior thalamic radiation (ATR), left inferior fronto-occipital fasciculus (IFOF), left uncinate fasciculus (UF), left superior longitudinal fasciculus (SLF), and left frontal aslant tract (FAT) (Li et al., 2017). Furthermore, left IFOF, left ATR, and left UF appeared to have unique contributions to semantic fluency when compared to phonological fluency (Li et al.,
2017).
When considering the previous lesion studies it needs to be noted that, apart from some notable exceptions (Baldo et al., 2006, 2010; Biesbroek et al., 2016; Chouiter et al., 2016), most studies investigated only one of the two verbal fluency variants (i.e. either semantic or phonological fluency) (Milner, 1964; Benton, 1968; Butler et al., 1993). Furthermore, for the semantic fluency most commonly one or two items were used whereas for the phonological fluency the letters F, A, S have been assessed most frequently (Table 2.1). As a result, differences in difficulty level between semantic and phonological items were ignored in previous studies. Furthermore, most studies used the region-based approach to investigate the impact of lesions on the performance of the verbal fluency task. That is patients were grouped with regard to their lesion location and then compared with regard to the verbal fluency task applied in the respective study. As the temporal and frontal lobe are both large entities with different and functionally heterogeneous subregions, voxel-wise analyses may reveal more specific results with regard to the underlying neural correlates for semantic and phonological fluency (Chapter 4). Although some studies have investigated both verbal fluency variants
(Table 2.1) findings are primarily based on qualitative measures and not quantitative
48 Chapter 2 Theoretical Background measures. Hence, only associations between semantic and phonological and temporal and frontal brain areas, respectively, have been reported (Milner, 1964; Baldo et al., 2006).
Positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) studies revealed that greater activity in the inferior frontal cortex was associated with phonological fluency, whereas more activation in the left temporal cortex was associated with semantic fluency (Demonet et al., 1992; Mummery et al., 1996; Gourovitch et al., 2000;
Costafreda et al., 2006; Birn et al., 2010; Katzev et al., 2013; Wagner et al., 2014; see Table
2.2 for an overview of neuroimaging studies). For example, one of the earliest studies used
PET in healthy subjects to investigate the differences in regional cerebral blood flow for semantic and phonological fluency (Mummery et al., 1996). By contrasting phonological fluency with semantic fluency the authors found left frontal activation in BA 44/6, whereas the reversed contrast revealed activations in the left temporal lobe (Mummery et al., 1996).
Another study that investigated the differences in regional cerebral blood flow using functional near-infrared spectroscopy (fNIRS) also found that semantic fluency was associated with greater activity in left temporal, whereas phonological fluency was associated with greater activity in left frontal cortex (Tupak et al., 2012).
However, this rather coarse classification of semantic and phonological fluency being more reliant on temporal and frontal brain areas, respectively, is not exclusively valid at present.
Recent studies investigating semantic and phonological fluency also found frontal involvement for semantic processing. A fMRI study using both letter and category cues found greater activity in the left inferior frontal gyrus (LIFG) when contrasting phonological fluency with semantic fluency, whereas for the reversed contrast greater activation was found in left middle frontal gyrus (Birn et al., 2010). Furthermore, the authors found greater activation in the left hemisphere for both semantic and phonological fluency when compared to automatic speech (Birn et al., 2010). Other authors have been concerned with the investigation of the involvement of the frontal lobe in semantic verbal fluency (Costafreda et al., 2006). More
Chapter 2 Theoretical Background 49 precisely, a functional segregation within the frontal lobe and especially of the left inferior frontal gyrus (LIFG) has been investigated by several authors (Costafreda et al., 2006; Heim et al., 2008; Katzev et al., 2013). In their meta-analysis Costafreda and colleagues (2006) suggested that the posterior-dorsal part of the LIFG (BA 44) is associated with phonological fluency whereas the anterior-ventral part of the LIFG (BA 45) is associated with semantic fluency. This functional dissociation could not be found in a systematic fMRI experiment by
Heim and colleagues (2008). These seemingly discordant findings could later be integrated by
Katzev and colleagues (2013) who found a functional segregation in LIFG when taking item difficulty and individual ability into account.
Although verbal fluency seems to be left-lateralized (Birn et al., 2010) there is also evidence that the right hemisphere plays a role to some extent (Perret, 1974; Tupak et al., 2012).
However, these results are still controversial and need further examination (Perani et al.,
2003; Chouiter et al., 2016; Li et al., 2017).
As has been mentioned in the previous section (Section 2.3.3), several functional neuroimaging studies have also investigated only one of the two verbal fluency variants (e.g.
Phelps et al., 1997; Herrmann et al., 2003; Hirshorn & Thompson-Schill, 2006; Allen et al.,
2008) and hence, differences in difficulty level between semantic and phonological fluency have not been considered in these studies. Moreover, the control tasks that have been used in these studies are very heterogeneous, making a direct comparison between studies even more difficult (Table 2.2). Just like the lesion studies, most studies have only reported associations between semantic and phonological fluency and activity in temporal and frontal brain areas, respectively. And quantitative analyses investigating the suggested double dissociation are very rare (Katzev et al., 2013).
With regard to the aspect of item difficulty it further needs to be noted that due to a general difference in task difficulty between semantic and phonological fluency, establishing the proposed double dissociation is considerably hampered by a potential resource artefact
50 Chapter 2 Theoretical Background
(Shallice, 1988; Davis, 2010). Although semantic fluency is generally reported to be easier than phonological fluency (see Section 2.5, Figure 2.6) none of the previous studies have explicitly controlled for this difference. Therefore, frontal lesions or frontal brain activation associated with performance in phonological fluency may only reflect higher task demands on general processes that are not specific for phonological fluency (Chapter 4; Shallice, 1988;
Davis, 2010).
In the second study of this thesis these limitations were considered and quantitative analyses were employed to confirm the suggested double dissociation between semantic and phonological fluency and temporal and frontal brain areas, respectively. To this end, data from a large sample of chronic left hemisphere stroke patients, who were assessed on measures of both semantic and phonological fluency, were investigated to study interaction effects in non-parametric lesion data (Chapter 4).
As verbal fluency is often impaired after stroke and this thesis includes a sample of stroke patients, a description of stroke and its impairments will be given in the following section.
After that a detailed description of a German version of the verbal fluency task developed by
Katzev and colleagues (2013), which is the primary neuropsychological measure used in the studies included in this thesis, will be given.
Chapter 2 Theoretical Background 51
Table 2.2. Overview of functional neuroimaging studies investigating verbal fluency
Study Author Year Imaging Verbal Fluency Task Control Task Activated Brain Areas Procedure
Semantic Fluency Phonological Fluency
1 Mummery 1996 PET Vegetables, Land S, B, M, P N.A. phon: BA 44/6 et al. Animals, Fruits, sem: LH temporal lobe, specifically Toys, Clothes, inferolateral and anteromedial cortex Tools, Weapons 2 Paulesu et 1997 fMRI Animals, Fruits, N.S. (30sec, covert) RS phon: posterior opercula portion of LIFG al. Kitchen Utensils sem: left retrosplenial region (30 sec, covert) 3 Phelps et al. 1997 fMRI N.A. generate 8 different Repeat left-sided activation. words and then go to Opposite LH SFG, LIFG, cingulate gyrus, DLPFC, the next letter in the (generating anterior cingulate alphabet (e.g. first R the antonym) followed by S) 4 Curtis et al. 1998 fMRI N.A. F, A, S (covert) internally SMA, LIFG, MFG, left insula articulate the word ‘rest’
5 Schlösser et 1998 fMRI N.A. F, A, S, T, N (covert) forward LH PFC al. counting 6 Gaillard et 2000 fMRI N.A. C, F, L (40sec, adults vs. posterior IFG, DLPFC, cingulate cortex al. covert) children 7 Gourovitch 2000 PET Sports, Animals, W, F, A, S, P, R alternate phon-sem: LIFG and inferior parietal cortex et al. Fruits, Jobs, between days Vegetables, Body of the week Parts and months of the year 8 Abrahams 2003 fMRI N.A. T, A, B, G, F naming of BA 46, 9, 44 and 45 et al. line drawings 9 Herrmann 2003 NIRS N.A. A, F, S RS LH and RH PFC et al. 10 Perani et al. 2003 fMRI Animals, Tools C, F, P (covert) RS phon: LH BA 44,45, 47, 9,6, 40 and RH BA
52 Chapter 2 Theoretical Background
(covert) 45,46 sem: LH BA 45, 18 11 Weiss et al. 2003 fMRI N.A. B, A, F, S (covert) RS LH and RH PFC (BA 47 and 46), cingulate (BA 32), RH cerebellum 12 Hirshorn & 2006 fMRI 1. free generation N.A. non-semantic greater LIFG activation during switching Thompson- (generate names baseline task compared to free recall and self-reported Schill of as many (covertly clustering category count exemplars) backwards by 2. switching two from a (generate category three-digit exemplars such number) that each item was from a different sub-category) 13 Allen et al. 2008 fMRI N.A. F, A, S, B, M, H, G, N.A. precentral/premotor almost exclusively in left L hemisphere language areas precentral/ premotor cortex, inferior frontal gyrus/frontal operculum, basal ganglia, thalamus, medial SMA 14 Heim et al. 2008 fMRI Birds, Mammals, B, F, K, M, SH, T syntactic phon: LH BA 45 and 44 Food, Weapons, (Phonemes) fluency task sem: LH BA 45 Tools, Toys (generate nouns with masculine gender), free generation 15 Schecklma 2008 fNIRS N.A. A, F, H, K, M, S reciting LH frontal activation nn et al. weekdays in consecutive manner 16 Hugdahl et 2009 fMRI 1. States in the N.A. imagine lying LH medial frontal gyrus al. USA, UK Soccer on the beach, clubs, male names looking at the sky (participants should keep a
Chapter 2 Theoretical Background 53
2. female names, constant countries in visual gestalt Africa, carnivore of this animals in memory in Scandinavia the working memory buffer) 17 Birn et al. 2010 fMRI N.S. N.S. months of the phon: LIFG, bilateral superior parietal cortex, year in and in the bilateral ventral occipitotemporal chronological cortex centered on the occipital temporal sulcus order starting sem: left fusiform gyrus and LH MFG from January 18 Tupak et al. 2012 fNIRS Animals, Fruits, A, F, S consecutively phon: frontal cortex Flowers repeat the sem: temporal cortex Clothes, Sports, G, P, E days of the Occupations week 19 Katzev et 2013 fMRI Vehicles, T, B, S, K, V, N, D, F N.A. phon: BA 44 al. Quadrupeds, sem: BA 45 Musical Instruments, Professions, Fluids, Toys, Furniture, Vegetables
Note. PET, positron emission tomography; fMRI, functional magnetic resonance imaging; TMS, transcranial magnetic stimulation; fNIRS functional near-infrared spectroscopy; N.A., not available; N.S., not specified; RS, resting state; BA, Broadman Area; sem, semantic verbal fluency; phon, phonological verbal fluency; LH, left hemisphere; RH, right hemisphere; SFG, superior frontal gyrus; LIFG, left inferior frontal gyrus; PFC, prefrontal cortex; DLPFC, dorsolateral prefrontal cortex; MFG, middle frontal gyrus; SMA, supplementary motor area. Note that the number of applied items for phonological and semantic fluency differ in some of the studies.
54 Chapter 2 Theoretical Background
2.4 Stroke
Stroke, which is known as the second leading cause of death and the third leading cause of disability worldwide, is defined as a global neurological impairment, with sudden onset of presumed vascular origin that persists for more than 24 hours (Petrea et al., 2009; Barker-
Collo et al., 2012; Johnson et al., 2016; World Health Organization, 2017). Stroke can occur in two different forms: in form of cerebral ischemia or cerebral hemorrhage. In the former, which is the most frequent form, the interruption of the blood supply of the brain is caused by a blockage of a blood vessel by a clot. In the latter, the interruption of blood supply is caused by the rupture of a blood vessel (Barker-Collo et al., 2012). The interruption of blood supply and of vital nutrients causes damage to the brain and if persistent leads to neuronal loss. A precursor but also risk factor, which is frequently confused with an ischemic insult, is the transient ischemic attack (TIA). As opposed to a stroke, neurological deficits and symptoms of a TIA are completely resolved within 24 hours. There are several other factors that have been associated with an elevated risk for stroke. Hypertension belongs to the highest risk factors, whereas other factors such as obesity, diabetes mellitus, cigarette smoking or physical inactivity have also been associated with an elevated likelihood of stroke (Goldstein et al.,
2011; Kim et al., 2016). The most common symptoms that can occur after stroke are motor dysfunction, such as clumsiness and loss of sensibility of fine finger movements or grip strength, but also complete paralysis of the side of the body contralateral to the hemisphere in which the stroke occurred (Barker-Collo et al., 2012; Xu et al., 2013). In addition to motor dysfunctions also language impairments leading to either speech production deficits or speech perception deficits or even both can emerge. These often differ in extent and severity from patient to patient, due to lesion location and size (Barker-Collo et al., 2012; Xu et al., 2013).
Other cognitive and neuropsychological impairments such as neglect, reduced processing speed, and executive dysfunctions as well as psychiatric impairments, such as depression have been frequently observed in stroke survivors (Barker-Collo et al., 2012; Xu, Ren, Prakash,
Chapter 2 Theoretical Background 55
Vijayadas, & Kumar, 2013; Fernandez et al., 2014). There are several factors that influence the occurrence of different deficits and their severity (Xu et al., 2013). These factors are the location and size of the lesion, hypoperfusion in the brain, functional deactivation of connected brain networks, as well as elevated pressure on the surrounding brain regions and tissues (Xu et al., 2013). For example stroke of the left hemisphere is most often associated with language and speech impairments whereas stroke of the right hemisphere most frequently leads to perceptual and visuospatial deficits (Barker-Collo et al., 2012; Lezak et al.,
2012).
These impairments have a direct impact on the quality of life and daily functioning (Patel et al., 2002). Although cognitive, motor, and language functions can improve in some cases during the first year after stroke onset with effective therapy, in other cases these impairments may persist. Some patients report impairments even decades after stroke onset (Barker-Collo et al., 2012). Therefore, adequate therapeutic interventions are necessary to improve quality of life in these patients.
2.5 German Version of the Verbal Fluency Task
The verbal fluency task used in the studies that were included in this thesis is a German version that was proposed by Katzev et al. (2013) employing a 2 × 2 factorial design of semantic cues (categories, e.g. vehicles) and phonological cues (letters, e.g. S) that were further classified as being of an easy or hard difficulty level based on extensive pilot testing
(cf. Katzev et al., 2013). For the pilot testing 16 preselected letters from the Mannheim
Corpus of German language, which comprises frequency estimations for a large set of existing words in spoken German, was used for the phonological cues (Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands; http://celex.mpi.nl; Katzev et al., 2013).
For the semantic cues, preselection included 8 categories with a presumably high task demand
(i.e. average retrieval of 8-10 category members) and 8 categories with a presumably low task
56 Chapter 2 Theoretical Background demand (i.e. average retrieval of 12-14 category members; cf. Katzev et al., 2013). During the pilot experiment, subjects were asked to generate as many nouns/category members within a time limit of one minute. For the proper experiment, four different items were presented for each cue type (semantic vs. phonological) and difficulty level (easy vs. hard), yielding a total of 16 items (8 categories and 8 letters). Based on the behavioral results of the pilot experiment cue types within each cell of the experimental design had an almost comparable level of empirical difficulty (Katzev et al., 2013). Furthermore, the items were chosen such that the items of hard difficulty in the semantic fluency and the easy items in the phonological fluency were of roughly equal difficulty. This is possible because the phonological fluency task is generally more difficult than the semantic verbal fluency task. The effects of cue type for the mean number of words produced can be found in Figure 2.6 (from Katzev et al., 2013).
The 16 items were presented in the same order for both healthy adults and stroke patients investigated in this thesis (Schmidt et al., 2017). Participants started with the semantic cues
(first, items for the easy condition: vehicles/means of transport [German: ‘Transportmittel’], quadrupeds [‘Vierbeiner’], musical instruments [‘Musikinstrumente’], professions [‘Berufe’]; second, items for the hard condition: fluids [‘Flüssigkeiten’], toys [‘Spielzeuge’], furniture
[‘Möbelstücke’], vegetables [‘Gemüsearten’]) followed by the phonological cues (first, items for the easy condition: T, B, S, K; second, items for the hard condition: V, N, D, F). For all study groups, which participated in the studies included in this thesis, the instructions for the verbal fluency task were given orally by the experimenter. They were told that the verbal fluency task would comprise two different parts (semantic and phonological) and that they were to generate as many nouns as possible within a time limit of one minute. The specific task rules were further explained with the aid of an exemplar category (e.g. food or groceries
[German: Lebensmittel]) and exemplar letter (e.g. E). For both conditions, participants were instructed that only nouns common in German should be said. No words should be produced twice and no proper names should be generated. Words beginning or ending with the same
Chapter 2 Theoretical Background 57 word stem were not allowed. For the phonological condition participants were told to generate words beginning with the given letter and that no verbs, adjectives, filler words, or numbers were allowed. Category and letter cues were displayed on a computer screen during the experiment and an acoustic cue indicated the beginning and end of the time interval in which participants were required to generate words. The total number of words for each cue was recorded and served as behavioral outcome measure for the studies reported in this thesis.
Since the psychometric properties of this German version of the verbal fluency task proposed by Katzev and colleagues (2013) have not been investigated in detail, the third study of this thesis aimed at investigating the psychometric properties as well as its test-retest reliability. In the following, findings from previous studies investigating the psychometric properties of common verbal fluency tasks as well as the rational of the third study (Chapter 5) will be further highlighted.
Figure 2.6. Effects of cue type (phonological vs. semantic) and task demand (high vs. low) (Katzev et al., 2013)
58 Chapter 2 Theoretical Background
2.6 Psychometric Properties of Verbal Fluency Tasks
The terms psychometric properties, quality criteria, or measurement properties are synonyms for the criteria required for the evaluation of the quality of an assessment or neuropsychological test (Cicchetti, 1994; Bortz & Döring, 2006). In this context, the classical test theory describes three main criteria. These are objectivity, reliability, and validity. As specific aspects of the second (i.e. reliability) and the third (i.e. validity) criterion were the main focus of the third study included in this thesis, these will be discussed in more detail below. For the sake of completeness, a brief explanation for the criterion ‘objectivity’ will be given first. Objectivity relates to the idea that a person’s performance on a cognitive task can be assessed independently of the experimenter (Bortz & Döring, 2006). Test results should not vary because of different testing conditions (e.g. different experimenter or different testing locations) (Bortz & Döring, 2006).
Reliability refers to the degree of accuracy with which a neuropsychological test assesses a particular cognitive function and hence, evaluates to which degree the measurement is free from measurement errors (Cicchetti, 1994; Bortz & Döring, 2006). Whereas objectivity is a necessary precondition of reliability, reliability is a necessary precondition of validity and there are several forms of reliability (Geisinger, 2013). These are inter-rater or inter-observer reliability, test-retest reliability, parallel-forms reliability, and the internal consistency reliability (Geisinger, 2013). As the test-retest reliability as well as the internal consistency reliability were the main focus of the third study included in this thesis they will be explained in greater detail in the following. First, test-retest reliability examines the degree of agreement between two (or more) assessments of a patient or participant, who was assessed with the same measurement instrument or neuropsychological test at two (or more) different measurement time points with a certain time interval between measurements (Cicchetti, 1994;
Bortz & Döring, 2006). Second, internal consistency reliability defines the extent to which items in a given test are correlated (Cicchetti, 1994). There are several different ways to
Chapter 2 Theoretical Background 59 measure internal consistency reliability with the most common known as Cronbach’s alpha
(Cronbach, 1951). Validity refers to the consistency of a cognitive task or empirical measurement with a logical measurement concept (Bortz & Döring, 2006). Therefore, validity refers to the degree to which a test claims to measure what it is supposed to measure
(Cronbach & Meehl, 1955). There are several forms of validity assessments (e.g. face, discriminant, concurrent and criterion validity etc.) of which those most relevant to this thesis will be discussed. These are concurrent validity and criterion validity. Concurrent validity refers to the extent to which a neuropsychological test correlates with an earlier neuropsychological test measuring the same or similar construct. Criterion validity refers to the relation of an individual’s standing on a current criterion (Cicchetti, 1994; Sireci & Sukin,
2013). If the measured criterion of interest is a group assignment, such as for example a comparison between patients and a control group, a quasi-experimental design can be used to assess a test’s concurrent validity (Sireci & Sukin, 2013). Hence, criterion-related concurrent validity is measured by testing for significant differences between groups (e.g. patients vs. controls).
After the theoretical introduction into the definition of the term ‘psychometric properties’ has been given, the findings of the current literature on the psychometric properties of the verbal fluency task will be presented in the following.
Despite the frequent use of the verbal fluency task also in clinical settings, only few studies have focused on its psychometric properties in general (e.g. Tombaugh, Kozak, & Rees, 1999;
Cohen & Stanczak, 2000; Bird et al., 2004) and their test-score reliability (Ruff et al., 1996;
Schmand et al., 2008) and the test-retest reliability (e.g. Lemay, Bedard, Rouleau, &
Tremblay, 2004; Ross et al., 2007) in particular. The validity of the verbal fluency task as a tool to assess both verbal ability (Cohen et al., 1999; Levelt et al., 1999) and executive ability
(Baldo & Shimamura, 1998; Miyake et al., 2000; Schwartz & Baldo, 2001) has been confirmed by several studies (Luo et al., 2010; Shao et al., 2014).
60 Chapter 2 Theoretical Background
However, with regard to the test-retest reliability of the verbal fluency task, extant findings are highly inconsistent, ranging from low (e.g. Bird et al., 2004) to adequate (e.g. Cohen &
Stanczak, 2000) to high indices of stability (e.g. Vlaar & Wade, 2003). These different findings might be due to the use of different versions and variations of the verbal fluency task with different types of items that have been reported in the literature. As has already been highlighted before different versions and variants with differing amounts of cues for the semantic and phonological fluency have been used in previous lesion and functional neuroimaging studies (Tables 2.1 and 2.2). That is, most commonly only one semantic (e.g. animals) and three phonological cues (e.g. F, A, S) were assessed (Benton, 1968; Loring et al., 1994; Herrmann et al., 2003; Schweizer et al., 2010). As none of the previous studies controlled for the differences in difficulty between semantic and phonological fluency (see
Section 2.3.3; Chapter 3), diminished measures of the test-retest reliability might be due to a restriction of range due to floor or ceiling effects. The third study of this thesis investigated the test-score reliability, internal consistency, criterion-related concurrent validity, and the test-retest reliability of a German version of the verbal fluency task with semantic and phonological cues, taking item difficulty into account (Chapter 5).
2.7 Present Studies and Aims of the Thesis
Verbal fluency is among the most widely used neuropsychological measures to assess language abilities and executive functions. Despite its frequent use in both healthy as well as clinical populations (Henry & Crawford, 2004; Costafreda et al., 2006; Wagner et al., 2014) there are still several inconsistencies and contradicting findings regarding the verbal fluency task and its underlying cognitive components as well as neural correlates. First, it is still debated whether semantic and phonological fluency rely on common or distinct sets of cognitive processes. Findings from behavioral, lesion, and functional neuroimaging studies suggest that semantic and phonological fluency are reliant on distinct brain structures and
Chapter 2 Theoretical Background 61 probably measure distinct cognitive processes, are challenged by factor analytic approaches suggesting that semantic and phonological fluency measure common cognitive processes
(Schmidt et al., 2017). The first study included in this thesis aimed at clarifying these seemingly inconsistent findings by using exploratory and confirmatory factor analysis with measures of semantic and phonological fluency only (Chapter 3; Schmidt et al., 2017).
Findings from lesion and functional neuroimaging studies investigating the neural correlates of verbal fluency have concordantly associated semantic fluency and phonological fluency with temporal and frontal brain structures, respectively. However, these associations are primarily based on qualitative analyses. Statistical methods, confirming the suggested double dissociation between semantic and phonological and temporal and frontal brain areas, respectively, have been lacking so far (Baldo et al., 2010). This is mainly because most studies investigated only one of the two verbal fluency variants in one and the same sample
(Milner, 1964; Mummery et al., 1996; Tucha et al., 1999). Furthermore, none of the previous studies controlled for differences in item difficulty between semantic and phonological fluency, hampering the chance for establishing dissociations (Chapter 4; Shallice, 1988).
By using quantitative analyses the second study of this thesis aimed at confirming this suggested double dissociation between semantic and phonological fluency and temporal and frontal brain areas, respectively (Chapter 4).
The third study is concerned with the psychometric properties of a German version of the verbal fluency task employed in the preceding studies. In addition to the assessment of its test-score reliability, indices for the internal consistency, its criterion-related concurrent validity, and its test-retest reliability will also be investigated (Chapter 5).
Now that the theoretical background has been established, in the following chapters of the individual studies a more detailed derivation of the specific research question addressed in each study will be provided.
62 Chapter 2 Theoretical Background
Research questions and overall aims of the thesis:
1. Do semantic and phonological fluency measure common or clearly distinct sets of cognitive processes?
2. Can the suggested double dissociation between semantic and phonological fluency and temporal and frontal brain areas be confirmed in statistical terms?
3. What are the psychometric properties of a German version of the verbal fluency task? Can we suggest a reliable task that is suitable for daily clinical practice, which includes both semantic and phonological items and further considers item difficulty?
Chapter 3 First Study – Factor analysis of the verbal fluency task 63
“I have no special talent. I am only passionately curious.“ Albert Einstein
Chapter 3
First Study – Factor analysis of the verbal fluency task
Both behavioral and lesion studies suggest that semantic and phonological fluency require distinct cognitive processes. However, factor analytic approaches challenge this assumption as has been outlined in Chapter 2. Therefore, the present study aimed at investigating whether an exploratory factor analysis reveals a coherent two-factor solution if explicitly tested in semantic and phonological verbal fluency only. As will be shown, a two-factor solution best explains the variance of the verbal fluency task in both healthy adults and stroke patients.1
1 The content of this chapter has been published in: Schmidt, C. S., Schumacher, L. V., Römer, P., Leonhart, R., Beume, L., Martin, M., Dressing, A., Weiller, C. & Kaller, C. P. (2017). Are semantic and phonological fluency based on the same or distinct sets of cognitive processes? Insights from factor analyses in healthy adults and stroke patients. Neuropsychologia , 99 , 148-155. The content of the article has not been changed, except for the numbering of sections, figures, and tables. © 2017 Elsevier Ltd. Reprinted with permission.
64 Chapter 3 First Study – Factor analysis of the verbal fluency task
3.1 Abstract
Verbal fluency for semantic categories and phonological letters is frequently applied to studies of language and executive functions. Despite its popularity, it is still debated whether measures of semantic and phonological fluency reflect the same or distinct sets of cognitive processes. Word generation in the two task variants is believed to involve different types of search processes. Findings from the lesion and neuroimaging literature further suggest a stronger reliance of phonological and semantic fluency on frontal and temporal brain areas, respectively. This evidence for differential cognitive and neural contributions is, however, strongly challenged by findings from factor analyses, which have consistently yielded only one explanatory factor.
As all previous factor-analytical approaches were based on very small item sets, this apparent discrepancy may be due to methodological limitations. In this study, we therefore applied a
German version of the verbal fluency task with 8 semantic (i.e. categories) and 8 phonological items (i.e. letters). An exploratory factor analysis with oblique rotation in N = 69 healthy young adults indeed revealed a two-factor solution with markedly different loadings for semantic and phonological items. This pattern was corroborated by a confirmatory factor analysis in a sample of N = 174 stroke patients. As results from both samples also revealed a substantial portion of common variance between the semantic and phonological factor, the present data further demonstrate that semantic and phonological verbal fluency are based on clearly distinct yet shared sets of cognitive processes.
Keywords: verbal fluency, cognitive processes, exploratory factor analysis, confirmatory factor analysis
Chapter 3 First Study – Factor analysis of the verbal fluency task 65
3.2 Introduction
Verbal fluency (e.g., Benton, 1968; Borkowski, Benton, & Spreen, 1967; Milner, 1964) is one of the most frequently used neuropsychological measures of language abilities and executive functioning (Chouiter et al., 2016; Lezak, Howieson, & Loring, 2004; Moscovitch, 1994;
Shao, Janse, Visser, & Meyer, 2014; Strauss, Sherman, & Spreen, 2006). This is particularly indicated by the vast and increasing number of more than 4100 publications listed in PubMed
(http://www.ncbi.nlm.nih.gov/pubmed; Fig. 3.1) that have assessed verbal fluency in a wide variety of clinical as well as healthy populations (for reviews, see Abwender, Swan,
Bowerman, & Connolly, 2001; Alvarez & Emory, 2006; Costafreda et al., 2006; Henry &
Crawford, 2004; Martin & Fedio, 1983; Metternich, Buschmann, Wagner, Schulze-Bonhage,
& Kriston, 2014; Sarkis et al., 2013).
1600
1200
800
400 Published Journal Articles in Medline (#) Published Journal Articles in Medline
0
before1977 19771982 1987 1992 1997 2002 2007 2012
¢ £ ¤ ¥ ¦ § ¨ 1981 1986 1991 1996 2001 2006 2011 2016
Five¡ Year Intervals Figure 3.1. Bibliometric overview of 4136 published journal articles on verbal fluency listed in PubMed (http://www.ncbi.nlm.nih.gov/pubmed; literature survey on December 31 st , 2016, search phrase: verbal fluency [Title/Abstract]) between 1965 and 2016 collapsed in five-year intervals.
66 Chapter 3 First Study – Factor analysis of the verbal fluency task
Verbal fluency is typically studied by requiring the subject to generate as many words as possible for a given category cue (semantic fluency) or letter cue (phonological fluency) within a pre-set time interval (e.g. 60s; Lezak et al., 2004; Strauss et al., 2006). In general, semantic fluency is usually easier than phonological fluency (Lezak et al., 2004) and both are assumed to differ in the type of search processes required for successful retrieval (Katzev,
Tuscher, Hennig, Weiller, & Kaller, 2013). That is, phonological fluency is believed to involve a serial search based on systematic syllabification of initial letters (Mummery et al.,
1996; Rende et al., 2002). By contrast, semantic fluency is most likely driven by association chains and spreading activations within cue-related subcategories (Gruenewald & Lockhead,
1980), thus requiring additional control processes such as generating and actively shifting between semantic sub-categories (Rosen & Engle, 1997; Troyer et al., 1997; Reverberi et al.,
2006), as well as selecting appropriate items from competing alternatives (Thompson-Schill et al., 1998).
In line with these proposed differences in cognitive processing during semantic and phonological fluency, the extant lesion and neuroimaging literature suggests a dissociation in the neural resources, with semantic fluency relying more on temporal brain areas and phonological fluency relying more on frontal brain areas. As such, patients with lesions in frontal regions reveal greater deficits in phonological fluency as compared to healthy controls or patients with non-frontal lesions, whereas patients with lesions in temporal regions show greater deficits in semantic verbal fluency (Baldo, Schwartz, Wilkins, & Dronkers, 2006,
2010; Borkowski, Benton, & Spreen, 1967; Jurado, Mataro, Verger, Bartumeus, & Junque,
2000; Szatkowska, Grabowska, & Szymanska, 2000; Thompson-Schill, et al., 1998; Troyer,
Moscovitch, Winocur, Alexander, & Stuss, 1998; see Henry & Crawford, 2004, for a meta- analytic review). Furthermore, greater task-related activation in frontal brain areas is associated with phonological fluency, whereas greater activation in temporal regions is
Chapter 3 First Study – Factor analysis of the verbal fluency task 67 associated with semantic fluency (e.g. Birn et al., 2010; Demonet et al., 1992; Gourovitch,
2000; Meinzer et al., 2009; Schlösser et al., 1998).
The potential dissociation between the cognitive processes involved and their underlying neural correlates associated with semantic and phonological verbal fluency is strongly challenged by findings from factor-analytic approaches: Several studies have suggested that semantic and phonological fluency primarily measure the same set of cognitive processes, given that inter-individual variation in performance in phonological and semantic fluency items consistently loads on a single factor (Ardila, Rosselli, & Bateman, 1994; Bizzozero et al., 2013; Unsworth, Spillers, & Brewer, 2011; Whiteside et al., 2016). Potential limitations of these previous factor analyses may lie in the very limited and partly disparate number of items used in assessing semantic and phonological fluency. For example, Whiteside and colleagues
(2016) as well as Bizzozero and colleagues (2013) used three phonological letters (F, A, S, and F, P, L, respectively) but only one semantic category (animals), while Unsworth and colleagues (2011) used two semantic (animal, supermarket) and two phonological letters (F,
S). Likewise, Ardila and colleagues (1994) used four phonological letters (F, A, S, M) but only two semantic categories (animals, fruits). In addition to the limited and disparate number of items, all previous studies compared measures of semantic and phonological verbal fluency in relation to other cognitive constructs, such as tests of executive function, language, working memory capacity, or processing speed (Ardila, Rosselli, & Bateman, 1994;
Bizzozero et al., 2013; Unsworth, Spillers, & Brewer, 2011; Whiteside et al., 2016). However, semantic and phonological verbal fluency can be expected to share common cognitive processes at least to some extent due to the general procedure of the fluency task, particularly in comparison to other cognitive functions. Thus, statistical models including a variety of other cognitive constructs lack a direct and unbiased comparison of verbal fluency sub-tasks and may not allow firm conclusions to be drawn about whether semantic and phonological fluency measure distinct or common cognitive processes.
68 Chapter 3 First Study – Factor analysis of the verbal fluency task
In this study, we addressed these potential limitations and investigated whether an exploratory factor analysis (EFA) reveals a two-factor rather than a one-factor solution (i) if explicitly tested in an analysis restricted to measures of phonological and semantic fluency and (ii) if this analysis is based on a larger and equal number of phonological and semantic items. To this end, we used a German version of the verbal fluency task with 16 items (8 categories and
8 letters; cf. Katzev et al., 2013). In a first exploratory step we analyzed verbal fluency data from a sample of healthy young adults (N = 69) in an EFA and demonstrated that semantic and phonological items indeed load on two separate factors, hence suggesting distinct sets of cognitive processes for semantic and phonological fluency. Furthermore, verbal fluency is often assessed in clinical populations with language and/or executive function deficits (Baldo et al., 2006, 2010; Birn et al., 2010; Henry & Crawford, 2004), so that analyses on the nature of verbal fluency processes is also highly relevant for neuropsychological studies. Thus, to probe the generalizability of results to a common clinical population, in a second confirmatory step we verified the results of the EFA using confirmatory factor analysis (CFA) in an independent sample of N = 174 stroke patients.
3.3 Methods
3.3.1 Participants
3.3.1.1 Healthy Subjects
For the exploratory factor analyses (EFA) in healthy young adults, N = 75 students were recruited from the University of Freiburg via information leaflets. All participants were right- handed and had normal or corrected-to-normal vision. Further inclusion criteria were age between 19 and 26 years, and German as a native language. Exclusion criteria were current or historical psychiatric or neurological illness, use of psychotropic medication, less than 8 years of education, and color blindness. Color blindness constituted an exclusion criterion, because the Tower of London-Freiburg version (TOL-F; Kaller et al., 2016) was also administered to
Chapter 3 First Study – Factor analysis of the verbal fluency task 69 each participant (cf. Köstering et al., 2015). Exclusion criteria were assessed using an in- house questionnaire on socio-demographic data. All students participated twice in the same measurements with a re-test interval of one week. Written informed consent was obtained before participation. The experiment was conducted in compliance with the Helsinki
Declaration of the World Medical Association (http://www.wma.net) and local ethical standards. Before participation, subjects were screened with respect to inclusion and exclusion criteria. One of the participants was excluded after the first session because of signs of depressive symptoms (score of 17) as measured with the Beck Depression Inventory-II (BDI-
II; Beck, Steer, & Brown, 1996).
Prior to the main analysis, individual data were inspected for outliers. Specifically, the difference between the total number of words produced at the first and second sessions were separately computed for the two fluency tasks (i.e. semantic and phonological fluency). Five participants were excluded from further analyses as their difference score for at least one of the two fluency tasks deviated for more than 1.5 times the interquartile range from the median difference score of the sample (see 3.3.2; Tukey, 1977). In consequence, the final sample comprised N = 69 healthy young adults (N = 47 female) with a mean age (±SD) of 23.07 ±
2.03 years (range, 19.04−26.48 years) and a mean education (±SD) of 16.10 ± 2.14 years
(range, 12 −23 years).
3.3.1.2 Stroke Patients
For the confirmatory factor analyses (CFA) in the clinical sample, N = 189 chronic stroke patients were recruited from the Department of Neurology at the University Medical Center
Freiburg and tested at least 5 months post-stroke as part of a larger study on the recovery after ischemic stroke. The main patient-specific inclusion criterion was first presentation of an ischemic stroke without a hemorrhagic event. Exclusion criteria at the acute stage were age over 90 years, inability to tolerate MRI examination or clinical testing due to poor general
70 Chapter 3 First Study – Factor analysis of the verbal fluency task health status, previous infarcts, previous intracerebral hemorrhage, previous traumatic brain injury, contemporary re-infarct, bilateral infarcts, major cognitive impairment (e.g., dementia), illiteracy, hearing and visual deficits, alcohol abuse, and contraindications for MRI examination such as claustrophobia or cardiac pacemaker. Every eligible participant was asked to participate and, once consented, tested at the Department of Neurology. The study was approved by local ethics authorities and conducted in compliance with the Helsinki
Declaration of the World Medical Association (http://www.wma.net).
Four patients were excluded from the present analyses because of severe aphasia (i.e. patients were unable to speak) which would confound the assessment of verbal fluency. Another 9 patients were excluded because they either did not complete, or were unable to perform, the task (i.e. task abortion at the request of the patient). Given the influence of education on performance in verbal fluency tasks (Straus, Sherman, & Spreen, 2006; Tombaugh et al.,
1999), another 2 patients were excluded due to an unusually low educational attainment of less than 8 years. Prior to the main analysis, individual data were inspected for outliers following the procedure described for healthy subjects. However, no further patients had to be excluded. In consequence, the final sample comprised N = 174 chronic stroke patients (N =
61 female) with a mean age (±SD) of 64.4 ± 13.7 years (range, 22.4−87.5 years), a mean education (±SD) of 13.2 ± 3.4 years (range, 8−23), and an average post-stroke duration (±SD) of 18.3 ± 19.1 months (range, 5−73.5 months). A total of N = 105 patients with left hemisphere and N = 69 patients with right-hemisphere strokes were included. The stroke territory concerned in most of these cases (n = 147) was that of the middle cerebral artery.
3.3.2 Verbal Fluency Task
Participants were administered a German version of the verbal fluency task employing a 2×2 factorial combination of semantic cues (categories, e.g. vegetables) and phonological cues
(letters, e.g. V) that were further classified as being of an easy or hard difficulty level. This
Chapter 3 First Study – Factor analysis of the verbal fluency task 71 classification was based on pilot testing in a preceding study (cf. Katzev et al., 2013) and ensured that, despite differences in the general difficulty between phonological and semantic items, letter and category cues within each cell of the resulting factorial design had an almost comparable level of empirical difficulty (Supplementary Table S3.1). Four different items were presented for each cue type (semantic vs. phonological) and difficulty level (easy vs. hard), yielding a total of 16 items (8 categories and 8 letters). Items and presentation order were identical for all healthy participants and stroke patients and did also not differ between the two testing sessions in the healthy sample. Both participants and patients started with the semantic condition (first, items for the easy condition: vehicles/means of transport [German:
‘Transportmittel’], quadrupeds [‘Vierbeiner’], musical instruments [‘Musikinstrumente’], professions [‘Berufe’]; second, items for the hard condition: fluids [‘Flüssigkeiten’], toys
[‘Spielzeuge’], furniture [‘Möbelstücke’], vegetables [‘Gemüsearten’]) followed by the phonological condition (first, items for the easy condition: T, B, S, K; second, items for the hard condition: V, N, D, F).
Instructions for the verbal fluency task were given orally by the experimenter (CS, LVS, PR).
Participants were told that the verbal fluency task would comprise two different parts
(semantic and phonological fluency) and that they were to generate as many nouns as possible within a time limit of 60s following either a category or a letter. Task rules were explained with the help of example items. In the semantic condition, the example item was
‘Lebensmittel’ (English: food or groceries). First of all, participants were told that only words common in German should be said (e.g., milk, butter, bread, cheese). Second, no words should be produced twice and participants were not allowed to say proper names or brand names (e.g., Pepsi). Finally, no words beginning or ending with the same word stem were valid (e.g., milk , milk powder). For the phonological condition, additional rules were explained with the help of the example letter ‘E’. Participants were told that words generated in the second part should begin with the given letter and only nouns should be said (e.g., egg,
72 Chapter 3 First Study – Factor analysis of the verbal fluency task eye, elephant). That is, besides proper names and brand names, verbs, adjectives, filler words, or numbers were also not allowed for these trials.
During trials either the corresponding category or letter was displayed on a computer screen and acoustic cues indicated the beginning and end of the 60 s response interval. The total number of correct words for each item was recorded and served as outcome measures for data analyses.
3.3.3 Data Analyses
Analyses were conducted using SPSS Statistics 23 and AMOS 23 (IBM Corp., Armonk, NY,
USA). To first investigate whether semantic and phonological fluency indeed measured two distinct sets of cognitive processes, an exploratory factor analysis (EFA) was carried out on the data of the healthy sample. In particular, semantic and phonological items entered a principal component analysis. As semantic and phonological fluency can be expected to share at least some cognitive processes such as language, working memory, and attention, it was hypothesized that the extracted factors may be correlated and, accordingly, an oblique rotation
(Promax, Hendrickson & White, 1964) was used instead of an orthogonal rotation (Field,
2005). The number of factors was determined using Horn’s parallel analysis and Velicer’s minimum average partial (MAP) test (Horn, 1965; O’Connor, 2000; Velicer, 1976). In Horn’s parallel analysis, factor extraction is based on the comparison of raw data eigenvalues with random data eigenvalues. Raw data eigenvalues that are larger than the random data eigenvalues are retained as factors (cf. Horn, 1965). Factor extraction of the MAP test is based on the matrix of partial correlations with a rule to stop factor extraction (cf. Velicer,
1976). To validate the factor structure derived from the EFA in the healthy sample, a confirmatory factor analysis (CFA) was then performed in the sample of stroke patients using
AMOS Version 23.
Chapter 3 First Study – Factor analysis of the verbal fluency task 73
3.4 Results
3.4.1 Exploratory Factor Analysis (EFA) in Healthy Participants
A principal component analysis was conducted on the verbal fluency data of healthy young adults (cf. 3.3.1.1) using Promax rotation. Before the analysis, the Kaiser-Meyer-Olkin
(KMO) measure of sampling adequacy and Bartlett’s test of sphericity were inspected. Both the KMO measure of .868 and a significant Bartlett’s test of sphericity ( χ²(120) = 539.954, p
<.001) indicated that the implementation of the analysis was appropriate. Velicer’s minimum partial (MAP) test as well as Horn’s parallel test suggested the extraction of two factors. Note that the same decision would have been obtained by the default criterion implemented in
SPSS (i.e. eigenvalues > 1) as well as by the inspection of the scree plot (Supplementary Fig.
S3.1).
ABSemantic items Phonological items Factor 1 Factor 1 1 Factor 2 1 Factor 2
0.8 0.8
0.6 0.6
0.4 0.4 Factor Loading Factor Loading
0.2 0.2
0 0
vehiclesquadrupedsmusicalprofessions instruments vegetables fluids toys furniture T B S K V N D F
Figure 3.2. Factor solutions of the exploratory factor analysis in healthy young adults illustrating the pattern matrix (A) for the semantic items and (B) for the phonological items.
As is illustrated in Figure 3.2, the results of the EFA revealed that the first factor had strong loadings from the items T, B, S, K, V, N, D, and F and accounted for 42.5% of variance in the verbal fluency data. The second factor had strong loadings from vehicles, quadrupeds,
74 Chapter 3 First Study – Factor analysis of the verbal fluency task musical instruments, professions, fluids, toys, furniture, and vegetables and accounted for further 13.5% of incremental variance. Both factor 1 and 2 were correlated by r = .508, thus indicating a common share of explained variance of about 25.8%. Depicting the different two factors’ unique and common shares of explained variance in a Venn diagram (Fig. 3.3) demonstrates that 56.1% of variation in inter-individual differences in verbal fluency could be accounted for by two partly correlated factors, presumably reflecting unique as well as overlapping cognitive processes for semantic and phonological fluency.
semantic
common variance
total phonological
Figure 3.3. Venn diagram depicting the two factors’ unique and common shares of explained variance for the variation in inter-individual differences in the verbal fluency task. The white circle, labeled with ‘total’, depicts the total variance (100%) of the data for the verbal fluency task in healthy young adults. The two light grey circles, labeled with ‘phonological’ and ‘semantic’, depict their respective proportions of explained variance (phonological [factor 1] = 42.5%; semantic [factor 2] = 28.1%). The dark grey area labeled common variance depicts the explained common variance of both factor 1 and factor 2 (14.5%). Therefore, the unique variance explained by factor 1 (phonological) is 42.5%-14.5% = 28.0% and by factor 2 (semantic) is 28.1%-14.5% = 13.6%. Total explained variance by the two factors is 14.5%+28%+13.6% = 56.1%.
3.4.2 Confirmatory Factor Analysis (CFA) in Stroke Patients
Based on the results of the EFA, a confirmatory factor analysis (CFA) was used to validate these findings in an independent sample of N = 174 stroke patients (cf. 3.3.1.2). More
Chapter 3 First Study – Factor analysis of the verbal fluency task 75 specifically, we hypothesized that items for semantic and phonological fluency load on two different (but partly correlated) latent factors and that, consequently, a CFA model including such a two-factor solution fits the data significantly better than a model with a one-factor solution.
Table 3.1. Results of the confirmatory factor analysis (CFA) for the clinical sample of 174 stroke patients
χ² DF p- χ²/DF TLI CFI GFI AGFI RMSEA value rho2
Model 1 93.565 103 .736 .908 1.005 1.000 .936 .916 <.001
Model 2 225.137 104 <.001 2.165 .938 .946 .809 .750 .082
Note: DF = degrees of freedom; TLI = Tucker-Lewis index; CFI = comparative fit index; GFI = goodness- of-fit index; AGFFI = adjusted goodness-of-fit index; RMSEA = root mean square error of approximation.
In detail, model 1 comprised a two-factor model with the 16 items assigned either to the phonological or to the semantic factor (see Supplementary Fig. S3.2). All semantic and phonological items significantly loaded on their corresponding factor, with loadings ranging from .659 to .908. The correlation between the two factors was r = .875, indicating about
76.6% of shared variance. The indices for goodness of fit (GFI) and for adjusted goodness of fit (AGFI) were well above the acceptable level of .85 (Table 3.1; Schermelleh-Engel,
Moosbrugger, & Müller, 2003). Other model fit indices such as χ2 and χ2/df, the comparative fit index (CFI), and the root-mean-square error of approximation (RMSEA) also suggested a very good fit of the data to the model (Table 3.1).
In order to test whether the two-factor model 1 was superior to a one-factor solution, in model
2 the correlation between factor 1 and factor 2 was set to 1 to constrain the two factors to be exactly the same. For model 2, the GFI and the AGFI were markedly lower as compared to model 1 (Table 3.1). Furthermore, all other model fit indices (i.e., χ2, χ2/df, CFI, and RMSEA)
76 Chapter 3 First Study – Factor analysis of the verbal fluency task suggested that the null hypothesis of a good model fit be rejected. The comparison between both models further indicated that the two-factor solution in model 1 was significantly better than the one-factor solution in model 2 by a change in χ2(df = 1) of 131.571 (p < .001). Taken together, the CFA hence confirmed that, despite their overlapping variance, semantic and phonological verbal fluency also measure distinct sets of cognitive processes.
Finally, for direct comparability, we repeated the EFA approach in the stroke sample. Results were essentially the same as for the healthy young adults, yielding a two-factor solution with all semantic items loading on one factor and all phonological items loading on the other (see
Supplementary Fig. S3.3).
As the stroke sample comprised several patients with aphasia who possibly performed worse in both the semantic and phonological verbal fluency due to a general language impairment, we further repeated the above CFA in a subsample of n = 134 stroke patients without signs of aphasia in the acute phase, so as to rule out any potential bias. The results of this analysis were highly comparable to those obtained in the whole sample (see Supplements).
To further understand the increased correlation between the semantic and phonological factor in the stroke patients (r = .875) as compared to the healthy controls (r = .508), partial correlation was computed. Specifically, available data on the patients’ Montreal Cognitive
Assessment score (MoCA; Nasreddine et al., 2005) was used as control variable, as this test reflects a measure of the general cognitive status. When controlling for the MoCA, a considerably reduced correlation between the two factors (r = .575) was revealed, thus closely matching the findings in the healthy controls. However, it should be noted that the MoCA also includes verbal fluency as a sub-test. To additionally control for this issue, partial correlation was re-computed using a MoCA score excluding the verbal fluency sub-test. Again, a substantially reduced correlation between the two factors (r = .616) was observed. The same results were obtained when repeated in the sample of patients without signs of aphasia (r =
.553). Taken together, a large proportion of the increased common variance of the two factors
Chapter 3 First Study – Factor analysis of the verbal fluency task 77 in stroke patients can be explained by their general cognitive status. The unique and common shares of explained variance of the two factors are depicted in a Venn diagram (Fig. 3.4).
common variance of AB phonological & semantic semantic common variance of phonological, semantic & MoCA semantic
common variance
total
phonological MoCA phonological
Figure 3.4. Venn diagram depicting the two factors’ unique and common shares of explained variance in stroke patients for the variation in inter-individual differences in (A) the verbal fluency task and (B) the MoCA. (A) The white circle, labeled with ‘total’, depicts the total variance (100%) of the data for the verbal fluency task in stroke patients. The two light grey circles, labeled with ‘phonological’ and ‘semantic’, depict their respective proportions of explained variance of the verbal fluency data (phonological [factor 1] = 63.6%; semantic [factor 2] = 47.3%). The dark grey area, labeled ‘common variance’, depicts the explained common variance of both factor 1 and factor 2 (40.9%). Therefore, the unique variance explained by factor 1 (phonological) is 63.6%-40.9% = 22.6% and by factor 2 (semantic) is 47.3%-40.9% = 6.4%. Total explained variance by the two factors is 40.9%+22.6%+6.4% = 69.9%. (B) The white circle, labeled with ‘MoCA’, depicts the total variance (100%) in MoCA scores of stroke patients. The dark grey area, labeled with ‘common variance of phonological & semantic’, depicts the shared variance of both factor 1 and factor 2 (58.5%). The lighter grey area, labeled with ‘common variance of ‘phonological, semantic & MoCA’, depicts the share of common variance of factor 1 and factor 2 and MoCA scores (39%). The unique variance of MoCA scores explained by factor 1 (phonological) is 40.5%-39% = 1.5% and by factor 2 (semantic) is 53.4%-39% = 14.4%. Note that the amount of variance (i.e. the size of each circle) of MoCA scores, factor 1 (phonological), and factor 2 (semantic) was set to be equal for visualization purposes.
78 Chapter 3 First Study – Factor analysis of the verbal fluency task
3.5 Discussion
Results of the exploratory factor analysis (EFA) in the healthy sample and the subsequent confirmatory factor analysis (CFA) in the stroke sample both indicated that a two-factor solution, in which all semantic items load on one factor and all phonological items load on the other, best described the variation of inter-individual differences in both verbal fluency types and that this solution was significantly superior to a one-factor solution. The present results hence constitute first evidence from factor-analytic approaches that semantic and phonological verbal fluency indeed measure two distinct sets of cognitive processes. In this respect, the present data close an existing gap in the literature by resolving the discrepancy between the results from previous factor analyses favoring a one-factor solution (Ardila,
Rosselli, & Bateman, 1994, Unsworth, Spillers, & Brewer, 2011; Whiteside et al., 2016) and the evidence from the lesion and neuroimaging literature suggesting a two-factor solution (see e.g., Birn et al., 2010; Gourovitch, 2000; Henry & Crawford, 2004; Indefrey & Levelt, 2004;
Wagner et al., 2014; for overviews).
At first glance, the present results contradict those found in previous factor analyses.
However, the preference of a one-factor solution in these studies (Ardila, Rosselli, &
Bateman, 1994; Bizzozero et al., 2013; Unsworth, Spillers, & Brewer, 2011; Whiteside et al.,
2016) might be due to several methodological reasons (see Introduction; i.e. limited and partly disparate number of items for semantic and phonological verbal fluency; potential bias by inclusion of other cognitive constructs in the analyses). Furthermore, as the present results emphasize both common and distinct shares of variance in semantic and phonological fluency, they rather seem to extend the findings from these previous studies by complementing prior evidence for common cognitive processes with additional evidence for distinct sets of cognitive processes.
Chapter 3 First Study – Factor analysis of the verbal fluency task 79
However, the present findings cannot answer the question regarding the nature of cognitive processes that are required for both fluency sub-tasks and the type of processes that are exclusively involved in either semantic or phonological fluency.
With regard to common cognitive processes, working memory, self-monitoring, cognitive flexibility as well as sustained attention have been frequently associated with both types of verbal fluency (Baldo et al., 2006; Rosen & Engle, 1997; Robinson et al., 2012; Schwartz et al., 2003; Troyer et al. 1998). For instance, executive control processes for constant monitoring of words already said and semantic sub-categories or letters already processed as well as switching between sub-categories/letters is required for successful retrieval in both semantic and phonological fluency (Henry & Crawford, 2004).
In addition to these shared cognitive processes, different or specific processes are required for one fluency task but not the other. These may be related to the different search strategies required for semantic and phonological fluency (Unsworth et al., 2011). Whereas semantic fluency is based on search through conceptual or semantic memory and hence is dependent on an intact integrity of semantic memory, phonological fluency is based on search through lexical or phonological memory that is dependent on proper syllabification (Henry &
Crawford, 2004; Mummery et al., 1996; Rende et al., 2002). Moreover, semantic verbal fluency is based on a serial search process in which, for the initially specified category, a first step is to search for sub-categories and then for members of these particular sub-categories. In phonological fluency successful retrieval is based on systematic syllabification of initial letters and hence, automatic retrieval influenced by semantic association chains needs to be suppressed (Katzev et al., 2013). Since knowledge is organized in semantic networks, which further promotes the retrieval strategy by semantic (sub-)categories, semantic fluency is commonly easier than phonological fluency. In close relation, studies using a dual-task methodology also found evidence for dissociable cognitive processes underlying semantic and phonological fluency. For instance, Moscovitch (1994) demonstrated that only letter fluency
80 Chapter 3 First Study – Factor analysis of the verbal fluency task was impaired when participants engaged in a concurrent finger-tapping task that presumably utilizes the frontal lobes, thus further arguing for differences in the neural correlates of the two types of fluency (also see below). Rende, Ramsberger, and Miyake (2002) used a dual- task paradigm to test which sub-components of working memory (e.g. phonological loop, visuospatial sketchpad, and central executive) are differentially involved in semantic and phonological fluency, respectively. They found that a concurrent task that primarily involved the phonological loop (e.g. articulatory suppression) resulted in decreased performance only in the phonological fluency task, whereas a concurrent task primarily involving the visuospatial sketchpad (e.g. cube comparison) resulted in decreased performance only in the semantic fluency task (Rende et al., 2002). In addition, a concurrent task that involved frequent switching between mental sets (e.g. arithmetic switching) and thus presumably engaged the central executive equally decreased both semantic and phonological fluency
(Rende et al., 2002). These findings are in line with the proposed differences in search strategy for the two types of verbal fluency in that they demonstrate dissociable recruitment of working memory processes by semantic versus phonological fluency and reverberate the shared recruitment of executive control processes presumably needed for both types of verbal fluency. On a statistical level, this is reflected by the results of this study of a two-factor solution with partial correlation between the two factors.
Considering the neural basis of semantic and phonological verbal fluency, several functional neuroimaging as well as lesion and behavioral studies showed that semantic and phonological verbal fluency can be attributed to the temporal and frontal lobes, respectively (Baldo et al.,
2006, 2010; Benton, 1968; Costafreda et al., 2006; Gourovitch et al., 2000; Henry &
Crawford, 2004, Thompson-Schill et al., 1998). However, other regions such as parietal cortex, insula, putamen, and cerebellum have also been implicated in both types of verbal fluency (Baldo et al., 2006; Indefrey & Levelt, 2004; Schweizer, Gillingham, Cusimano &
Stuss, 2010). Thompson-Schill and colleagues (1997) suggested that the left inferior frontal
Chapter 3 First Study – Factor analysis of the verbal fluency task 81 gyrus (LIFG) is critical for the selection processes, whereas Robinson and colleagues (2012) suggested that fluency tasks with greater selection demands from multiple competing responses will be impaired following LIFG damage. Katzev and colleagues (2013) tested this assumption and demonstrated that differences in activation of sub-regions within the LIFG can be attributed to differences in task demand and individual ability (Katzev et al., 2013;
Robinson, Shallice, Bozzali, & Cipolotti, 2012; Thompson-Schill et al., 1997; see Costafreda et al., 2006 for a review on the role of LIFG in verbal fluency; see also Wagner et al., 2014), further indicating that distinct sets of cognitive processes are required for the two types of verbal fluency tasks. Also studies investigating qualitative features of the verbal fluency task such as clustering and switching (Troyer et al., 1997) found differential involvement of areas within the frontal lobe (Reverberi, Laiacona, & Capitani, 2006; Schwartz & Baldo, 2001), corroborating the current study and previous findings.
With regard to the present findings in the clinical group, the higher correlation between the two factors for stroke patients as compared to the healthy participants suggests that the patients probably exhibited deficits in those cognitive processes that are required for both semantic and phonological fluency such as working memory and attention (Mok et al., 2004;
Patel, Coshall, Rudd, & Wolfe, 2002; Tatemichi, et al., 1994). This is supported by the considerably reduced correlation between the two factors when controlling for the MoCA score (see 3.4.2), which measures a patient’s general cognitive status. However, the MoCA score also represents inter-individual differences that are due to differences in age or level of education but independent of the effects of the stroke. Although the MoCA score is partially controlled for these variables, both specific and unspecific cognitive impairments are quantified with this test. In close relation, the patient sample was considerably older as compared to the group of healthy controls, which may imply that age-related decrements in processing speed, executive function, working memory, and access to lexico-semantic operations, irrespective of stroke severity, were likely to be present in this group and may
82 Chapter 3 First Study – Factor analysis of the verbal fluency task have augmented the shared variance between the two factors (Baciu et al., 2016; Rosen &
Engle, 1997; Salthouse, 2009). In this respect, a negative correlation between age and performance in the verbal fluency task in the stroke sample (r = -.342) indicated that younger patients performed better than older patients, further corroborating these assumptions.
There are several limitations of the present study. Firstly, the two samples differed with respect to various socio-demographic variables. Specifically, the healthy sample consisted of young and mainly female university students, whereas the stroke sample comprised older and mainly male patients with a broad variation in educational attainment. These differences might be due to the fact that, at least in our experience, women are more likely to voluntarily participate as healthy control subjects in research studies. By contrast, due to age- and sex- related differences in the incidence of stroke (Appelros et al., 2009; Petrea et al., 2009;
Reeves et al., 2008; Roquer, Campello, & Gomis, 2003; Wyller, 1999), stroke patients at an average age of 64 years, as recruited in our sample, are more likely to be male. Furthermore, the difference between the two groups concerning education can be attributed to the recruitment of healthy young adults from among university students, whereas no such restrictions applied to the stroke sample. As age, education, and gender have been reported to influence performance in semantic and phonological fluency (e.g., Loonstra et al., 2001;
Straus, Sherman, & Spreen, 2006; Tombaugh et al., 1999; van der Elst et al., 2006), it can be assumed that the two samples also differed with respect to their overall verbal fluency ability and probably also with respect to general language abilities. However, as the results from the exploratory factor analysis in the young healthy adults and the confirmatory factor analysis in the older stroke sample converged on the same conclusion, the different characteristics of the two samples may in turn be taken as an indication for the generalizability of the present findings.
A second limitation concerns the procedure of the verbal fluency task, which was in part different from other studies. Letter and category cues are commonly given orally but were
Chapter 3 First Study – Factor analysis of the verbal fluency task 83 visually presented in the present study. This difference is, however, unlikely to have biased the results, as all participants were able to read. Another difference from commonly applied procedures is that only nouns were allowed for responding to the phonological fluency condition. This task rule was adopted from Katzev and colleagues (2013) who used frequency estimations for German nouns (exclusive of compound nouns) with given initial letters derived from the Mannheim Corpus of the German language as an indication of the expected difficulty level of individual letter cues. In consequence, instructions for the phonological condition restricted correct answers to nouns only, and as the present study used the same items as Katzev et al. (2013), this restriction was also applied here (see also Heim et al., 2008,
2009). This difference in the task procedures may have resulted in increased item difficulty and increased task demands for phonological cues compared to the common assessment of phonological fluency. However, given that previous evidence from the lesion, neuroimaging, and cognitive literature already suggested a potential dissociation between semantic and phonological verbal fluency in the underlying neural and cognitive processes (Baldo et al.,
2006, 2010; Benton, 1968; Costafreda et al., 2006; Gourovitch et al., 2000; Henry &
Crawford, 2004, Thompson-Schill et al., 1998), it seems unlikely that the present findings of a two-factor solution are artificially introduced by restricting the phonological fluency to nouns only. Yet, future studies should explicitly compare the impact of different types of instructions for phonological fluency on the number of words generated and on the factor- analytic pattern of relationships between semantic and phonological fluency.
3.6 Conclusion
Taken together, these findings considerably expand previous studies that investigate whether semantic and phonological verbal fluency measure the same or distinct sets of cognitive processes by providing explanations for inconsistent findings between the extant neuroimaging and lesion literature and results from factor-analytic approaches. Both EFA and
84 Chapter 3 First Study – Factor analysis of the verbal fluency task
CFA with a large number of items revealed that semantic and phonological verbal fluency measure both clearly distinct, yet shared sets of cognitive processes. Further studies may aim at investigating the nature of these distinct and common cognitive processes and whether the observed sharing of unique and common variance of the verbal fluency task for the two factors can be attributed to circumscribed neural networks. Although specific cognitive sub- functions involved thus remain to be fully characterized, it is now clear that semantic and phonological fluency are not sufficiently described by assuming one cognitive process.
3.7 Acknowledgement
The authors report no conflicts of interest. This research was supported by a grant from the
BrainLinks-BrainTools Cluster of Excellence (project #36 to C.W., C.P.K, and M.M.) funded by the German Research Foundation (DFG; grant # EXC 1086). CSMS and LVS received scholarship funds from the State Law on Graduate Funding of the University of Freiburg,
Germany. The authors thank Dr. Ruoyi Sun for her valuable comments on the manuscript.
Chapter 3 First Study – Factor analysis of the verbal fluency task 85
3.8 References
Abwender, D. A., Swan, J. G., Bowerman, J. T., & Connolly, S. W. (2001). Qualitative analysis of verbal fluency output: review and comparison of several scoring methods. Assessment , 8(3), 323-338. Alvarez, J. A., & Emory, E. (2006). Executive function and the frontal lobes: a meta-analytic review. Neuropsychology Review , 16(1), 17-42. Appelros, P., Stegmayr, B., & Terént, A. (2009). Sex Differences in Stroke Epidemiology. A Systematic Review, 40 (4), 1082-1090. Ardila, A., Rosselli, M., & Bateman, J. R. (1994). Factorial structure of cognitive activity using a neuropsychological test battery. Behav Neurol, 7 (2), 49-58. Baldo, J. V., Schwartz, S., Wilkins, D., & Dronkers, N. F. (2006). Role of frontal versus temporal cortex in verbal fluency as revealed by voxel-based lesion symptom mapping. Journal of the International Neuropsychological Society, 12 , 896-900. Baldo, J. V., Schwartz, S., Wilkins, D. P., & Dronkers, N. F. (2010). Double dissociation of letter and category fluency following left frontal and temporal lobe lesions. Aphasiology , 24(12), 1593-1604. Benton, A. L. (1968). Differential behavioral effects in frontal lobe disease. Neuropsychologia , 6(1), 53-60. Birn, R. M., Kenworthy, L., Case, L., Caravella, R., Jones, T. B., Bandettini, P. A., & Martin, A. (2010). Neural systems supporting lexical search guided by letter and semantic category cues: a self-paced overt response fMRI study of verbal fluency. Neuroimage, 49 (1), 1099-1107. Bizzozero, I., Scotti, S., Clerici, F., Pomati, S., Laiacona, M., & Capitani, E. (2013). On which abilities are category fluency and letter fluency grounded? A confirmatory factor analysis of 53 Alzheimer's dementia patients. Dement Geriatr Cogn Dis Extra, 3(1), 179-191. Borkowski, J. G., Benton, A. L., & Spreen, O. (1967). Word fluency and brain damage. Neuropsychologia , 5(2), 135-140. Chouiter, L., Holmberg, J., Manuel, A. L., Colombo, F., Clarke, S., Annoni, J. M., & Spierer, L. (2016). Partly segregated cortico-subcortical pathways support phonologic and semantic verbal fluency: A lesion study. Neuroscience, 329 , 275-283. Costafreda, S. G., Fu, C. H., Lee, L., Everitt, B., Brammer, M. J., & David, A. S. (2006). A systematic review and quantitative appraisal of fMRI studies of verbal fluency: role of the left inferior frontal gyrus. Human Brain Mapping , 27(10), 799-810. Demonet, J. F., Chollet, F., Ramsay, S., Cardebat, D., Nespoulous, J. L., Wise, R., . . . Frackowiak, R. (1992). The anatomy of phonological and semantic processing in normal subjects. Brain, 115 ( Pt 6) , 1753-1768. Field, A. (2005). Discovering Statistics using SPSS (2nd ed.): SAGE Publications Inc. Gourovitch, M. L., Kirkby, B. S., Goldberg, T. E., Weinberger, D. R., Gold, J. M., Esposito, G., . . . Berman, K. F. (2000). A comparison of rCBF patterns during letter and semantic fluency. Neuropsychology, 14 (3), 353-360. Gruenewald, P. J., & Lockhead, G. R. (1980). The free recall of category examples. Journal of Experimental Psychology: Human Learning and Memory , 6(3), 225-240.
86 Chapter 3 First Study – Factor analysis of the verbal fluency task
Heim, S., Eickhoff, S. B., & Amunts, K. (2008). Specialisation in Broca's region for semantic, phonological, and syntactic fluency?. Neuroimage , 40 (3), 1362-1368. Heim, S., Eickhoff, S. B., & Amunts, K. (2009). Different roles of cytoarchitectonic BA 44 and BA 45 in phonological and semantic verbal fluency as revealed by dynamic causal modelling. Neuroimage , 48 (3), 616-624. Hendrickson, A. E., and P. O. White. 1964. Promax: a quick method for rotation to oblique simple structure. British Journal of Statistical Psychology , 17, 65-70. Henry, J. D., & Crawford, J. R. (2004). A meta-analytic review of verbal fluency performance following focal cortical lesions. Neuropsychology , 18(2), 284-295. Indefrey, P., & Levelt, W. J. (2004). The spatial and temporal signatures of word production components. Cognition, 92 , 101-144. Jurado, M. A., Mataro, M., Verger, K., Bartumeus, F., & Junque, C. (2000). Phonemic and semantic fluencies in traumatic brain injury patients with focal frontal lesions. Brain Inj, 14 , 789-795. Katzev, M., Tuscher, O., Hennig, J., Weiller, C., & Kaller, C. P. (2013). Revisiting the functional specialization of left inferior frontal gyrus in phonological and semantic fluency: the crucial role of task demands and individual ability. Journal of Neuroscience , 33(18), 7837-7845. Köstering, L., Nitschke, K., Schumacher, F. K., Weiller, C., & Kaller, C. P. (2015). Test- retest reliability of the Tower of London Planning Task (TOL-F). Psychol Assess, 27 (3), 925-931. Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsychological Assessment . New York: Oxford University Press. Loonstra, A. S., Tarlow, A. R., & Sellers, A. H. (2001). COWAT metanorms across age, education, and gender. Applied Neuropsychology , 8(3), 161-166. Martin, A., & Fedio, P. (1983). Word production and comprehension in Alzheimer's diseáse: The breakdown of semantic knowledge. Brain and Language , 19(1), 124-141. Meinzer, M., Flaisch, T., Wilser, L., Eulitz, C., Rockstroh, B., Conway, T., . . . Crosson, B. (2009). Neural signatures of semantic and phonemic fluency in young and old adults. J Cogn Neurosci, 21 (10), 2007-2018. Metternich, B., Buschmann, F., Wagner, K., Schulze-Bonhage, A., & Kriston, L. (2014). Verbal Fluency in Focal Epilepsy: A Systematic Review and Meta-analysis. Neuropsychology Review , 24(2), 200-218. Milner, B. (1964). Some effects of frontal lobectomy in man. In: J.M. Warren & K. Akert (Eds.), The frontal granular cortex and behavior (pp. 313-334). New York: McGraw- Hill. Mok, V. C. T., Wong, A., Lam, W. W. M., Fan, Y. H., Tang, W. K., Kwok, T., Hui, A. C. F., & Wong, K. S. (2004). Cognitive impairment and functional outcome after stroke associated with small vessel disease. Journal of Neurology, Neurosurgery & Psychiatry, 75 , 560-566. Mummery, C. J., Patterson, K., Hodges, J. R., & Wise, R. J. (1996). Generating 'tiger' as an animal name or a word beginning with T: differences in brain activation. Proceedings in the Royal Society B: Biological Sciences , 263(1373), 989-995. Nasreddine, Z. S., N. A. Phillips, V. Bedirian, S. Charbonneau, V. Whitehead, I. Collin, J. L. Cummings and H. Chertkow (2005). "The Montreal Cognitive Assessment, MoCA: a
Chapter 3 First Study – Factor analysis of the verbal fluency task 87
brief screening tool for mild cognitive impairment." J Am Geriatr Soc , 53(4): 695-699. O'Connor, B. P. (2000). SPSS and SAS programs for determining the number of components using parallel analysis and velicer's MAP test. Behav Res Methods Instrum Comput, 32 (3), 396-402. Patel, M. D., Coshall, C., Rudd, A. G., & Wolfe, C. D. (2002). Cognitive impairment after stroke: clinical determinants and its associations with long-term stroke outcomes. J Am Geriatr Soc, 50 , 700-706. Petrea, R. E., Beiser, A. S., Seshadri, S., Kelly-Hayes, M., Kase, C. S., & Wolf, P. A. (2009). Gender Differences in Stroke Incidence and Poststroke Disability in the Framingham Heart Study. Stroke, 40 (4), 1032-1037. Raboutet, C., Sauzeon, H., Corsini, M. M., Rodrigues, J., Langevin, S., & N'Kaoua, B. (2010). Performance on a semantic verbal fluency task across time: dissociation between clustering, switching, and categorical exploitation processes. Journal of Clinical and Experimental Neuropsychology , 32(3), 268-280. Reeves, M. J., Bushnell, C. D., Howard, G., Gargano, J. W., Duncan, P. W., Lynch, G., ... Lisabeth, L. (2008). Sex differences in stroke: epidemiology, clinical presentation, medical care, and outcomes. The Lancet Neurology, 7 (10), 915-926. Rende, B., Ramsberger, G., & Miyake, A. (2002). Commonalities and differences in the working memory components underlying letter and category fluency tasks: a dual-task investigation. Neuropsychology , 16(3), 309-321. Reverberi, C., Laiacona, M., & Capitani, E. (2006). Qualitative features of semantic fluency performance in mesial and lateral frontal patients. Neuropsychologia , 44(3), 469-478. Roquer, J., Campello, A. R., & Gomis, M. (2003). Sex Differences in First-Ever Acute Stroke. Stroke, 34 (7), 1581-1585. Rosen, V. M., & Engle, R. W. (1997). The role of working memory capacity in retrieval. Journal of Experimental Psychology: General , 126(3), 211-227. Salthouse, T.A. (2009). Major issues in cognitive aging. Oxford University Press, New York. Schweizer, T. A., Alexander, M. P., Susan Gillingham, B. A., Cusimano, M., & Stuss, D. T. (2010). Lateralized cerebellar contributions to word generation: a phonemic and semantic fluency study. Behav Neurol, 23 , 31-37. Schermelleh-Engel, K., Moosbrugger, H., & Müller, H. (2003). Evaluating the fit of structural equation models: Test of significance and descriptive goodness-of-fit measures. Methods of Psychological Research - Online , 8(2), 23-74. Shao, Z., Janse, E., Visser, K., & Meyer, A. S. (2014). What do verbal fluency tasks measure? Predictors of verbal fluency performance in older adults. Front Psychol, 5 , 772. Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A Compendium of Neuropsychological Tests . New York: Oxford University Press. Szatkowska, I., Grabowska, A., & Szymanska, O. (2000). Phonological and semantic fluencies are mediated by different regions of the prefrontal cortex. Acta Neurobiol Exp (Wars), 60 , 503-508. Tatemichi, T. K., Desmond, D. W., Stern, Y., Paik, M., Sano, M., & Bagiella, E. (1994). Cognitive impairment after stroke: frequency, patterns, and relationship to functional abilities. Journal of Neurology, Neurosurgery & Psychiatry, 57, 202-207. Thompson-Schill, S. L., Swick, D., Farah, M. J., D’Esposito, M., Kan, I. P., & Knight, R. T. (1998). Verb generation in patients with focal frontal lesions: A neuropsychological test
88 Chapter 3 First Study – Factor analysis of the verbal fluency task
of neuroimaging findings. Proceedings of the National Academy of Sciences , 95(26), 15855-15860. Tombaugh, T. N., Kozak, J., & Rees, L. (1999). Normative Data Stratified by Age and Education for Two Measures of Verbal Fluency: FAS and Animal Naming. Archives of Clinical Neuropsychology, 14 (2), 167-177. Troyer, A. K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two components of verbal fluency: Evidence from younger and older healthy adults. Neuropsychology, 11 (1), 138-146. Tukey, J. W. (1977). Exploratory Data Analysis. Addison-Wesley. Unsworth, N., Spillers, G. J., & Brewer, G. A. (2011). Variation in verbal fluency: a latent variable analysis of clustering, switching, and overall performance. Q J Exp Psychol (Hove), 64 (3), 447-466. Van Der Elst, W., Van Boxtel, M. P., Van Breukelen, G. J., & Jolles, J. (2006). Normative data for the Animal, Profession and Letter M Naming verbal fluency tests for Dutch speaking participants and the effects of age, education, and sex. Journal of the International Neuropsychological Society , 12 (01), 80-89. Wagner, S., Sebastian, A., Lieb, K., Tuscher, O., & Tadic, A. (2014). A coordinate-based ALE functional MRI meta-analysis of brain activation during verbal fluency tasks in healthy control subjects. BMC Neurosci, 15 , 19. Whiteside, D. M., Kealey, T., Semla, M., Luu, H., Rice, L., Basso, M. R., & Roper, B. (2016). Verbal Fluency: Language or Executive Function Measure? Appl Neuropsychol Adult, 23 (1), 29-34. Wyller, T. B. (1998). Stroke and gender. The journal of gender-specific medicine: JGSM: the official journal of the Partnership for Women's Health at Columbia , 2(3), 41-45.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 89
“Nothing in life is to be feared. It is only to be understood.“ Marie Curie
Chapter 4
Second Study – Dissociating neural correlates of verbal fluency
Previous lesion and functional neuroimaging studies suggest that semantic and phonological verbal fluency differentially rely on temporal and frontal brain areas, respectively. The present study aims at statistically confirming the suggested double dissociation in a large sample of left hemisphere stroke patients. As will be shown, stroke patients with lesions in temporal regions are more impaired on semantic fluency, whereas patients with lesions in frontal regions are more impaired on phonological fluency, respectively.2
2 The content of this chapter is in preparation for publication as: Schmidt, C. S., Nitschke, K., Bormann, T., Römer, P., Kümmerer, D., Martin, M., Umarova, R. M., Leonhart, R., Dressing, A., Musso, M., Willmes, K., Weiller, C., Kaller, P. C. (in preparation, March 28th, 2018). Dissociating the frontal and temporal correlates of phonological and semantic fluency in a large sample of left hemisphere stroke patients.
90 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
4.1 Abstract
Previous lesion studies suggest that semantic and phonological fluency are differentially subserved by distinct brain regions in the left temporal and the left frontal cortex, respectively. However, as of yet, this often implied double dissociation has not been explicitly investigated due to mainly two reasons: (i) the lack of sufficiently large samples of brain- lesioned patients that underwent assessment of the two fluency variants and (ii) the lack of methodical tools to assess interactions of non-normally distributed behavioral data. In addition, previous studies did not control for task resource artifacts potentially introduced by the generally higher task difficulty of phonological compared to semantic fluency.
We addressed these issues by task-difficulty adjusted assessment of semantic and phonological fluency in 85 chronic patients with ischemic stroke of the left middle cerebral artery. For classical region-based lesion-behavior mapping patients were grouped with respect to their primary lesion location. Building on the extension of the non-parametric Brunner-
Munzel rank-order test to multi-factorial designs, analyses revealed a significant two-way interaction for cue type (semantic vs. phonological) by lesion location (left temporal vs. left frontal vs. other as stroke control group). Subsequent contrast analyses further confirmed the proposed double dissociation by demonstrating that (i) compared to stroke controls, left temporal lesions led to significant impairments in semantic but not in phonological fluency, whereas left frontal lesions led to significant impairments in phonological but not in semantic fluency, and that (ii) patients with frontal lesions showed significantly poorer performance in phonological than in semantic fluency, whereas patients with temporal lesions showed significantly poorer performance in semantic than in phonological fluency.
The anatomical specificity of these findings was further assessed in voxel-based lesion- behavior mapping analyses using the multi-factorial extension of the Brunner-Munzel test.
Voxel-wise analyses identified circumscribed parts of left inferior frontal gyrus and left superior and middle temporal gyrus that significantly double-dissociated with respect to their
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 91 differential contribution to phonological and semantic fluency, respectively. Furthermore, a main effect of lesion with significant impairments in both fluency types was found in left inferior frontal regions adjacent to but not overlapping with those showing the differential effect for phonological fluency.
The present study hence not only provides first explicit evidence for the often implied anatomical double dissociation in verbal fluency but also clearly underlines that its formulation constitutes an oversimplification as parts of left frontal cortex appear to contribute to both semantic and phonological fluency.
Keywords: verbal fluency; brain lesion; frontal lobe; temporal lobe; double dissociation
92 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
4.2 Introduction
Verbal fluency is one of the most frequently used neuropsychological measures of language abilities and executive functioning (Moscovitch, 1994; Strauss et al., 2006; Lezak, Howieson,
Bigler, & Tranel, 2012; Shao et al., 2014), requiring the examinee to generate as many words as possible to a given category cue (semantic fluency) or letter cue (phonological fluency) within a pre-set time interval (e.g. 60 s; Strauss et al., 2006; Lezak et al., 2012).
Previous lesion studies revealed that patients with left frontal lesions produce fewer words in phonological fluency tasks as compared to healthy controls (e.g., Perret, 1974; Pendleton et al., 1982; Miller et al., 1984; Janowsky et al., 1989; Stuss et al., 1994; Gershberg &
Shimamura, 1995; Tucha et al., 1999; Channon & Crawford, 2000; Jurado et al., 2000; Baldo et al., 2001) and to patients with non-frontal lesions (e.g., Milner, 1964; Perret, 1974;
Helmstaedter et al., 1998), whereas patients with lesions in left temporal areas produce fewer words in semantic fluency tasks as compared to healthy controls (Martin et al., 1990; Troyer et al., 1998; Luckhurst & Lloyd-Jones, 2001). Based on this evidence, it hence appears that semantic and phonological fluency are differentially subserved by left temporal and left frontal lobe, respectively (Baldo et al., 2006, 2010). Yet, none of the previous studies has directly investigated this implied double dissociation by statistically assessing the respective interaction between the type of fluency task (semantic vs. phonological) and lesion location
(left temporal vs. left frontal).
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 93
Table 4.1. Overview of performance scores for semantic and phonological fluency reported in previous lesion studies
Semantic Fluency Phonological Fluency Effect of Performance Performance Task Study Authors Year Sample Description Items M ± SD Items M ± SD Difficulty
1 Joanette et 1986 RH Vascular Lesion (n = 35) Animals, Furniture ∑ 32.5 ± 10.5 B, R ∑ 20.0 ± 11.3 sem > phon al. HC (n = 20) Animals, Furniture ∑ 43.2 ± 7.9 B, R ∑ 24.5 ± 13.1 sem > phon
2 Loring et 1994 LH Temporal Lobectomy Animals 13.0 ± 2.8 F, A, S ∑ 23.2 ± 7.8 sem > phon al. (n = 12, pre-operative) RH Temporal Lobectomy Animals 16.9 ±5.1 F, A, S ∑ 28.8 ± 8.7 sem > phon (n = 11, pre-operative)
3 Vilkki et al. 1994 LH Anterior Lesion (n = 19) Animals (20) 1 176 ± 196 sec S (10) 1 122 ± 194 sem > phon 1 sec LH Posterior Lesion (n = 16) Animals (20) 1 175 ± 242 sec S (10) 1 68 ± 76 sec phon > sem 1 RH Anterior Lesion (n = 10) Animals (20) 1 111 ± 94 sec S (10) 1 57 ± 58 sec sem > phon 1 RH Posterior Lesion (n = Animals (20) 1 76 ± 26 sec S (10) 1 30 ± 11 sec phon > sem 1 15)
4 Goulet et 1997 RH Brain Damage (n =15) Animals, Clothes, Ø 12.71 ± 3.52 P, M, T, V, L, N Ø 10.62 ± 3.83 sem > phon al. Sports, Vegetables, Tools, Weapons HC (n =15) Animals, Clothes, Ø 15.53 ± 4.13 P, M, T, V, L, N Ø 14.69 ± 7.57 sem > phon Sports, Vegetables, Tools, Weapons
5 Baldo et al. 1998 LH + RH Frontal Lesion Animals, Fruits, Ø 11.7 F, A, S Ø 7.6 sem > phon (n = 12) Occupations HC (n = 12) Animals, Fruits, Ø 18.5 F, A, S Ø 16.7 sem > phon Occupations
6 Jurado et al. 2000 TBI all (n = 13) Animals, ∑ 44.0 ± 13.8 F, A, S ∑ 37.2 ± 12.6 sem > phon Supermarket Items RH TBI Lesion (n = 3) Animals, ∑ 31.7 ± 1.5 F, A, S ∑ 36.3 ± 4.2 sem > phon Supermarket Items LH TBI Lesion (n = 3) Animals, ∑ 57.0 ± 13.2 F, A, S ∑ 40.0 ± 25.2 sem > phon
94 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
Supermarket Items Bilateral TBI Lesion (n = 7) Animals, ∑ 43.7 ± 12.6 F, A, S ∑ 36.3 ± 9.7 sem > phon Supermarket Items HC (n = 26) Animals, ∑ 51.4 ± 10.6 F, A, S ∑ 47.7 ± 12.2 sem > phon Supermarket Items
7 Szatkowska 2000 LH DLPFC (n = 6) Animals N.A. K N.A. sem ≠ phon et al. RH DLPFC (n = 6) Animals N.A. K N.A. *** LH Ventromedial PFC (n = Animals N.A. K N.A. across all 6) patients RH Ventromedial PFC (n = Animals N.A. K N.A. 6)
8 Baldo et al. 2006 LH Temporal and Frontal Fruits, Animals, Ø 30.5 ± 22.1 F, A, S (90 sec) Ø 18.8. ± 17.6 sem > phon Lesion (n = 48) Supermarket Items (90 sec)
9 Baldo et al. 2010 LH Temporal Lesion (n = 1) Fruits, Animals, ∑ 6 F, A, S ∑ 34.5 phon > sem Supermarket Items LH Frontal Lesion (n = 1) Fruits, Animals, ∑ 35 F, A, S ∑ 2 sem > phon Supermarket Items
10 Robinson et 2012 LH + RH Posterior Lesions Fruits, Vegetables Ø 14.2 ± 5.2 S 13.8 ± 5.8 sem > phon al. (n = 20) LH Frontal Lesion (n = 20) Fruits, Vegetables Ø 12.4 ± 6.1 S 6.8 ± 5.2 sem > phon RH Frontal Lesion (n = 27) Fruits, Vegetables Ø 14.3 ± 6.0 S 12.6 ± 4.2 sem > phon HC (n = 35) Fruits, Vegetables Ø 20.5 ± 5.5 S 16.9 ± 4.7 sem > phon
11 Biesbroek 2016 Ischemic Stroke Patients (n Animals (2min) 23.1 ±10.4 A (1min) 8.0 ± 4.3 sem > phon et al. = 93) N (1min) 7.7 ± 4.4 sem > phon
12 Chouiter et 2016 LH Lesion (n = 108) Animals N.A. M or S N.A. sem > phon al. *** RH Lesion (n = 83) Animals N.A. M or S N.A.
13 Li et al. 2017 Stroke Patients (n=51) Animals, Fruits and Ø 8.39 /Da4/ 2, /Bu4/ 2 Ø 5.67 sem > phon Vegetables, Tools HC (n=39) Animals, Fruits and Ø 17.26 /Da4/ 2, /Bu4/ 2 Ø 11.88 sem > phon Vegetables, Tools
Note. M ± SD, mean ± standard deviation; HC, healthy control; LH, left hemisphere; RH, right hemisphere; TBI, traumatic brain injury; N.A., not available; *** p < .001. Note that the number of applied items for phonological and semantic fluency differ in some of the studies. If reported, the sum ( ∑) and the average (Ø) symbols indicate the type of
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 95 aggregation applied to the performance scores to allow for relative comparisons. 1 In this study performance was assessed as the time subjects needed to generate 20 and 10 semantically (animals) and phonologically (S) cued words, respectively. 2 In this study Chinese syllables were used as phonological cues.
96 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
Furthermore, establishing this proposed double dissociation is considerably hampered by a potential task resource artefact (Shallice, 1988; Davies, 2010) due to general differences in task difficulty between semantic and phonological fluency for which – to the best of our knowledge – none of the previous studies has explicitly controlled for. Semantic fluency is commonly reported to be easier than phonological fluency, as becomes evident from the descriptive overview of performance scores reported in previous lesion studies (see Table 1), although this was not tested statistically in most studies. As this pattern can be observed across various samples of brain lesioned patients as well as healthy controls (Table 1; see also
Katzev et al., 2013; Schmidt et al., 2017), frontal lesion sites associated with poorer performance in phonological fluency may hence simply reflect higher task demands on general processes, which nonetheless subserve both types of verbal fluency rather than a functional specificity for phonological fluency (Shallice, 1988; Davies, 2010).
Finally, left frontal and temporal association cortices constitute large structures subsuming different, functionally heterogeneous subregions. Impairments in semantic and phonological fluency are therefore likely to anatomically dissociate with circumscribed lesions at the level of functionally specific subregions rather than at the coarse level of the overall left temporal and left frontal lobe. In turn, left frontal and/or temporal lobes may also entail subregions, which subserve common processes (Schmidt et al., 2017) and whose integrity is associated with successful performance in both types of verbal fluency (cf. Biesbroek et al. 2016;
Chouiter et al., 2016). In this respect, evidence from neuroimaging studies suggests contributions of frontal areas in semantic fluency (e.g., Frith et al., 1991; Hirshorn &
Thompson-Schill, 2006; Katzev et al., 2013; see Costafreda et al., 2006, and Wagner et al.,
2014, for meta-analyses). Likewise, several lesion studies reported semantic fluency to be affected not only in patients with left temporal but also with left frontal lesions (e.g., Stuss et al., 1996, 1998, 1999; Baldo et al., 1998; Baldo & Shimamura, 1998; Rogers et al., 1998;
Troyer et al., 1998; Szatkowska et al., 2000; Sylvester & Shimamura, 2002; Reverberi et al.,
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 97
2006; Robinson et al., 2012; see Henry & Crawford, 2004, for a meta-analytic review).
Applying voxel-based lesion-behavior mapping (VLBM; Bates et al., 2003; Rorden et al.,
2007) as a complement to the common region-based approaches may hence reveal more specific insights into the differential functional anatomy underlying semantic and phonological fluency, but it requires larger samples of patients and was therefore only very rarely applied (see Baldo et al., 2006; Almairac et al., 2015; Biesbroek et al., 2016; Chouiter et al., 2016, for notable exceptions).
Taken together, as of now, empirical support for the proposed double dissociation of semantic vs. phonological fluency and left temporal vs. left frontal cortex is promising but very limited, as an explicit statistical confirmation is still lacking and would also have to account for the above raised issues of task difficulty and anatomical specificity. In addition, systematic assessment of the effects of left frontal and left temporal brain lesions to phonological and semantic fluency in a non-parametric analysis does not only require a large sample of well- described patients but also calls for appropriate statistical tools to investigate interaction effects in non-normally distributed behavioral data of brain-lesioned patients (cf. Nitschke et al., forthcoming).
In the present study, we therefore addressed all these challenges by explicitly studying the proposed double dissociation between fluency type (semantic vs. phonological) and lesion location (temporal vs. frontal) in a large sample of 85 chronic left hemisphere stroke patients
(see Section 4.3.1). To control for a potential task resource artefact, we administered a
German version of the verbal fluency task that comprised a subset of items with comparable task difficulty for both the semantic and phonological fluency condition (cf. Katzev et al.,
2013; see Section 4.3.2). In a first series of classical region-based analyses patients were grouped with respect to their primary lesion location (i.e., left frontal, left temporal, and neither of them as stroke control group). Non-parametric factorial analyses were conducted based on an extension of the Brunner-Munzel rank-order test towards multi-factorial designs
98 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
(Brunner & Munzel, 2000, 2002; see Section 4.3.5). To further understand the hypothesized interaction effects between lesion location and fluency type, subsequent contrast analyses followed an established predefined operationalization of a double dissociation (see Section
4.3.6). In a second series of voxel-based analyses we applied a newly developed toolbox
(NIX-toolbox; Nitschke et al., forthcoming; cf. Dressing et al., 2018) which also builds on the multi-factorial extension of the Brunner-Munzel rank-order test. This allowed us to directly assess interaction effects in anatomically more specific VLBM analyses. Besides establishing the implied double dissociation, the present study further aimed at revealing common brain regions, which are comparably crucial for both semantic and phonological fluency (Biesbroek et al. 2016; Chouiter et al., 2016).
4.3 Materials and Methods
4.3.1 Participants
Chronic stroke patients were recruited from the Department of Neurology at the University
Medical Center Freiburg and tested at least 5 months post-stroke as part of a larger study on their recovery after ischemic stroke (e.g., Dressing et al., 2018; Beume et al., 2017; Martin et al. 2016). The patient-specific inclusion criterion was first presentation of an ischemic stroke of the middle cerebral artery of the left hemisphere without a hemorrhagic event. Exclusion criteria in the acute phase were an age over 90 years, illiteracy, as well as previous infarcts, intracerebral hemorrhage, traumatic brain injury or contemporary re-infarct. Any major cognitive impairment (e.g. dementia), hearing and visual deficits, or alcohol abuse constituted further exclusion criteria. Patients with an inability to tolerate the MRI examination or neuropsychological testing were also excluded from the study. Every eligible patient was asked to participate and, once consented, tested at the Department of Neurology. The study was approved by the local ethics committee and conducted in compliance with the Helsinki
Declaration of the World Medical Association (http://www.wma.net).
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 99
From the resulting sample of 101 attending chronic stroke patients, 4 patients were excluded because of severe aphasia (i.e. patients were unable to speak). Another 5 patients were excluded because they either did not complete or were unable to perform the task (i.e. task abortion at the request of the patient). Data of one patient was excluded because of technical difficulties with the segmentation and normalization of the lesion data and 4 other patients were excluded because of very small lesions (i.e. lesion volume of less than 1ml). Given the influence of education on performance in verbal fluency tasks (Tombaugh et al., 1999; Strauss et al., 2006) another 2 patients were excluded due to an extraordinary low educational attainment of less than 8 years. Prior to the main analysis, individual data were inspected for outliers within the patient sample. In detail, the total number of words produced for the semantic and the phonological fluency and their respective interquartile ranges were separately computed but revealed no patients with deviant performance.
The final sample comprised 85 chronic stroke patients (23 female) with a mean age (±SD) of
63.97 ± 14.20 years (range 22.4 - 85.8 years), a mean education (±SD) of 13.27 ± 3.46 years
(range 8 - 23 years), and an average post-stroke duration (±SD) of 16.98 ± 18.20 months
(range 5.0 - 66.8 months). The lesion overlay of the final overall sample is displayed in Figure
4.1A.
100 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
Figure 4.1. Lesion overlays (A) for the overall sample of stroke patient s (n = 85, maximum overlap = 25) as well as for the subsample of patients with lesions mainly (B) in left temporal cortex (n = 13, maximum overlap = 10) and (C) left frontal cortex (n = 24, maximum overlap = 14), and (D) the stroke control patients with th eir main lesion sites neither in temporal nor frontal cortex (n = 48, maximum overlap = 20).
4.3.2 Verbal Fluency Task
Participants were administered a German version of the verbal fluency task (Katzev et al.,
2013). The task comprised 8 semantic cues (ca tegories, e.g. vegetables) and 8 phonological cues (letters, e.g. V) that were further classified as being easy or hard (cf. Katzev et al., 2013).
Four different items were presented for each combination of cue type (semantic vs. phonological) and level of difficulty (easy vs. hard), yielding a total of 16 tasks (i.e., 8 semantic categories and 8 letters). Given that semantic fluency is commonly easier than
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 101 phonological fluency (Katzev et al., 2013), applying this manipulation of task difficulty was done to allow for matching the two different fluency types for difficulty, which constitutes a central prerequisite for avoiding task resource artefacts (cf. Shallice, 1988) and for identifying a functionally specific contribution of left temporal and left frontal brain regions independent of differences in task difficulty. In this respect, only the four semantic hard (thereafter semantic) items and the four phonological easy (thereafter phonological) items were used for the present analyses as these were comparable in task difficulty (see also Katzev et al., 2013).
Items and presentation order were identical for all patients (easy semantic, hard semantic, easy phonological, hard phonological items). Instructions for the verbal fluency task were given orally by the experimenter (CS or PR). Participants were told that the verbal fluency task would comprise two different parts (semantic and phonological) and that they were to generate as many nouns as possible within a time limit of 60 s following either a category or a letter. Task rules were explained with exemplar items (i.e., category: Lebensmittel [English: food or groceries]; letter: E). For both conditions only words common in German should be said. No words should be produced twice, no proper names, and no words beginning or ending with the same word stem were allowed. In the phonological condition, legitimate responses were limited to nouns to make the task more comparable to the semantic task. The total number of correct words for the semantic and for the phonological condition was recorded and served as outcome measures for data analyses.
4.3.3 Magnetic Resonance Imaging (MRI)
All participants were administered T1-weighted high-resolution anatomical brain imaging in the acute stroke phase on a 3-Tesla TIM TRIO whole-body MRI scanner (SIEMENS,
Erlangen, Germany) applying T1-weighted magnetization-prepared rapid gradient echo
(MPRAGE) imaging with the following scan acquisition parameters: repetition time (TR),
2200 ms; echo time (TE), 2.15 ms; inversion time (TI), 1100 ms; voxel size, 1 × 1 × 1 mm 3;
102 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
176 sagittal slices. Furthermore, fluid attenuated inversion recovery (FLAIR) images (TR,
9000 ms; TE, 93.0 ms, flip angle, 140°; voxel size, 0.94 × 0.94 × 5.00 mm 3, 23 axial slices) were taken from all patients. In addition, diffusion-weighted images (DWI) were acquired in the acute phase for the delineation of lesion location and size with the following standard sequence: TR, 3700 ms; TE, 100 ms, flip angle, 90°; voxel size, 1.2 × 1.2 × 5 mm 3; 23 axial slices; 3 diffusion-encoding gradient directions with a b-factor of 1000 s/mm 2. The same MRI sequences were also acquired in the chronic phase to control for potential re-infarction.
4.3.4 Lesion Demarcation and Analysis
First, lesions were roughly delineated in the diffusion-weighted images (DWI) using a customized region-of-interest toolbox implemented in SPM8 (r4667; http://www.fil.ion.ucl.ac.uk/spm/software/spm8) running on MATLAB (R2012a; The
MathWorks, Inc., Natick, MA). Individual intensity thresholds were applied to find the best match between the binary lesion map and the diffusion-restricted area. The resulting lesion map was subsequently inspected with MRIcron
(http://people.cas.sc.edu/rorden/mricron/index.html) and manually adjusted if necessary (cf.
Hoeren et al., 2014; Martin et al., 2016). For spatial normalization of the lesion map, the underlying DWI scan was co-registered to the anatomical T1 scan. For four patients no T1 scan with sufficient quality for normalization purposes was available. In these cases, co- registration and normalization was based on the acute FLAIR. High-resolution T1 (or FLAIR) images were segmented using the VBM8 toolbox (r435; http://dbm.neuro.uni- jena.de/vbm/download/) for SPM8. Deformation field parameters for nonlinear normalization into the stereotactic Montreal Neurological Institute (MNI) standard space were then computed using the DARTEL (diffeomorphic anatomical registration through exponentiated lie algebra; Ashburner, 2007) approach implemented in VBM8. Normalization quality of lesion maps was visually checked (cf. Hoeren et al., 2014; Martin et al., 2016).
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 103
4.3.5 Data Analysis
In a first series of region-based lesion-behavior mapping analyses, the hypothesized double dissociation was investigated by classifying patients according to the primary location of their lesion. Classification was done using in-house custom software written in MATLAB (The
MathWorks, Inc., Natick, MA) based on the predefined regions-of-interest (ROIs) for the frontal and temporal lobes (Lancaster et al., 2000) as implemented in the WFU PickAtlas
(Maldjian et al., 2003) available for SPM8. Intersections between the atlas’ ROIs and the individual lesions were calculated. The ROI containing the highest lesion volume was regarded as primary lesion location. Patients with the primary lesion location neither in the frontal nor in the temporal lobe were regarded as stroke control group. Analyses comprised the extension of the non-parametric Brunner-Munzel rank test (Brunner &Munzel, 2002) for multi-factorial designs with within- and between-subjects factors. The ANOVA-type test statistic was chosen. The region-based analysis comprised the between-subjects factor lesion location (i.e. frontal vs. temporal vs. stroke controls) and the within-subject factor cue type
(i.e. semantic vs. phonological). The non-parametric Brunner-Munzel rank test was chosen for the analyses (i) as the distributions within the cells of the factorial design in some cases deviated from normality, and (ii) to keep the approach for the region-based analyses identical to the subsequent voxel-based analyses for which single voxels most often violate the assumptions of a normal distribution and hence for parametric testing (Rorden, Karnath, &
Bonilha, 2007). Given that the factor lesion location comprised three levels, a significant main effect was planned to be followed up by two predefined contrasts that tested for worse performance in the frontal and temporal subgroups against the stroke control patients (one- tailed with an α-level of .05). Significant effects for the interaction were planned to be followed up by predefined contrast analyses as explicated in more detail in Section 2.6 (see below). Contrasts were computed using the non-parametric Brunner-Munzel rank test.
104 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
In a second series of voxel-based analyses, differential VLBM analyses for semantic and phonological fluency were performed using the recently developed NIX-toolbox (Nitschke et al., forthcoming; https://github.com/kainitschke/NIX). The NIX (‘Non-parametric Interaction
Effects’)-toolbox is an open-source toolbox implemented in MATLAB which enables voxel- wise testing of interaction effects in non-normally distributed data. As for the region-based analysis (see above), also the VLBM analyses were based on the ANOVA-type extension of the non-parametric Brunner-Munzel rank test (Brunner & Munzel, 2002) for multi-factorial designs with within- and between-subjects factors. More specifically, the present VLBM analyses comprised the between-subjects factor lesion (i.e. lesion vs. no lesion) and the within-subjects factor cue type (i.e. semantic vs. phonological fluency. Only voxels with 5 or more patients per group (e.g. lesion vs. no lesion) with a cluster size of 10 voxels (33.75 mm 3) or larger and passing an uncorrected threshold of puncorr < .001 were considered significant (cf.
Dressing et al., 2018; Martin et al., 2016). Significant interaction effects of cue type by lesion were planned to be followed up using predefined contrasts for voxel-wise establishing single dissociations (see also section 4.3.6). These contrast analyses were conducted using the non- parametric Brunner-Munzel rank test as implemented in the NIX toolbox.
Given that significant interaction effects of the factors cue type by lesion in these voxel-wise analyses only allow to reveal single dissociations, the full picture of a double dissociation was tested in a separate analysis based on the data from two clusters, namely the main left frontal and the main left temporal cluster showing a significant interaction. As explicated below, patients either had a lesion in the main frontal or the main temporal cluster or in neither of them, but none of the patients had a lesion overlapping these two main clusters. Similar to the region-based analyses, the resulting model hence comprised a two-factorial design with the between-subjects factor lesion location (i.e. frontal vs. temporal vs. stroke controls) and the within-subject factor cue type (i.e. semantic vs. phonological). As the factor lesion location comprised three levels, a significant main effect was planned to be followed up by two
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 105 predefined contrasts that tested for worse performance in the frontal and temporal subgroups against the stroke control patients (one-tailed; see also above). The rationale behind predefined contrasts for following up a significant interaction effect of lesion location by cue type is specified below (see section 4.3.6).
4.3.6 Operationalization for Establishing Single and Double Dissociations
Several concepts were previously suggested for establishing a double dissociation to infer separate localizations for different cognitive functions (e.g. Teuber, 1955; Shallice, 1988;
Crawford et al., 2003; Davies, 2010). A double dissociation emerges if a brain-injured patient
P1 is impaired on task T A but performs normal on task T B whereas a patient P 2 with a different lesion site shows the opposite pattern with normal performance on task T A and impairment in task T B (see also Davies, 2010). However, as this definition of a classical double dissociation
(Teuber 1955; Shallice, 1988) crucially hinges on the applied criterion for the normal range, it was extended by the requirement for a significant within-patient comparison (Crawford et al.,
2003; Davies, 2010) so that, in addition to the pattern above, patient P 1 performs task T B significantly better than task T A and that patient P 2 performs task T A significantly better than task T B.
In the present paper, we therefore adopted the combined approach of testing for a significant classical double dissociation (Shallice, 1988) that also requires significant within-patient comparisons (Crawford et al., 2003). More specifically, the implied double dissociation between left frontal and left temporal brain lesions and differential performance impairments in phonological and semantic fluency, respectively, was tested in both the region-based and the voxel-based lesion-behavior mapping analyses as follows: In a first step, we tested for a significant cross-over interaction effect between the two factors cue type (semantic vs. phonological) and lesion location (frontal vs. temporal vs. stroke controls). Provided a significant interaction was revealed, we then tested the requirements for a classical double
106 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency dissociation in terms of a priori defined contrasts between the levels of the factor lesion location (stroke controls > frontal and stroke controls > temporal) separately for the two levels of the factor cue type . In addition, we also formulated and tested one-sided contrasts reflecting the required within-patient comparison separately for the two levels of interest of the factor lesion location , namely phonological < semantic fluency for frontal patients and semantic < phonological fluency for temporal patients. That is, given that all these contrasts were based on a priori hypotheses and formulated in a directed fashion, the respective statistical tests were consequently one-tailed. Within-patient comparisons of semantic and phonological fluency in the stroke control patients were however formulated in a two-tailed fashion given that no a priori expectations existed. In this respect, the stroke control patients hence served as an unbiased assessment of potential differences in task difficulty between the two cue types.
For the VLBM analyses, the single dissociations emerging at the voxel-level were planned to be followed up in a similar fashion: Providing a significant interaction of the factors cue type by lesion , we then tested for single dissociations (see section 2.5) using two a priori defined contrasts that tested one-sided for worse performance in the presence of a lesion separately for the two levels of the factor cue type . Within-patient comparisons for the two levels of cue type were only conducted in patients with a lesion in a given voxel with one-sided contrasts being complementary for temporal (semantic > phonological) and frontal voxels (phonological > semantic).
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 107
4.4 Results
4.4.1 Region-Based Lesion-Behavior Analysis
In order to investigate the implied double dissociation in classical region -based lesion- behavior analyses, patients were classified with respect to the primary location of their lesion.
Of the 85 chronic stroke patients n = 24 and n = 13 had a lesion primarily in the left frontal and the left temporal lobe, respectively, whereas the remaini ng patients had primarily parietal
(n = 17) and subcortical lesions (n = 31) and were considered as stroke control group (n = 48) in the following. The individual lesion overlays of the three lesion location groups (left temporal, left frontal, stroke cont rols) are depicted in Figures 1B, 1C, and 1D, respectively.
The factorial extension of the non -parametric Brunner-Munzel rank-order test with the between-subjects factor lesion location , the within-subjects factor cue type (semantic vs. phonological), and the total number of words produced as dependent variable revealed a significant two-way interaction of cue type × lesion location (F 2,83 = 6.310, p = .003) whereas the main effects of lesion location (F 2,83 = 1.875, p = .161) and cue type (F 2,83 = 2.495, p =
.117) failed to reach significance.
Figure 4.2. Illustrations of the double dissociation based on the significant two -way interaction cue type × lesion location as revealed in (A) the region-based lesion -behavior
108 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency analysis and in (B) the voxel-based lesion-behavior mapping analysis. Mean rank scores +/- SEM.
For the significant interaction effect as illustrated in Figure 4.2A, we first tested whether the above formulated contrasts fulfilled the requirements for a classical double dissociation :
Patients with a left temporal lesion (red) as compared to patients in the stroke control group
(gray) produced significantly fewer words in the semantic fluency condition (mean rank score
± standard error of the mean, M RS ± SEM, 71.89 ± 12.70 vs. 97.91 ± 5.49, p = .049) but showed no significant differences in the phonological fluency condition (M RS ± SEM, 83.89 ±
15.44 vs. 93.15 ± 8.05, p = .257). The opposite pattern emerged for the patients with a left frontal lesion (blue) who – when compared to patients in the stroke control group (gray) – produced significantly fewer words in the phonological fluency condition (M RS ± SEM, 54.46
± 10.47 vs. 93.15 ± 8.05, p = .004), but showed no significant difference in the semantic fluency condition (M RS ± SEM, 84.69 ± 8.09 vs. 97.91 ± 5.49, p = .108). In addition, we further computed the above formulated contrasts for the within-patient comparison of the two fluency types separately for the left frontal and the left temporal patients: Patients with a left temporal lesion (red) showed a strong trend for producing fewer words in the semantic as compared to the phonological fluency condition (M RS ± SEM, 71.89 ± 12.70 vs. 83.89 ±
15.44, p = .054), whereas patients with a left frontal lesion (blue) produced significantly fewer words in the phonological as compared to the semantic fluency condition (M RS ± SEM, 54.46
± 10.47 vs. 84.69 ± 8.09, p < .001). For patients in the stroke control group (gray), no significant difference was found between the semantic compared to the phonological fluency condition (M RS ± SEM, 97.91 ± 5.49 vs. 93.15 ± 8.05, p = .771) (see Fig. 2A).
Given that the stroke controls comprised a heterogeneous selection of patients with parietal and subcortical lesions, the present results might be at least partly driven by the larger variance in the lesion distribution of the reference group of stroke controls (n = 48) compared
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 109 to the patients of interest with lesions in either left frontal (n = 24) or left temporal lobe (n =
13). In order to preclude this potential bias, we repeated the above analyses with only taking either the patients with parietal (n = 17) or the patients with subcortical lesions (n = 31) as reference group of stroke controls (see Supplementary Materials). These control analyses corroborated the double dissociation which did also hold for separately testing in the two subgroups of patients with parietal and subcortical lesions.
Taken together, the results of the region-based lesion-behavior analysis fully conform to the requirements for establishing a double dissociation both in the classical sense as well as based on significant within-patient comparisons. Lesions to the left temporal and left frontal lobe hence differentially affect performance in semantic and phonological fluency, respectively
(Fig. 4.2A). Notably, this double dissociation cannot be attributed to potential differences in task difficulty between the two types of verbal fluency as these were effectively controlled for
(Fig. 4.2A, stroke control group). The anatomical specificity of these findings was elucidated in subsequent voxel-based lesion-behavior mapping analyses.
4.4.2 Voxel-Based Lesion-Behavior Mapping (VLBM) Analyses
In a second series of voxel-based analyses a VLBM analysis with the factor lesion (yes vs. no) and cue type (semantic vs. phonological) revealed voxels with significant main effect of lesion as well as voxels with a significant interaction of cue type × lesion .
The spatial distribution of the significant main effect of lesion (Fig. 4.3; depicted in magenta) comprised voxels mainly in left frontal cortex that showed a significant difference between patients with and without a lesion irrespective of cue type (semantic or phonological). Note that in order to identify voxels showing a main effect that was not solely driven by a potentially underlying
(subthreshold) interaction effect, the assessment of the significant main effect (p < .001) was restricted only to voxels concurrently showing a p-value for an interaction effect above p > .05. For most clusters of voxels patients with a lesion performed worse as compared to patients without a lesion.
Subsequent contrast analyses further confirmed that in a majority of clusters this pattern was evident
110 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency irrespective of the type of fluency, although the relative effect size s often appeared stronger for phonological fluency. Detailed information on the location, extent, and direction of the main effect of lesion is reported in Table 4.2.
Figure 4.3. Overview of the voxel -wise results in the voxel-based lesion -behavior mapping analysis. Voxels colored in magenta indicate the distribution of the main effect of lesion with significant differences between patients with and without a lesion irrespective of the type of verbal fluency. Voxels colored in red and blue indicate the d istribution of the significant interaction effect of lesion by cue type and the two resulting single dissociations with their opposing impairments in semantic and phonological fluency, respectively. Note that only voxels passing a threshold of p uncorr < .0 01 are displayed. The three panels below the brain rendering illustrate the observed patterns of lesion effects on performance in semantic and phonological fluency for each of the main clusters for the main effect of lesion (magenta) as well as for the dir ections of the interaction effect (red, blue).
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 111
Table 4.2. Overview of clusters with significant voxels for the main effect of lesion
Behavior
Cluster Descriptives Patient Groups Semantic Fluency Phonological Fluency Contrasts
w/o w/ (2) < (4) < (2) < (4) < Peak Anatomical Distribution Size w/o lesion (1) w/ lesion (2) w/o lesion (3) w/ lesion (4) lesion lesion (1) (3) (4) (2)
# x,y,z WFU Pick Atlas labels k n n MRS ±SEM MRS ±SEM MRS ±SEM MRS ±SEM p p p p
1 -42.0 inferior frontal gyrus 3268 74 11 93.23 ± 4.59 69.77 ± 12.84 87.89 ± 6.57 33.14 ± 7.56 .068 <.001 n.a. n.a. 12.0 (20.3%) 37.5 insula (16.8%) middle frontal gyrus (11.6%) precentral gyrus (9.2%) lateral front-orbital gyrus (5.5%)
2 -16.5 putamen (13.0%) 100 80 5 91.82 ± 4.59 64.20 ± 6.24 82.06 ± 6.47 60.80 ± 8.66 <.001 .290 n.a. n.a. 4.5 caudate nucleus (11.0%) 6.0 thalamus (11.0%) globus pallidus (5.0%)
3 -46.5 angular gyrus (57.3%) 96 79 6 87.36 ± 4.51 127.5 ± 9.65 76.41 ± 6.30 138.67 ± .999 .999 n.a. n.a. -43.5 postcentral gyrus (42.7%) 8.49 48.0
4 -33.0 uncus (41.7%) 24 79 6 92.04 ± 4.51 65.83 ± 16.50 83.84 ± 6.42 40.83 ± .081 .027 n.a. n.a. 6.0 superior temporal gyrus 10.93 -24.0 (33.3%)
5 -40.5 superior temporal gyrus 22 77 8 93.94 ± 4.41 54.13 ± 14.74 84.34 ± 6.47 46.75 ± .016 .026 n.a. n.a. 0.0 (59.1%) 15.12 -18.0
112 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
6 -15.0 caudate nucleus (94.1%) 17 76 9 92.34 ± 4.70 72.06 ± 10.45 85.72 ± 6.53 39.28 ± .028 .006 n.a. n.a. 13.5 10.02 13.5
7 -19.5 putamen (64.3%) 14 79 6 90.84 ± 4.65 81.75 ± 11.16 83.87 ± 6.39 40.42 ± .189 .024 n.a. n.a. 3.0 13.62 -15.0
8 -39.0 middle occipital gyrus 14 79 6 88.16 ± 4.57 116.92 ± 77.71 ± 6.41 121.58 ± .999 .999 n.a. n.a. -82.5 (78.6%) 11.66 11.56 16.5 inferior occipital gyrus (14.3%) 9 -55.5 angular gyrus (85.7%) 14 80 5 87.23 ± 4.42 137.50 ± 9.28 77.11 ± 6.25 139.90 ± .999 .999 n.a. n.a. -42.0 supramarginal gyrus 11.71 49.5 (14.3%)
10 -42.0 middle frontal gyrus 13 80 5 90.68 ± 4.57 82.40 ± 15.63 83.56 ± 6.35 36.70 ± .304 .043 n.a. n.a. 33.0 (92.3%) 12.00 21.0
11 -40.5 not assignable, temporal 11 78 7 92.97 ± 4.43 59.29 ± 16.95 84.56 ± 6.46 39.00 ±9.85 .051 .013 n.a. n.a. -13.5 white matter -15.0
12 -13.5 caudate nucleus (18.2%) 11 80 5 91.63 ± 4.56 67.20 ± 11.86 83.88 ± 6.35 33.10 ± 5.95 .013 .003 n.a. n.a. 6.0 22.5
Note. w/o, without; w/, with; k, cluster size in voxels; M RS ±SEM, mean rank score ± standard error of the mean; n.a., not assessed. Coordinates of peak voxels (x,y,z) are provided in MNI space. Anatomical labels were specified based on the predefined regions-of-interest (ROIs) for the frontal and temporal lobes (Lancaster et al., 2000) as implemented in the WFU PickAtlas (Maldjian et al., 2003) available for SPM8. The percentage indicates the amount of overlap for the cluster with the ROI. Note that clusters consisted of voxels passing a threshold of punc < .001 and had a minimum size of k > 10 voxels. Note that in clusters #3, 8, and 9 patients with a lesion performed better than patients without a lesion, thus resulting in p-values > .999 for the predefined contrasts testing for a difference in the opposite direction. Clusters revealing significant (p < .05) or at least marginally significant (p < .10) one-sided contrasts (w/ lesion < w/o lesion) for both cue types are highlighted with p-values in bold font.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 113
The spatial distribution for the significant two-way interaction for cue type × lesion revealed anatomically distinct patterns of two different single dissociations : Performance differences between patients with vs. without a lesion in semantic but not in phonological fluency (Fig.
4.3; depicted in red) were primarily located in left temporal brain areas, whereas differential performance differences between patients with vs. without a lesion in phonological but not in semantic fluency (Fig. 4.3; depicted in blue) were primarily found in left frontal brain areas.
Detailed information on the location, extent, and direction of the two-way interaction for cue type × lesion as well as on the contrast analyses relevant for establishing single dissociations is presented in Table 4.3.
In order to further assess whether the single dissociations in left frontal and left temporal areas indeed reflected a double dissociation, the data from the two largest clusters in left frontal and left temporal cortex from the significant cue type × lesion interaction were used to test for the three-way interaction of lesion (yes vs. no) × location (temporal vs. frontal) × fluency type
(semantic vs. phonological) (see Table 4.3). However, as there were no patients with a lesion in both the frontal and temporal cluster, the 2 × 2 × 2 design was incomplete and therefore rearranged into a 2 × 3 design with cue type (semantic vs. phonological) × lesion location
(patients with a lesion in main left temporal cluster, patients with a lesion in main left frontal cluster, remaining patients as stroke controls). The analysis with the extended Brunner-
Munzel rank-order test for this 2×3 design revealed non-significant main effects of cue type
(F 2,83 = 2.498, p = .114) and lesion location (F 2,83 = 1.726, p = .189) but, most importantly, a significant two-way interaction for cue type × lesion location (F 2,83 = 20.083, p < .001) fulfilling the requirements of a classical double dissociation : As illustrated in Fig. 4.2 B, patients with a lesion in the main left temporal cluster (red) as compared to the stroke control patients (gray) produced significantly fewer words in the semantic fluency condition (M RS ±
SEM, 45.5 ± 17.85 vs. 94.13 ± 4.78, p = .018), whereas no significant difference was found for the phonological fluency condition (M RS ± SEM, 78.5 ± 27.99 vs. 89.85 ± 6.18, p = .414).
114 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
In contrast, patients with a lesion in the main left frontal cluster (blue) as compared to the stroke control patients (gray) produced significantly fewer words in the phonological fluency condition (M RS ± SEM, 30.5 ± 6.07 vs. 89.85 ± 6.81, p < .001) but not in the semantic fluency condition (M RS ± SEM, 86.5 ± 10.88 vs. 94.13 ± 4.78, p = .400). The within-patient comparisons further revealed that patients with a lesion in the main left temporal cluster (red) produced significantly fewer words in the semantic as compared to the phonological fluency condition (M RS ± SEM, 45.5 ± 17.85 vs. 78.5 ± 27.99, p <.001), whereas the opposite pattern was found for the patients with a lesion in the main left frontal cluster (blue) who produced significantly fewer words in the phonological as compared to the semantic fluency condition
(M RS ± SEM, 30.5 ± 6.07 vs. 86.5 ± 10.88, p < .001). For the stroke control patients (gray), no significant difference was found between the semantic and the phonological fluency condition
(M RS ± SEM, 94.13 ± 4.78 vs. 89.85 ± 6.81, p = .561) (Fig. 2B).
Taken together, the results of the voxel-based lesion-behavior mapping analysis not only confirm the pattern of a double dissociation (Fig. 4.2B) that was established in the region- based analysis (Fig. 4.2A) but also substantially extend these findings due to the higher anatomical specificity of the VLBM approach. That is, differentially impaired performance in semantic fluency compared to phonological fluency was particularly observed for lesions of the left middle and superior temporal gyri (Fig. 4.3, red), whereas the opposite pattern with differentially impaired performance in phonological fluency compared to semantic fluency was observed following lesions of left inferior frontal gyrus (IFG) (Fig. 4.3, blue) with a particular focus on pars opercularis but also on pars orbitalis. Beyond this double dissociation, it also became evident that lesions in other parts of left IFG, namely pars triangularis (Fig.
4.3, magenta), led to impairments in both types of verbal fluency.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 115
Table 4.3. Overview of clusters with significant voxels for the interaction cue type × lesion
Behavior
Cluster Descriptives Patient Groups Semantic Fluency Phonological Fluency Contrasts
w/o w/ (2) < (4) < (2) < (4) < Peak Anatomical Distribution Size w/o lesion (1) w/ lesion (2) w/o lesion (3) w/ lesion (4) lesion lesion (1) (3) (4) (2)
# x,y,z WFU Pick Atlas labels k n n MRS ±SEM MRS ±SEM MRS ±SEM MRS ±SEM p p p p
1 -60.0 superior temporal gyrus (44.2%) 2895 80 5 92.98 ± 4.36 45.50 ± 17.85 80.95 ± 6.32 78.50 ± 27.99 .018 .414 <.00 n.a. -55.5 middle temporal gyrus (34.8%) 1 21.0
2 -54.0 precentral gyrus (40.0%) 916 73 12 90.79 ± 4.81 86.54 ±10.88 89.08 ± 6.58 30.50 ± 6.07 .400 <.00 n.a. <.001 1.5 inferior frontal gyrus (21.1%) 1 24.0 middle frontal gyrus (6.2%)
3 -48.0 inferior frontal gyrus (68.9%) 180 80 5 88.99 ± 4.61 109.40 ± 6.79 84.11 ± 6.31 28.00 ± 8.64 >.99 .008 n.a. <.001 31.5 9 -12.0
4 -57.0 middle temporal gyrus (63.2%) 76 80 5 92.98 ± 4.36 45.50 ± 17.85 80.95 ± 6.32 78.50 ± 27.99 .018 .414 <.00 n.a. -10.5 inferior temporal gyrus ( 35.5%) 1 -30.0
5 -40.5 superior temporal gyrus(85.5%) 62 79 6 92.22 ± 4.51 63.50 ± 15.86 79.51 ± 6.42 97.92 ± 20.32 .064 >.99 <.00 n.a. 21.0 9 1 -31.5
6 -37.5 middle frontal gyrus (73.2%) 56 78 7 90.83 ± 4.67 83.07 ± 11.65 86.05 ± 6.32 22.36 ± 6.42 .241 <.00 n.a. <.001 -3.0 1 52.5
7 -52.5 inferior frontal gyrus (64.1%) 39 80 5 89.94 ± 4.61 94.30 ± 11.73 83.55 ± 6.38 36.90 ± 6.18 >.99 .010 n.a. <.001 24.0 middle frontal gyrus (35.9%) 9
116 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
7.5
8 -57.0 middle temporal gyrus (65.7%) 35 80 5 90.02 ± 4.39 93.00 ± 27.93 79.38 ± 6.29 103.70 ± >.99 >.99 <.00 n.a. -63.0 middle occipital gyrus (28.6%) 27.39 9 9 1 16.5
9 -45.0 superior temporal gyrus( 94.1%) 34 80 5 92.98 ± 4.36 45.50 ±17.85 80.95 ± 6.32 78.50 ± 27.99 .018 .414 <.00 n.a. 18.0 1 -31.5
10 -31.5 not assignable, frontal white 34 76 9 91.25 ± 10.47 81.28 ± 9.76 87.02 ± 9.98 28.33 ± 6.32 .151 <.00 n.a. <.001 1.5 matter 1 31.5
11 -40.5 middle frontal gyrus (83.9%) 31 80 5 90.06 ± 4.59 92.30 ± 14.50 84.77 ± 6.23 17.40 ± 7.96 >.99 .002 n.a. <.001 13.5 9 36.0
12 -36.0 inferior frontal gyrus (33.3%) 24 80 5 88.99 ± 4.61 109.40 ± 6.79 84.11 ± 6.31 28.00 ± 8.64 >.99 .008 n.a. <.001 43.5 9 -1.5
13 -52.5 middle temporal gyrus (100.0%) 17 78 7 93.40 ± 4.43 54.43 ± 15.04 81.01 ± 6.50 78.50 ± 18.29 .015 .418 <.00 n.a. -64.5 1 1.5
14 -55.5 superior temporal gyrus (50.0%) 12 80 5 92.98 ± 4.36 45.50 ± 17.85 80.95 ± 6.32 78.50 ± 27.99 .018 .414 <.00 n.a. 6.0 middle temporal gyrus (50.0%) 1 -28.5
15 -34.5 middle frontal gyrus (16.7%) 12 80 5 89.59 ± 4.60 99.90 ± 12.06 84.37 ± 6.28 23.80 ± 8.38 >.99 .004 n.a. <.001 10.5 9 49.5
Note. w/o, without; w/, with; k, cluster size in voxels; M RS ±SEM, mean rank score ± standard error of the mean; n.a., not assessed. Coordinates of peak voxels (x,y,z) are provided in MNI space. Anatomical labels were specified based on the predefined regions-of-interest (ROIs) for the frontal and temporal lobes (Lancaster et al., 2000) as implemented in the WFU PickAtlas (Maldjian et al., 2003) available for SPM8. The percentage indicates the amount of overlap for the cluster with the ROI. Note that clusters consisted of voxels passing a threshold of punc < .001 and had a minimum size of k > 10 voxels. Clusters in frontal and temporal cortex significantly (p < .05) or at least marginally significantly (p < .10) conforming to the criteria for single dissociations in the expected directions are highlighted with p-values in bold font.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 117
4.5 Discussion
The present study provided conclusive evidence for the proposed double dissociation between differential impairments in semantic and phonological fluency and lesions in left temporal and left frontal lobe (Fig. 4.2). To our knowledge this is the first study that explicitly addressed and statistically confirmed this double dissociation in a large sample of chronic stroke patients using both region-based as well as voxel-based lesion-behavior mapping analyses. Our findings hence corroborate previous assumptions from both the lesion and neuroimaging literature which suggested this double dissociation (e.g., Mummery et al., 1996; Baldo et al.,
2006, 2010). However, the present study goes substantially beyond the mere experimental confirmation of a since long proposed double dissociation: The results from the VLBM analysis further revealed a significant main effect for lesion particularly in the pars triangularis of the left IFG, showing that patients with a lesion in these voxels performed worse than patients without a lesion irrespective of the type of verbal fluency (semantic or phonological) (see Fig. 4.3). Taking the double dissociation as evidence for a coarse attribution of cognitive processing in semantic and phonological fluency to the functions of the temporal and frontal cortex, respectively, seems hence to be an oversimplification.
4.5.1 Double Dissociation between Left Frontal and Temporal Lesions and Impairments in
Phonological and Semantic Fluency
The present study’s main objective concerned the statistical confirmation of double- dissociating differential contributions of left frontal and temporal cortex in phonological and semantic fluency that was previously implied but has not been explicitly proven. In this respect, one might of course argue that - in view of the partial evidence for the underlying single dissociations (e.g., Milner, 1964; Perret, 1974; Henry & Crawford, 2004; Baldo et al.,
2006, 2010) - incidental proof for the double dissociation is little surprising but coercive.
However, although previous lesion studies have demonstrated that comparing healthy controls
118 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency to patients with left frontal and temporal lesions resulted in impairments in phonological
(Perret, 1974; Pendleton et al., 1982; Miller et al., 1984; Janowsky et al., 1989; Stuss et al.
1994; Gershberg & Shimamura, 1995; Tucha et al., 1999; Channon & Crawford, 2000; Jurado et al., 2000; Baldo et al., 2001; Sylvester & Shimamura, 2002) and semantic fluency (Martin et al., 1990; Troyer et al., 1998; Luckhurst & Lloyd-Jones, 2001), respectively, the meta- analysis of Henry and Crawford (2004) has pointed out that the sensitivity of semantic fluency for frontal lesions is comparable to that of phonological fluency (cf. Stuss et al., 1996,
1998, 1999; Baldo et al., 1998; Baldo & Shimamura, 1998; Rogers et al., 1998; Troyer et al.,
1998; Szatkowska et al., 2000; Robinson et al., 2012; Biesbroek et al., 2016). In addition, previous research has provided only little direct evidence for differential patterns in terms of significant within- or between-patient comparisons that was neither independently nor fully covering the overall pattern of the proposed double dissociation (but see Baldo et al., 2006,
2010, for notable exceptions). In this respect, patients with frontal lesions were reported to have worse performance in phonological fluency compared to semantic fluency (Rogers et al.,
1998) and compared to patients with temporal lesions (Milner, 1964; Perret, 1974;
Helmstaedter et al., 1998). Taken together, the (implicit) evidence for the double dissociation from previous studies has been less clear than it may appear at first glance. Furthermore, while most previous lesion studies relied on region-based lesion-behavior mapping on the level of lobes, the findings on contributions of left frontal cortex in semantic fluency contrasting the proposed pattern of a double dissociation particularly indicate the need for analyses at a higher spatial resolution to resolve the equivocality of potentially distinct and shared neural correlates in the frontal lobe.
As of yet, only four studies have applied voxel-based lesion-behavior mapping (VLBM) analyses either in chronic (Baldo et al., 2006) or (sub)acute stroke patients (Biesbroek et al.,
2016; Chouiter et al., 2016) or in patients with low-grade glioma (Almairac et al., 2015) that were assessed with both semantic and phonological fluency. In line with the present results,
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 119
Baldo et al. (2006) revealed differential associations of impaired performance in phonological and semantic fluency with lesions primarily in left frontal (mainly precentral gyrus and IFG pars opercularis) and left temporal cortex (mainly superior and middle temporal gyrus), respectively, but they provided only qualitative subtraction maps for visualization of the differential lesion distributions and did not explicitly test the significance of the apparent double dissociation. Additional analyses by Baldo et al. (2006) further implicated that performance in hard semantic items also relied on left frontal cortex. Almairac et al. (2015) reported associations of deficits in semantic fluency with lesions in the deep white matter underlying the left IFG and left superior temporal gyrus, the deep sylvian fissure, the posterior orbito-frontal areas, the striatum and in the insula, but no significant results for phonological fluency. Biesbroek et al. (2016) found again overlapping anatomical correlates in the left insula and left frontal (inferior frontal, middle frontal, and precentral gyri, rolandic operculum) cortex and discordant correlates of semantic and phonological fluency in left temporal (inferior temporal, lingual, and fusiform gyri, medial temporal lobe) and left frontal cortex (middle frontal gyrus), respectively, but did also not statistically compare these differential patterns. Chouiter et al. (2016) reported associations with lesions shared by both fluency types that however concerned left parietal (angular and parts of supramarginal gyrus) and left temporal cortex (superior and middle temporal gyri). In addition, semantic fluency was differentially associated with lesions in middle temporal gyrus and the pallidum, whereas phonological fluency was associated with lesions in anterior middle and superior temporal areas, the rolandic operculum and the supramarginal gyrus (Chouiter et al., 2016).
Although none of these studies directly assessed the double dissociation in statistical terms, the findings from these VLBM studies either are fully in line (Baldo et al., 2006) or at least partly concur (Almairac et al., 2015; Biesbroek et al., 2016; Chouiter et al., 2016) with the anatomy of the present differential patterns. Yet, as systematic differences of task difficulty between phonological and semantic fluency have not been explicitly controlled for in any of
120 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency these studies (see also Table 4.1), it cannot be formally excluded that findings of differential contributions in these VLBM studies could be fully or partly driven by the consistently reported higher task difficulty of phonological fluency and might hence reflect an unspecific task-resource artifact (cf. Shallice, 1988; Davies, 2010). Further note, that the same argument holds true for all other fluency studies that applied region-based lesion-behavior mapping approaches without controlling for the immanent differences in task difficulty between both types of verbal fluency (Table 4.1).
Taken together and to the best of our knowledge, the present study hence comprises the first explicit evidence for the double dissociation between differential contributions of left frontal and left temporal cortex in phonological and semantic fluency, respectively. The present findings also highlight the value of applying non-parametric factorial analyses in the context of lesion studies. In this respect, by exploiting the higher spatial resolution of voxel-based compared to region-based lesion-behavior mapping analyses, the present study further allowed to resolve apparent discrepancies on distinct versus shared contributions of left frontal areas in both fluency types.
4.5.2 Distinct and Shared Contributions of Left IFG in Phonological and Semantic Fluency
The present results indicate that the specificity and/or sensitivity of frontal lesions for deficits in the two different types of verbal fluency appears to hinge on their precise location. That is, while lesions in the pars opercularis (and partly in pars orbitalis) of the IFG lead to an isolated impairment in phonological fluency, lesions in pars triangularis lead to general impairments in both semantic and phonological fluency (Fig. 4.3). In this respect, various researchers have suggested that the left frontal cortex and in particular the left IFG is not only involved in phonological but also in semantic fluency (e.g., Troyer et al., 1998; Henry & Crawford, 2004;
Reverberi et al., 2006; Robinson et al., 2012; Biesbroek et al., 2016). However, these region-
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 121 based lesion-behavior mappings did not allow differentiating between shared or distinct neural correlates (or both) of phonological and semantic fluency in the left frontal lobe.
Anatomically more specific, Costafreda and colleagues (2006) proposed a functional segregation of the left IFG in semantic and phonological processing. Based on a systematic review of functional MRI studies of verbal fluency, they suggested that the posterior-dorsal part of the left IFG (Brodmann area [BA] 44; roughly corresponding to pars opercularis) is associated with phonological, whereas the anterior-ventral part of the left IFG (BA 45; roughly corresponding to pars triangularis) is associated with semantic fluency (Costafreda et al., 2006). Testing the proposed dissociation directly in a within-subjects fMRI experiment,
Heim and colleagues (2008) only found overlapping activations in BA 45 for phonological and semantic fluency, whereas the former led to specific activation in BA 44 (see also
Wagner et al., 2014). Although this seems partly in contrast with the proposal of Costafreda et al. (2006), it nonetheless appears to mirror the present findings as well as those of Biesbroek et al. (2016), who also reported common impairments in semantic and phonological fluency following lesions in left anterior IFG (pars triangularis, BA 45; see also Robinson et al., 2012) and specific phonological impairments following lesions in left posterior IFG (pars opercularis, BA 44).
Already Baldo et al. (2006) and Reverberi et al. (2006) suggested that the degree of frontal involvement in semantic fluency may rely on the task demands exerted by different categories as an efficient strategy may particularly affect fluent word retrieval from smaller categories
(i.e. hard semantic items). Taking item difficulty and individual ability into account in a recent fMRI study in healthy participants, Katzev and colleagues (2013) were able to outline the determinants of the proposed differential involvement of BA 44 and 45 for phonological and semantic fluency, respectively (cf. Costafreda et al., 2006): Higher activation for semantic than phonological fluency was evident in posterior BA 45 particularly in low-performing respondents and in anterior BA 45 only for semantic hard items, whereas higher activation for
122 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency phonological than semantic fluency was robustly found in BA 44 but to a larger extent in high-performing respondents (Katzev et al., 2013). Compared against implicit baseline, BA
44 and posterior BA 45 were rather selectively activated by phonological and semantic fluency, respectively. In contrast, anterior BA 45 was activated by both types of fluency with the strongest activations for semantic hard items, intermediate levels for phonological hard and easy items, but with no significant activation for semantic easy items.
The present finding of impairments in (comparably difficult) semantic hard items and phonological easy items following left IFG lesions in pars triangularis (BA 45) hence concur with the fMRI data of Katzev et al. (2013). Although neither the focus nor part of the present analyses, one would further expect a differential impairment comparing semantic hard vs. easy items, since the latter did not yield any activation of BA 45 in the study of Katzev et al.
(2013). Exploratory VLBM analyses in semantic hard and easy items however failed to reveal such an interaction between lesion and item difficulty in addition to (or instead of) a main effect of lesion in left IFG pars triangularis (not shown).
Besides a potential lack of statistical power, the substantial age differences between the samples of Katzev et al. (mean ± SD, 26.1 ± 6.6 years) and the present study (63.97 ± 14.20 years) might also indicate that semantic easy items exert possibly higher task demands on strategically controlled retrieval in older adults (at least partly reliant on anterior BA 45) as compared to rather automatically triggered responses in younger adults (without essential recruitment of anterior BA 45). In accord with this, older adults have been reported to generate smaller clusters and to switch less frequently in semantic fluency than younger adults
(Troyer et al., 1997; Lanting et al., 2009; but see Mayr & Kliegl, 2000). Directly comparing fMRI activation of older and younger adults has further demonstrated higher activation for phonological compared to semantic fluency in BA 44 and anterior BA 45 in younger adults, whereas older adults failed to show this distinction but recruited these areas in both conditions to a similar extent (Meinzer et al., 2009). Further given that task difficulty of phonological
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 123 and semantic fluency was not matched with the latter being significantly easier than the former in the younger but not the older adults (Meinzer et al., 2009) finally closes the circle towards the differential findings of Katzev et al. (2013) in younger adults by implying that the performance of older healthy adults and most likely also older stroke patients (as in the present sample) relies on BA 45 not only for semantic hard but also for semantic easy items.
4.5.3 Cognitive Correlates Underlying Lesion-Specific Impairments in Verbal Fluency
Dissociable neural contributions in left frontal and left temporal cortex in verbal fluency presumably reflect the differences in the retrieval processes underlying the phonological and semantic condition (Katzev et al., 2013). The present differential effect for lesions in left IFG pars opercularis (or BA 44) leading to specific impairment in phonological fluency is hence likely to be related to the serial search based on systematic syllabification of initial letters
(Mummery et al., 1996; Rende et al, 2002; Henry & Crawford, 2004). In this respect, it was suggested that the left IFG performs sensorimotor encoding of auditory phonetic input
(Demonet et al., 1994; Bockheimer, 2002) that overlaps with processes of inner speech such as motor programming and articulation (Indefrey & Levelt, 2000) and presumably supports proper subvocal syllabification of initial letters in phonological fluency.
In contrast to the rather artificial kind of search in phonological fluency, retrieval of words in semantic fluency relies on the natural organization of conceptual knowledge stored in the temporal lobe (Gruenewald & Lockhead, 1980; Katzev et al., 2013). Already Gruenewald and
Lockhead (1980) proposed that retrieving words from a given semantic category is a two- stage process consisting of (i) the top-down identification of a task-relevant subcategory as the foundation for (ii) the subsequent bottom-up retrieval of appropriate category members triggered by automatic associations (or semantic proximity) within that subcategory.
Expanding on this and in an attempt to disentangle executive from semantic processes, Troyer et al. (1997) suggested that clustering in terms of the (automatic) word retrieval within
124 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency subcategories and (controlled) switching between these subcategories constitute two important components of semantic fluency (see also Abwender et al., 2001). This concept of switching and clustering is however not specific for semantic fluency but can also be applied to phonological fluency to disentangle executive from lexical processes (Troyer et al., 1998).
With respect to shared contributions in left frontal cortex, mainly two different types of control processes have been discussed to be associated with recruitment of left IFG during semantic fluency (Katzev et al., 2013): either (post-retrieval) selection from among competing co-activated representations or words (Hirshorn and Thompson-Schill, 2006; see also
Thompson-Schill et al. 1997, 1998) or top-down controlled semantic search when bottom-up
(automatic) retrieval (based on association chains; Collins & Loftus, 1975) is insufficient
(e.g., Badre et al., 2005; Wagner et al., 2001). While the present lesion data do not allow deciding between these two alternative explanations, the fMRI data of Katzev et al. (2013) were in favor for the latter interpretation of the functional role of anterior BA 45 (but did not exclude the former for posterior BA 45; see also Robinson et al., 1998). Furthermore, switching between semantic subcategories has been related to the integrity of left frontal lobe
(Troyer et al., 1998) and the functional activation of pars triangularis in particular (Hirshorn and Thompson-Schill, 2006), whereas semantic clustering has been associated with integrity of left temporal lobe (Troyer et al., 1998). It has been argued that not switching per se is impaired in patients with frontal lesion, but rather the application of an efficient search strategy as patients were found to exhibit increased switching and reduced semantic proximity
(Reverberi et al., 2006).
Given that anterior BA 45 was found to exert intermediate activation in phonological hard and easy items (Katzev et al., 2013) and that also continuous word generation in phonological fluency is at least to some extent semantically structured (Abwender et al., 2001; Schwartz et al., 2003; Azuma, 2004), the present finding of a main effect of lesion in pars triangularis that lead to general impairments in both semantic and phonological fluency may hence reflect
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 125 impairments in top-down controlled semantic search and/or retrieval. Alternatively, it may also signify overlapping deficits in domain-independent control processes as switching subcategories was also found to be affected in phonological fluency following left lateral frontal lesions (Troyer et al., 1998). For instance, the post-retrieval selection hypothesis
(Thompson-Schill et al., 1997, 1998) would predict deficits that are at least partly due to an impaired suppression of inappropriate responses automatically activated by semantic association chains which may similarly affect performance in phonological and semantic fluency (Katzev et al., 2013). In addition, overlapping impairments in both fluency types may also be directly caused or indirectly mediated by deficits in less specific domain-general control functions such as energization, self-monitoring, attention and processing speed (cf.
Biesbroek et al., 2016).
Finally, it should be noted that evidence from the clustering and switching approach as put forward by Troyer et al. (1997, 1998) may not allow to unequivocally disentangle executive control processes from genuine semantic processes (Mayr & Kliegl, 2000; Mayr, 2002; see also Reverberi et al., 2006). In particular, the two resulting components are not independent
(particularly if fluency is assessed in a time-restricted manner as usual; cf. Troyer et al., 1997) so that a reduced number of switches per se may not only reflect difficulties in (top-down controlled) accessing a new subcategory but may also depend on difficulties in (bottom-up driven automatic) semantic retrieval within subcategories (Mayr, 2002). In other words, the more time is spent on one category, the less time remains for switching to new categories.
Following Mayr (2002), this interpretation uncertainty can only be resolved by the analysis of detailed timing protocols (i.e., within- and between-cluster retrieval latencies) – thus indicating not only the need but also the methodology for future research.
126 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
4.5.4 Limitations
There are several limitations to the interpretation of the present results to be noted. With respect to the formal criteria for claiming a valid double dissociation, Shallice (1988) argued that Teuber’s (1955) formulation of a classical double dissociation (i.e. differential normal/impaired performance of two patients P 1 and P 2 in tasks T A and T B) is “not sufficient grounds for inferring separate subsystems” (p. 234) if the two tasks have different difficulties and may thus differentially call on given resources. He therefore proposed to consider the between-patient comparison with patient P 1 performing significantly worse than patient P 2 on task T A and with patient P 2 performing significantly worse than patient P 1 on task T B. In the present paper, we adopted an alternative but also established approach based on the extension of the tests for a classical double dissociation by significant within-patient comparisons
(Crawford et al., 2003; Davies, 2010) so that patient P 1 performs task T B significantly better than task T A and that patient P 2 performs task T A significantly better than task T B. This was done mainly for two reasons: First, by explicitly controlling for comparable levels of task difficulty between semantic and phonological fluency, we have directly addressed and resolved the problem of potential task resource artifacts put forward by Shallice (1988).
Second, and more pragmatically, by accounting for systematic effects in inter-individual variance the within-patient approach is superior to the between-patient approach in terms of statistical test power, thus reducing the Type II error of false negative findings. Based on
Monte-Carlo simulations McIntosh (2017) even argued that solely the significant within- patient comparison would be logically necessary to establish a (double) dissociation whilst sufficiently controlling for Type I error of false positive findings, whereas the here additionally applied tests for a classical (double) dissociation (i.e., impaired performance significant from normal on task T A but not on task T B in patient P 1 and vice versa in patient
P2; cf. Teuber, 1955; Crawford et al., 2003; Davies, 2010) would only constitute further qualifying information.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 127
Furthermore, the obvious ambiguity in attributing impairments in verbal fluency following circumscribed lesions (for instance, in anterior or posterior pars triangularis) to specific cognitive functions (e.g., switching semantic/phonological subcategries vs. post-retrieval selection) results at least partly from the fact that the majority of studies used rather coarse classifications on the level of lobes (e.g., frontal vs. non-frontal). This lacking spatial acuity hampered any direct reference between the present and previous results particularly as none of the few extant studies addressing impairments in verbal fluency on the level of processes had applied spatially resolved VLBM analyses. In addition, all considerations with reference to microstructural cytoarchitecture (i.e., Brodmann areas) are purely speculative as the present data and methodology permit only macroanatomical mappings. In this respect it has to be noted that the large inter-individual anatomical variability of BA 44 and 45 (Amunts et al.,
1999) may have also constrained the present analyses (see also Amunts & Willmes, 2006).
4.6 Conclusion
The present region-based and voxel-based lesion-behavior mapping analyses in a large sample of chronic stroke patients corroborate and extent previous findings and also resolve discrepancies in the extant literature on the neural correlates of verbal fluency by demonstrating distinct as well as shared contributions. More precisely, a significant double dissociation confirmed that the integrity of left superior and middle temporal gyri is specifically crucial for semantic fluency whereas the integrity of pars opercularis of left IFG specifically subserves phonological fluency. Furthermore, a main effect of lesion in pars triangularis of left IFG affirmed that its integrity is generally critical for word retrieval in both semantic and phonological fluency.
128 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
4.7 Acknowledgements
The authors report no conflicts of interest. The present research was partly supported by a grant of the BrainLinks-BrainTools Cluster of Excellence (C.W., C.P.K, project #36) funded by the German Research Foundation (DFG; grant # EXC 1086) and by a grant of the Müller-
Fahnenberg Foundation (C.P.K., C.W.). C.S.M.S. received scholarship funds from the State
Law on Graduate Funding of the University of Freiburg, Germany.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 129
4.8 References
Abwender, D. A., Swan, J. G., Bowerman, J. T., & Connolly, S. W. (2001). Qualitative analysis of verbal fluency output: Review and comparison of several scoring methods. Assessment , 8(3), 323-338. Almairac, F., Herbet, G., Moritz-Gasser, S., de Champfleur, N. M., & Duffau, H. (2015). The left inferior fronto-occipital fasciculus subserves language semantics: a multilevel lesion study. Brain Structure and Function , 220 (4), 1983-1995. Amunts, K., Schleicher, A., Bürgel, U., Mohlberg, H., Uylings, H., & Zilles, K. (1999). Broca's region revisited: cytoarchitecture and intersubject variability. Journal of Comparative Neurology , 412 (2), 319-341. Amunts, K., & Willmes, K. (2006). From intersubject variability in clinical syndromes to anatomical variability. Brain and Language , 96 (2), 147-150. Ashburner, J. (2007). A fast diffeomorphic image registration algorithm. Neuroimage , 38 (1), 95-113. Azuma, T. (2004). Working memory and perseveration in verbal fluency. Neuropsychology , 18 (1), 69. Badre, D., Poldrack, R. A., Paré-Blagoev, E. J., Insler, R. Z., & Wagner, A. D. (2005). Dissociable controlled retrieval and generalized selection mechanisms in ventrolateral prefrontal cortex. Neuron , 47 (6), 907-918. Baldo, J. V., Schwartz, S., Wilkins, D., & Dronkers, N. F. (2006). Role of frontal versus temporal cortex in verbal fluency as revealed by voxel-based lesion symptom mapping. Journal of the International Neuropsychological Society, 12 (06), 896-900. Baldo, J. V., Schwartz, S., Wilkins, D. P., & Dronkers, N. F. (2010). Double dissociation of letter and category fluency following left frontal and temporal lobe lesions. Aphasiology, 24 (12), 1593-1604. Baldo, J. V., & Shimamura, A. P. (1998). Letter and category fluency in patients with frontal lobe lesions. Neuropsychology , 12 (2), 259. Baldo, J. V., Shimamura, A. P., Delis, D. C., Kramer, J., & Kaplan, E. (2001). Verbal and design fluency in patients with frontal lobe lesions. Journal of the International Neuropsychological Society , 7(5), 586-596. Bates, E., Wilson, S. M., Saygin, A. P., Dick, F., Sereno, M. I., Knight, R. T., & Dronkers, N. F. (2003). Voxel-based lesion–symptom mapping. Nature Neuroscience ,6(5), 448. Beume, L. A., Martin, M., Kaller, C. P., Klöppel, S., Schmidt, C. S., Urbach, H., ... & Umarova, R. M. (2017). Visual neglect after left-hemispheric lesions: a voxel-based lesion–symptom mapping study in 121 acute stroke patients. Experimental brain research , 235 (1), 83-95. Biesbroek, J. M., van Zandvoort, M. J., Kappelle, L. J., Velthuis, B. K., Biessels, G. J., & Postma, A. (2016). Shared and distinct anatomical correlates of semantic and phonemic fluency revealed by lesion-symptom mapping in patients with ischemic stroke. Brain Structure and Function , 221 (4), 2123-2134. Bookheimer, S. (2002). Functional MRI of language: new approaches to understanding the cortical organization of semantic processing. Annual review of neuroscience , 25 (1), 151-188.
130 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
Brunner, E., & Munzel, U. (2000). The nonparametric behrens-fisher problem: Asymptotic theory and a small-sample approximation. Biometrical Journal, 42(1), 17–25. Brunner, E., & Munzel, U. (2002). Nichtparametrische Datenanalyse. Springer, Berlin, Heidelberg , 34 , 59. Channon, S., & Crawford, S. (2000). The effects of anterior lesions on performance on a story comprehension test: Left anterior impairment on a theory of mind-type task. Neuropsychologia , 38 (7), 1006-1017. Chouiter, L., Holmberg, J., Manuel, A. L., Colombo, F., Clarke, S., Annoni, J. M., & Spierer, L. (2016). Partly segregated cortico-subcortical pathways support phonologic and semantic verbal fluency: A lesion study. Neuroscience, 329 , 275-283. Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological review , 82 (6), 407. Costafreda, S. G., Fu, C. H., Lee, L., Everitt, B., Brammer, M. J., & David, A. S. (2006). A systematic review and quantitative appraisal of fMRI studies of verbal fluency: role of the left inferior frontal gyrus. Human Brain Mapping , 27 (10), 799-810. Crawford, J. R., Garthwaite, P. H., & Gray, C. D. (2003). Wanted: Fully operational definitions of dissociations in single-case studies. Cortex , 39 (2), 357-370. Davies, M. (2010). Double dissociation: Understanding its role in cognitive neuropsychology. Mind & Language , 25 (5), 500-540. Demonet, J. F., Price, C., Wise, R., & Frackowiak, R. S. J. (1994). A PET study of cognitive strategies in normal subjects during language tasks: Influence of phonetic ambiguity and sequence processing on phoneme monitoring. Brain , 117 (4), 671-682. Dressing, A., Nitschke, K., Kümmerer, D., Bormann, T., Beume, L., Schmidt, C. S., ... & Kaller, C. P. (2018). Distinct Contributions of Dorsal and Ventral Streams to Imitation of Tool-Use and Communicative Gestures. Cerebral Cortex , 28(2), 474-492.. Frith, C. D., Friston, K. J., Liddle, P. F., & Frackowiak, R. S. J. (1991). A PET study of word finding. Neuropsychologia , 29 (12), 1137-1148. Gershberg, F. B., & Shimamura, A. P. (1995). Impaired use of organizational strategies in free recall following frontal lobe damage. Neuropsychologia , 33 (10), 1305-1333. Gruenewald, P. J., & Lockhead, G. R. (1980). The free recall of category examples. Journal of Experimental Psychology: Human Learning and Memory , 6(3), 225. Helmstaedter, C., Gleissner, U., Zentner, J., & Elger, C. E. (1998). Neuropsychological consequences of epilepsy surgery in frontal lobe epilepsy. Neuropsychologia , 36 (7), 681-689. Henry, J. D., & Crawford, J. R. (2004). A meta-analytic review of verbal fluency performance following focal cortical lesions. Neuropsychology, 18 (2), 284-295. Hirshorn, E. A., & Thompson-Schill, S. L. (2006). Role of the left inferior frontal gyrus in covert word retrieval: neural correlates of switching during verbal fluency. Neuropsychologia , 44 (12), 2547-2557. Hoeren, M., Kümmerer, D., Bormann, T., Beume, L., Ludwig, V. M., Vry, M. S., ... & Weiller, C. (2014). Neural bases of imitation and pantomime in acute stroke patients: distinct streams for praxis. Brain , 137 (10), 2796-2810. Indefrey, P., & Levelt, W. J. (2000). The neural correlates of language production. In The new cognitive neurosciences; 2nd ed. (pp. 845-865). MIT press.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 131
Janowsky, J. S., Shimamura, A. P., Kritchevsky, M., & Squire, L. R. (1989). Cognitive impairment following frontal lobe damage and its relevance to human amnesia. Behavioral neuroscience , 103 (3), 548. Jurado, M. A., Mataro, M., Verger, K., Bartumeus, F., & Junque, C. (2000). Phonemic and semantic fluencies in traumatic brain injury patients with focal frontal lesions. Brain Inj, 14 (9), 789-795. Katzev, M., Tuscher, O., Hennig, J., Weiller, C., & Kaller, C. P. (2013). Revisiting the functional specialization of left inferior frontal gyrus in phonological and semantic fluency: the crucial role of task demands and individual ability. J Neurosci, 33 (18), 7837-7845. Lancaster, J. L., Woldorff, M. G., Parsons, L. M., Liotti, M., Freitas, C. S., Rainey, L., ... & Fox, P. T. (2000). Automated Talairach atlas labels for functional brain mapping. Human brain mapping , 10 (3), 120-131. Lanting, S., Haugrud, N., & Crossley, M. (2009). The effect of age and sex on clustering and switching during speeded verbal fluency tasks. Journal of the International Neuropsychological Society , 15 (2), 196-204. Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychological assessment (5th ed.). New York, NY, US: Oxford University Press. Li, M., Zhang, Y., Song, L., Huang, R., Ding, J., Fang, Y., ... & Han, Z. (2017). Structural connectivity subserving verbal fluency revealed by lesion-behavior mapping in stroke patients. Neuropsychologia , 101 , 85-96. Luckhurst, L., & Lloyd-Jones, T. J. (2001). A selective deficit for living things after temporal lobectomy for relief of epileptic seizures. Brain and Language , 79 (2), 266-296. Maldjian, J. A., Laurienti, P. J., Kraft, R. A., & Burdette, J. H. (2003). An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. Neuroimage , 19 (3), 1233-1239. Martin, M., Beume, L., Kümmerer, D., Schmidt, C. S., Bormann, T., Dressing, A., ... & Kaller, C. P. (2016). Differential roles of ventral and dorsal streams for conceptual and production-related components of tool use in acute stroke patients. Cerebral Cortex , 26 (9), 3754-3771. Martin, R. C., Loring, D. W., Meador, K. J., & Lee, G. P. (1990). The effects of lateralized temporal lobe dysfunction on normal and semantic word fluency. Neuropsychologia , 28 (8), 823-829. Mayr, U. (2002). On the dissociation between clustering and switching in verbal fluency: comment on Troyer, Moscovitch, Winocur, Alexander and Stuss. Neuropsychologia , 40 (5), 562-566. Mayr, U., & Kliegl, R. (2000). Complex semantic processing in old age: Does it stay or does it go?. Psychology and aging , 15 (1), 29. McIntosh, R. D. (2017). Simple dissociations for a higher-powered neuropsychology. PrePrint from: https://psyarxiv.com/mnhct/. Meinzer, M., Flaisch, T., Wilser, L., Eulitz, C., Rockstroh, B., Conway, T., . . . Crosson, B. (2009). Neural signatures of semantic and phonemic fluency in young and old adults. J Cogn Neurosci, 21 (10), 2007-2018.
132 Chapter 4 Second Study – Dissociating neural correlates of verbal fluency
Miller, E. (1984). Verbal fluency as a function of a measure of verbal intelligence and in relation to different types of cerebral pathology. British Journal of Clinical Psychology , 23 (1), 53-57. Milner, B. (1964). Some effects of frontal lobectomy in man. The frontal granular cortex and behavior. , 313-334. Mummery, C. J., Patterson, K., Hodges, J. R., & Wise, R. J. (1996). Generating 'tiger' as an animal name or a word beginning with T: differences in brain activation. Proceedings in the Royal Society B: Biological Sciences , 263(1373), 989-995. Pendleton, M. G., Heaton, R. K., Lehman, R. A., & Hulihan, D. (1982). Diagnostic utility of the Thurstone Word Fluency Test in neuropsychological evaluations. Journal of Clinical and Experimental Neuropsychology , 4(4), 307-317. Perret, E. (1974). The left frontal lobe of man and the suppression of habitual responses in verbal categorical behaviour. Neuropsychologia , 12 (3), 323-330. Rende, B., Ramsberger, G., & Miyake, A. (2002). Commonalities and differences in the working memory components underlying letter and category fluency tasks: a dual-task investigation. Neuropsychology , 16(3), 309-321. Reverberi, C., Laiacona, M., & Capitani, E. (2006). Qualitative features of semantic fluency performance in mesial and lateral frontal patients. Neuropsychologia , 44 (3), 469-478. Robinson, G., Blair, J., & Cipolotti, L. (1998). Dynamic aphasia: an inability to select between competing verbal responses?. Brain: a journal of neurology , 121 (1), 77-89. Robinson, G., Shallice, T., Bozzali, M., & Cipolotti, L. (2012). The differing roles of the frontal cortex in fluency tests. Brain , 135 (7), 2202-2214. Rogers, R. D., Sahakian, B. J., Hodges, J. R., Polkey, C. E., Kennard, C., & Robbins, T. W. (1998). Dissociating executive mechanisms of task control following frontal lobe damage and Parkinson's disease. Brain: a journal of neurology , 121 (5), 815-842. Rorden, C., Karnath, H. O., & Bonilha, L. (2007). Improving lesion-symptom mapping. Journal of Cognitive Neuroscience , 19 (7), 1081-1088. Schmidt, C. S., Schumacher, L. V., Römer, P., Leonhart, R., Beume, L., Martin, M., ... & Kaller, C. P. (2017). Are semantic and phonological fluency based on the same or distinct sets of cognitive processes? Insights from factor analyses in healthy adults and stroke patients. Neuropsychologia, 99, 148-155. Schwartz, S., Baldo, J., Graves, R. E., & Brugger, P. (2003). Pervasive influence of semantics in letter and category fluency: A multidimensional approach. Brain and Language , 87 (3), 400-411. Shallice, T. (1988). From neuropsychology to mental structure . Cambridge University Press. Shao, Z., Janse, E., Visser, K., & Meyer, A. S. (2014). What do verbal fluency tasks measure? Predictors of verbal fluency performance in older adults. Front Psychol, 5 , 772. Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A Compendium of Neuropsychological Tests New York: Oxford University Press. Stuss, D. T., Alexander, M. P., Hamer, L., Palumbo, C., Dempster, R., Binns, M., ... & Izukawa, D. (1998). The effects of focal anterior and posterior brain lesions on verbal fluency. Journal of the International Neuropsychological Society , 4(3), 265-278. Stuss, D. T., Alexander, M. P., Palumbo, C. L., Buckle, L., Sayer, L., & Pogue, J. (1994). Organizational strategies with unilateral or bilateral frontal lobe injury in word learning tasks. Neuropsychology , 8(3), 355.
Chapter 4 Second Study – Dissociating neural correlates of verbal fluency 133
Stuss, D. T., Craik, F. I., Sayer, L., Franchi, D., & Alexander, M. P. (1996). Comparison of older people and patients with frontal lesions: Evidence from word list learning. Psychology and Aging , 11 (3), 387. Stuss, D. T., Toth, J. P., Franchi, D., Alexander, M. P., Tipper, S., & Craik, F. I. (1999). Dissociation of attentional processes in patients with focal frontal and posterior lesions. Neuropsychologia , 37 (9), 1005-1027. Sylvester, C. Y. C., & Shimamura, A. P. (2002). Evidence for intact semantic representations in patients with frontal lobe lesions. Neuropsychology , 16 (2), 197. Szatkowska, I., Grabowska, A., & Szymanska, O. (2000). Phonological and semantic fluencies are mediated by different regions of the prefrontal cortex. Acta Neurobiol Exp (Wars), 60 (4), 503-508. Teuber, H. L. (1955). Physiological psychology. Annual review of psychology , 6(1), 267-296. Thompson-Schill, S. L., D’Esposito, M., Aguirre, G. K., & Farah, M. J. (1997). Role of left inferior prefrontal cortex in retrieval of semantic knowledge: a reevaluation. Proceedings of the National Academy of Sciences , 94 (26), 14792-14797. Thompson-Schill, S. L., Swick, D., Farah, M. J., D’Esposito, M., Kan, I. P., & Knight, R. T. (1998). Verb generation in patients with focal frontal lesions: A neuropsychological test of neuroimaging findings. Proceedings of the National Academy of Sciences, 95 (26), 15855-15860. Tombaugh, T. N., Kozak, J., & Rees, L. (1999). Normative data stratified by age and education for two measures of verbal fluency: FAS and animal naming. Archives of Clinical Neuropsychology , 14 (2), 167-177. Troyer, A. K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two components of verbal fluency: evidence from younger and older healthy adults. Neuropsychology , 11 (1), 138. Troyer, A. K., Moscovitch, M., Winocur, G., Alexander, M. P., & Stuss, D. (1998). Clustering and switching on verbal fluency: the effects of focal frontal- and temporal- lobe lesions. Neuropsychologia, 36 (6), 499-504. Tucha, O., Smely, C., & Lange, K. W. (1999). Verbal and figural fluency in patients with mass lesions of the left or right frontal lobes. Journal of Clinical and Experimental Neuropsychology , 21 (2), 229-236. Wagner, A. D., Paré-Blagoev, E. J., Clark, J., & Poldrack, R. A. (2001). Recovering meaning: left prefrontal cortex guides controlled semantic retrieval. Neuron , 31 (2), 329-338. Wagner, S., Sebastian, A., Lieb, K., Tüscher, O., & Tadi ć, A. (2014). A coordinate-based ALE functional MRI meta-analysis of brain activation during verbal fluency tasks in healthy control subjects. BMC neuroscience , 15 (1), 19.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 135
“Everybody is a genius. But if you judge a fish by its ability to climb a tree, it will live its whole life believing that it is stupid.“ Albert Einstein
Chapter 5
Third Study – Psychometric properties of the verbal fluency task
As outlined in Chapter 2 the psychometric properties and in particular the test-retest reliability of the verbal fluency task have only been sparsely investigated. Furthermore, findings across previous studies seem to be inconsistent. Therefore, the study presented in this chapter aimed at systematically investigating the psychometric properties of a German version of the verbal fluency task used for the neuropsychological assessment of verbal fluency in this thesis. As will be shown, results revealed adequate to high test-score reliability, split-half reliability, good criterion validity, and appropriate test-retest reliability. Furthermore, it will be shown that item difficulty is an important determinant of adequate verbal fluency assessment.3
3 The content of this chapter is in preparation for publication as: Schmidt, C. S., Schumacher, L. V., Römer, P., Nitschke, K., Schumacher, F. K., Musso, M., Weiller, C., Kaller, P. C. (in preparation, March 19th, 2018). Psychometric properties of the verbal fluency task: The impact of item difficulty.
136 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
5.1 Abstract
Objective: Verbal fluency (VF) is a classical test of language abilities and executive functioning and among the most frequently applied assessment tools in clinical neuropsychology. Yet, despite its popularity, its psychometric properties have only been sparsely investigated. Furthermore, existing findings are inconsistent, presumably because different task versions with varying item difficulties and numbers of items were employed across studies.
Methods: We used a German VF task with semantic and phonological items that were further classified as being of an easy or hard difficulty level. Test-score reliability, split-half reliability, and internal consistency were assessed in stroke patients and normal controls.
Furthermore, we investigated the test-retest reliability of semantic and phonological VF in a sample of normal young adults, taking item difficulty into account.
Results: For all samples, indices for test-score reliability ranged from high to very high.
Compared to a subset of normal controls matched for age, sex, and education, deficits emerged for stroke patients. In terms of test-retest reliability, intra-class correlations for the absolute agreement and the relative consistency in the average number of retrieved words per condition ranged from .568 to .708 and from .775 to .849, respectively, indicating that item difficulty is an important determinant for the reliable assessment of VF performance, with highest reliability for moderately difficult items.
Conclusion: The here proposed VF task exhibits good psychometric properties, with indicating better reliability for easy over hard items for both types of verbal fluency. For the clinical assessment of VF performance we propose a short version including both semantic and phonological items.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 137
5.2 Introduction
Verbal fluency tasks (e.g., Benton, 1968; Borkowski, Benton, & Spreen, 1967; Milner, 1964) require the examinee to generate as many words as possible to a given category cue (semantic fluency) or letter cue (phonological fluency) within a preset time interval (e.g. 60 s; Lezak,
Howieson, & Loring, 2004; Strauss, Sherman, & Spreen, 2006). Tests of verbal fluency are frequently employed in neuropsychological assessments of language abilities and executive functioning (Lezak et al., 2004; Strauss et al., 2006) in a wide variety of clinical as well as healthy populations (for reviews, see Abwender, Swan, Bowerman, & Connolly, 2001;
Alvarez & Emory, 2006; Costafreda et al., 2006; Henry & Crawford, 2004; Martin & Fedio,
1983; Metternich, Buschmann, Wagner, Schulze-Bonhage, & Kriston, 2014; Sarkis et al.,
2013). However, despite this apparent popularity of verbal fluency assessments with as of yet more than 4,000 published studies (cf. Schmidt et al., 2017), only very few studies have focused on the tasks’ psychometric properties in general (e.g. Bird et al., 2004; Cohen &
Stanczak, 2000; Tombaugh, Kozak, & Rees, 1999) and their test-score reliability (Ruff et al.,
1996; Schmand et al., 2008) and test-retest reliability (e.g. Lemay, Bedard, Rouleau, &
Tremblay, 2004; Ross et al., 2007) in particular.
More specifically, only two studies exist that addressed the test-score reliability of phonological fluency tasks (Ruff et al., 1996; Schmand et al., 2008; Table 5.1), whereas only one study (Melinder et al., 2005) considered semantic fluency. A reason for this may be that phonological and semantic fluency are most commonly assessed using just one (e.g. animals;
Bird et al., 2004) versus three items (e.g. F, A, S; Ruff et al., 1996), respectively.
Furthermore, as outlined in Table 5.1, extant findings for test-retest reliability of semantic and phonological fluency tasks are highly inconsistent, ranging from low (e.g. Bird et al., 2004) to adequate (e.g. Cohen & Stanczak, 2000) to high indices of stability (e.g. Vlaar & Wade,
2003).
138 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
Table 5.1. Chronological overview of studies reporting on the test-score and test-retest reliability of the verbal fluency task
Study Sample Verbal fluency task Comparison Reliability indices Test-retest reliability Task
Reference Participants N Age Cue Cues Duration Cue Cues Test Metric/ Coefficient Retest Metric/ Coefficient range type Type Type Interval Type (M ± SD)
1 Diesfeldt, Patients with 50 61-91 Sem [Clothing 3min N.A. N.A. N.A. N.A. N.A. 2-3 Pearson .88 1985 dementia years , Fruits] weeks ’s r syndrome (79.8 ± 6.7) Clothing 3min 2-3 Pearson .84 weeks ’s r Fruits 3min 2-3 Pearson .83 weeks ’s r
2 Roberts & Recently 17 30-83 Sem [Animals 60s N.A. N.A. N.A. N.A. N.A. 6-8 Pearson .91 Dorze, aphasic years , weeks ’s r 1994 patients (61.4 ± Clothing, 14.9) Food]
Chronic 16 34-76 Sem [Animals 60s N.A. N.A. N.A. N.A. N.A. 6-8 Pearson .94 aphasic years , weeks ’s r patients (53.3 ± Clothing, 11.3) Food]
3 Ruff et al., HCS 360 16-70 Phon F,A,S 60s N.A. N.A. internal Cronba .83 6 Pearson .74 1996 years consist ch’s α months ’s r ency
4 Chiu et HCS and Sem Animals 60s N.A. CMMS conver Pearson .64 1 week Pearson .67 al., 1997 demented E gent ’s r 's r subjects validity Sem Fruits 30s N.A. CMMS conver Pearson .47 1 week Pearson .45 E gent ’s r ’s r validity Sem Vegetabl 30s N.A. CMMS conver Pearson .44 1 week Pearson .58 es E gent ’s r ’s r
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 139
validity Sem ‘things 30s N.A. CMMS conver Pearson .59 1 week Pearson .68 that E gent ’s r ’s r make validity you happy’ Sem ‘things 30s N.A. CMMS conver Pearson .33 1 week Pearson .31 that E gent ’s r ’s r make validity you sad’ Sem [Animals N.A. CMMS conver Pearson .76 N.A. N.A. N.A. , Fruits, E gent ’s r Vegetabl validity es]
5 Basso, HCS (men 50 20-59 Phon [F-A-S] 60s N.A. N.A. N.A. N.A. N.A. 12 Pearson .80 Bornstein, only) years months ’s r & (32.5 ± Lang,199 9.27) 9
6 Dikmen, HCS 81 15-83 Phon [F-A-S] 60s N.A. N.A. N.A. N.A. N.A. 2.4- Pearson .72 Heaton, years 15.8 's r Grant, & (34.2 ± months Temkin, 16.7 ) (mean 1999 9.1))
7 Tombaug HCS 38 65.6 ± Phon [F-A-S] 60s N.A. N.A. N.A. N.A. N.A. 5.6 Pearson .74 h, Kozak, 9.7 years ’s r & Rees, 1999
8 Cohen & HCS 188 31.06 ± Phon [S, [5min, N.A. N.A. criterio BD vs. Hit rate: 6 Pearson .79 Stanczak, 10.38 (writi C] 4min] n HCS 88.3% weeks ’s r 2000 ng) validity , Sensitivity: predicti 96.6% ve validity Specificity : 37.5%
140 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
Brain- 296 43.34 ± Phon [S, [5min, Phon Word constru Pearson .66 N.A. N.A. N.A. damaged 17.01 (writi C] 4min] Fluenc ct ’s r ng) y Test- validity Form A Phon COWA constru Pearson .72 N.A. N.A. N.A. ct ’s r validity Phon FAS constru Pearson .81 N.A. N.A. N.A. ct ’s r validity Phon Word Not Phon COWA constru Pearson .52 N.A. N.A. N.A. Fluency reported ct ’s r Test- validity Form A Phon FAS constru Pearson .49 N.A. N.A. N.A. ct ’s r validity Phon COWA Not Phon FAS constru Pearson .85 N.A. N.A. N.A. reported ct ’s r validity
9 Harrison, HCS 90 Reported Sem Animals 1.5min N.A. N.A. N.A. N.A. N.A. 1-8 Pearson .68 Buxton, for weeks ’s r Husain, & overall Wise, sample 2000 (N=365): Females 41.3 ± 18.7; Males 40.2 ± 18. Phon B 60s N.A. N.A. N.A. N.A. N.A. 1-8 Pearson .73 weeks ’s r Phon [F-A-S] 60s N.A. N.A. N.A. N.A. N.A. 1-8 Pearson .82 weeks ’s r
10 Vlaar & Patients with 30 35-65 Phon [F-A-S] 60s N.A. N.A. N.A. N.A. N.A. 7-14 Pearson .85 Wade, multiple years days ’s r 2003 sclerosis
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 141
F 60s N.A. N.A. N.A. N.A. N.A. 7-14 Pearson .79 days ’s r A 60s N.A. N.A. N.A. N.A. N.A. 7-14 Pearson .71 days ’s r S 60s N.A. N.A. N.A. N.A. N.A. 7-14 Pearson .80 days ’s r
11 Ross, HCS 14 52-80 Sem Tools 60s N.A. N.A. N.A. N.A. N.A. T1-T2: Pearson .22 2003 years 14 days ’s r (67.35 ± 7.82) N.A. N.A. N.A. N.A. N.A. T1:T3: Pearson .40 28 days ’s r
N.A. N.A. N.A. N.A. N.A. T2:T3: Pearson .01 14 days ’s r
N.A. N.A. N.A. N.A. N.A. T1-T2: ICC .20 14 days (2,1) Phon F 60s N.A. N.A. N.A. N.A. N.A. T1-T2: Pearson .69 14 days ’s r N.A. N.A. N.A. N.A. N.A. T1:T3: Pearson .81 28 days ’s r
N.A. N.A. N.A. N.A. N.A. T2:T3: Pearson .80 14 days ’s r
N.A. N.A. N.A. N.A. N.A. T1-T2: ICC .76 14 days (2,1)
12 Bird, HCS 99 39-75 Sem Animals 60s N.A. N.A. N.A. N.A. N.A. 1 Pearson .56 Papadopo years month ’s r ulou, (57 ± Ricciardel 8.3) li, Rossor, & Cipolotti, 2004
Phon S 60s N.A. N.A. N.A. N.A. N.A. 1 Pearson.63
142 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
month ’s r
13 Melinder HCS 45 37.89 ± Sem [Animals 60s N.A. N.A. internal Cronba .70 1-2 Pearson .85 et al., 9.15 , consist ch’s α days ’s r 2005 Fruits,Bo ency dy Parts] Animals 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .74 ’s r Fruits 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .78 ’s r Body 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .84 parts ’s r Phon [F-A-S] 60s N.A. N.A. internal Cronba .78 1-2days Pearson .81 consist ch’s α ’s r ency F 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .64 ’s r A 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .51 ’s r S 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .85 ’s r [A-S] 60s N.A. N.A. N.A. N.A. N.A. 1-2days Pearson .76 ’s r
14 Woods et HCS 174 18-66 Sem Animals 60s Sem ‘things Conver Pearson .59 4.1- Pearson .73 al., 2005 years that gent ’s r 15.5 ’s r (38.8 ± people and months 11.8) do’ diverge nt validity (constr uct validity ) Phon F,A,S 60s Sem ‘things Conver Pearson .53 that gent ’s r people and do’ diverge nt validity
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 143
(constr uct validity )
15 Schmand HCS 200 17-89 Phon K,O,M 60s N.A. N.A. Internal Cronba .80 et al., years consist ch’s 2008 (53.2 ± ency alpha 17.9) Phon P,G,R 60s N.A. N.A. Internal Cronba .84 consist ch’s ency alpha Phon D,A,T 60s N.A. N.A. Internal Cronba .82 consist ch’s ency alpha
16 Fliessbach HCS 40 16-75 Phon P Not Sem Superm Constru Pearson .45 6-10 Pearson .69 , Hoppe, years (writing) reported arket ct ’s r weeks ’s r Schlegel, (NeuroC items validity (mean: Elger, & ogFX) (writin 8 Helmstaed g) weeks) ter, 2006
Phon P not Phon LPS 6 Constru Pearson .60 6-10 Pearson .69 (writing) reported (writin ct ’s r weeks ’s r (NeuroC g) validity (mean: ogFX) 8 weeks) Patients with 42 Phon P Not Sem Superm Constru Pearson .34 6-10 Pearson .69 epilepsy (writing) reported arket ct ’s r weeks ’s r (NeuroC items validity (mean: ogFX) (writin 8 g) weeks)
Phon LPS 6 Not Constru Pearson .49 (writing) reported ct ’s r validity
17 Ross et HCS 53 21.32 ± Phon [C-F-L] 60s N.A. N.A. N.A. N.A. N.A. Mean ICC .84 al., 2007 3.85 interval (3,1)
144 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
: 44.6 days
18 Schmand, HCS 100 17-89 Phon [D-A-T, 60s N.A. N.A. N.A. N.A. N.A. 1h Pearson .80 Groenink, years K-O-M, (twice ’s r & van den (53.2 ± P-G-R] during Dungen, 17.9) one test 2008 session )
19 Hurks, HCS, 41 Grade 3: Sem Animals 60s N.A. N.A. N.A. N.A. N.A. Twice ICC .49 2012 strategy 9.27 ± during (3,1) instruction 0.51 one test before Grade 4: session second 10.06 ± testing 0.65 Grade 5: 11.03 ± 0.46 Grade 6: 12.29 ± 0.37
HCS, story 40 Grade 3: Sem Animals 60s N.A. N.A. N.A. N.A. N.A. Twice ICC .61 telling 9.48 ± during (3,1) before 0.33 one test second Grade 4: session testing 10.06 ± 0.45 Grade 5: 11.24 ± 0.67 Grade 6: 12.22 ± 0.44
Note. If cue items are listed in squared brackets ([]) the reliability coefficients refer to the average word counts across items. HCS, healthy control sample; M, mean; SD, standard deviation; N.A., not applicable; Sem, semantic; Phon, phonological; CMMSE, Cantonese version of the Mini-Mental State Examination, LPS 6, Leistungsprüfsystem Untertest 6.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 145
Given that different versions and variations of the verbal fluency task with different types of items and different levels of item difficulty have been applied (Table 5.1) diminished measures of test-retest reliability may be caused by items with too low or too high levels of difficulty. That is, as semantic and phonological fluency differ with respect to the kind of search processes required for successful retrieval (Katzev, Tuscher, Hennig, Weiller, &
Kaller, 2013) semantic fluency is usually easier than phonological fluency (Lezak et al.,
2004). Moreover, also the difficulty levels between different semantic categories and between different letters show substantial variation (Katzev et al., 2013; Lezak et al., 2004). Therefore, items may result in potential floor or ceiling effects, leading to a restriction of range of performance scores and a reduced discrimination power.
In addition, the majority of the psychometric studies on verbal fluency have reported inadequate measures of reliability. For instance, although Cronbach’s alpha is the most often reported index for the assessment of internal consistency, it is not valid for the comparison of two halves of a single test or in case of a single test it is only valid if the scale is tau- equivalent (Sijtsma, 2009). If these assumptions are violated, which is often the case, alpha is only a lower bound to reliability that likely underestimates reliability and hence attains values that are outside the range of possible values for the reliability that can be derived from a single test administration (Sijtsma, 2009). Therefore, for an adequate assessment several other indices such as the greatest lower bound (glb; Sijtsma, 2009), lambda 4 ( λ4; Guttman, 1945), and omega total ( ωt; Revelle & Zinbarg, 2009) have been proposed.
Likewise, test-retest reliability of verbal fluency tasks has been commonly assessed in terms of Pearson’s product-moment correlations (e.g. Bird, Papadopoulou, Ricciardelli, Rossor, &
Cipolotti, 2004; Lemay et al., 2004; Roberts & Dorze, 1994), whereas only three studies
(Hurks, 2012; Lemay et al., 2004; Ross et al., 2007) (Table 5.1) reported more appropriate intra-class correlation coefficients (ICC) of relative consistency and absolute agreement (cf.
Shrout & Fleiss, 1979). The use of Pearson’s r in the investigation of test-retest reliability has
146 Chapter 5 Third Study – Psychometric properties of the verbal fluency task been highly discouraged, as it cannot detect systematic errors (Atkinson & Nevill, 1998;
Bates, Zhang, Dufek, & Chen, 1996; Weir, 2005). In addition, it does not provide any information about the absolute agreement between the outcomes of an individual over two repeated assessments (Muller & Buttner, 1994), which is however a prerequisite for reliably assessing a patient’s change over the course of an illness or as a result of a clinical intervention. Thus, in addition to the aforementioned inconsistency in their results, most of the previous psychometric studies on the test-retest reliability of the verbal fluency task do not provide the information that is relevant in the context of clinical evaluations.
Taken together, despite the verbal fluency task’s ubiquity in clinical and experimental neuropsychology, information on its test-score and test-rest reliability is either lacking or inconsistent and in the majority of studies based on inadequate psychometric indices. The present study therefore aimed at a comprehensive investigation of both types of reliability for a task variant comprising 8 semantic and 8 phonological fluency items. In an attempt to resolve the inconsistency between previous studies, we specifically assessed whether item difficulty and potentially resulting restrictions of range may have a particular impact on the task’s reliability. The here applied verbal fluency task therefore comprised a 2 × 2 factorial manipulation of cue type (semantic vs. phonological fluency) and item difficulty (easy vs. hard) (Katzev et al., 2013).
Reliability was not only assessed in an easily accessible young-adult sample of undergraduate students but also in two further and for clinical use more representative samples of normal old adults and chronic stroke patients. To overcome previous limitations, analyses were focused on indices reported to be appropriate for an adequate assessment of the test-score and test- retest reliability of a neuropsychological test, such as, for instance, the glb (Sijtsma, 2009) and
ICCs of relative consistency and absolute agreement (Shrout & Fleiss, 1979), respectively.
Finally, another objective was to propose a clinically useful selection of semantic and phonological items. To this end the criterion-related concurrent validity was calculated for
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 147 stroke patients and matched normal controls to determine items that were best to discriminate between these two groups.
5.3 Methods
5.3.1 Participants
5.3.1.1 Normal Young Adults
The sample of normal young adults consisted of 69 participants (47 female) with a mean age
(±SD) of 23.07 ± 2.03 years (range 19.04-26.48) and a mean education (±SD) of 16.10 ± 2.14 years (range 12-23) and comprised the same sample as described in Schmidt et al. (2017).
Inclusion criteria were an age between 19 and 26 years, right-handedness, normal or corrected-to-normal vision, and German as native language. Exclusion criteria were current or historical psychiatric or neurological illness or use of psychotropic medication. Participants were assessed twice at an interval of one week with the verbal fluency task. Prior to the analyses, individual data were inspected for outliers. In detail, the total number of words produced for the semantic and the phonological fluency and their respective interquartile ranges (IQRs) were separately computed (Tukey, 1977). Out of the initially recruited 75 participants, one participant was excluded after the first session due to signs of depressive symptoms (score of 17) as measured with the Beck Depression Inventory-II (BDI-II; Beck et al., 1996) and another five participants had to be excluded due to outlier performance.
5.3.1.2 Normal Old Adults
The sample of normal older adults comprised 40 participants (24 female) with a mean age
(±SD) of 61.07 ± 7.03 years (range 50.05 - 77.53) and a mean education (±SD) of 16.51 ±
3.84 years (range 10-23). Participants were recruited from advertisements published on the black board of the University Medical Center’s webpage and by word of mouth among community-dwelling citizens living independently in and around Freiburg. Inclusion criteria
148 Chapter 5 Third Study – Psychometric properties of the verbal fluency task were age between 50 and 90 years. Exclusion criteria were current or historical psychiatric or neurological illness or use of psychotropic medication. Participants underwent one assessment with the verbal fluency task. None of the participants had to be excluded due to outlier performance.
5.3.1.3 Stroke Patients
The third sample of the present study comprised 174 chronic stroke patients (61 female) which were recruited from the Dept. of Neurology at the University Medical Center Freiburg
(cf. Schmidt et al., 2017). All patients had a first presentation of an ischemic stroke without a hemorrhagic event. Exclusion criteria at the acute stage were an age over 90 years, inability to tolerate MRI examination or clinical testing due to poor general health status, previous infarcts, previous intracerebral hemorrhage, previous traumatic brain injury, contemporary re- infarct, bilateral infarcts, major cognitive impairment (e.g. dementia), illiteracy, hearing or visual deficits, alcohol abuse, and contraindications for MRI examination such as claustrophobia or cardiac pacemaker. From the initially recruited sample of 189 chronic stroke patients four were excluded because of severe aphasia (i.e. patients were unable to speak), nine were excluded because they either did not complete, or were unable to perform the task (i.e. task abortion at the request of the patient), and another two patients were excluded due to an unusually low educational attainment of less than 8 years (Strauss et al.,
2006; Tombaugh et al., 1999; Schmidt et al., 2017). Each patient underwent one assessment with the verbal fluency task and none had to be excluded due to outlier performance. The study was approved by local ethics authorities and conducted in compliance with the Helsinki
Declaration of the World Medical Association (http://www.wma.net).
The sample had a mean age (±SD) of 64.43 ± 13.73 years (range 22.41-87.50), a mean education (±SD) of 13.24 ± 3.35 years (range 8 - 23), and an average post-stroke duration
(±SD) of 18.33 ± 19.08 months (range 5.03-73.50). A total of 105 patients with left-
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 149 hemisphere and 69 with right-hemisphere strokes were included. The stroke territory concerned in most of these cases (n = 147) was that of the middle cerebral artery.
5.3.2 Verbal Fluency Task
Participants were administered a German version of the verbal fluency task (Katzev et al.,
2013; Schmidt et al., 2017). Briefly, the task comprised a 2×2 factorial combination of semantic cues (categories, e.g. vegetables) and phonological cues (letters, e.g. V) that were further classified as being of an easy or hard difficulty level (cf. Katzev et al., 2013). Four different items were presented for each combination of cue type (semantic vs. phonological) and difficulty level (easy vs. hard), yielding a total of 16 items. Items and presentation order were identical for all participants (easy semantic, hard semantic, easy phonological, hard phonological items) with the following items for the semantic easy condition; vehicles
[German: ‘Transportmittel’], quadrupeds [‘Vierbeiner’], musical instruments
[‘Musikinstrumente’], professions [‘Berufe’]; for the semantic hard condition; fluids
[‘Flüssigkeiten’], toys [‘Spielzeuge’], furniture [‘Möbelstücke’], vegetables [‘Gemüsearten’]; for the phonological easy condition; T, B, S, K; for the phonological hard condition: V, N, D,
F (cf. Katzev et al, 2013; Schmidt et al., 2017). Instructions for the verbal fluency task were given orally by the experimenter (CS, PR). Participants were told that the verbal fluency task would comprise two different parts (semantic and phonological) and that they were to generate as many nouns as possible within a time limit of 60 s following either a category or a letter. Task rules were explained with exemplar items (i.e., category: Lebensmittel [English: food or groceries]; letter: E). For both conditions only words common in German should be said. No words should be produced twice, no proper names, and no words beginning or ending with the same word stem were allowed. In the phonological condition, additional rules implied that also no verbs, adjectives, filler words, or numbers should be said.
150 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
5.3.3 Data Analyses
Analyses of variance were performed using IBM SPSS (version 23; IBM Corp., Armonk,
NY). Analyses of test-score and split-half reliability, internal consistency and test-retest reliability were performed with R (R Core Team. 2014; version 3.1.2) using the package psych (Revelle, 2014; version 1.5.1).
5.3.3.1 Test-Score Reliability, Split-Half Reliability, and Internal Consistency
Reliability was analyzed separately for the three different groups of participants and included the following indices: The greatest lower bound (glb; Bentler and Woodward, 1980) which indicates the greatest estimate of the lower bound to reliability was estimated as the main outcome measure for test-score reliability (Sijtsma, 2009). Furthermore, lambda 4 ( λ4;
Guttman, 1945) was assessed as the maximum estimate of the lower bound of the split-half reliability (Jackson & Agunwamba, 1977) and omega total ( ωt) was computed for the estimation of the total variance of a test (i.e. for the estimation of the test’s internal consistency; Revelle & Zinbarg, 2009). In addition, lambda 2 and 3 ( λ2, λ3; Guttman, 1945), with λ3 corresponding to Cronbach’s alpha ( α; Cronbach, 1951), were assessed for the sake of completeness and comparability with other studies.
The calculation of reliability indices was based on the total number of words produced as outcome measure. For the normal young adults who underwent two repeated testing sessions the indices of reliability were calculated separately as well as jointly for both testing sessions.
5.3.3.2 Test-Retest Effects and Reliability
Test-retest effects of repeated assessment within an one-week interval in the sample of normal young adults, possibly indicating any learning or practice effects between the two test sessions, were assessed by means of an RM-ANOVA with the within-subjects factors session
(test session 1 and 2), cue type (semantic and phonological), and item difficulty (easy and
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 151 hard), and the total number of words produced as dependent variable. Within-subjects effects were Greenhouse-Geisser-corrected for non-sphericity, if necessary. The alpha level was set to .05 and Bonferroni-corrected for all post-hoc pairwise comparisons.
Analysis of test-retest reliability of the total number of words produced as outcome measure further entailed the calculation of the intra-class correlation coefficient (ICC). The ICCs were computed using two-way random effects models (Shrout & Fleiss, 1979; Weir, 2005), once in terms of absolute agreement (assessing the consistency of individual scores) and once in terms of relative consistency of repeated measures (assessing the consistency of rank orders within the group), corresponding to the ICC(2,1) and ICC(3,1) types of Shrout and Fleiss (1979), respectively. In addition, Pearson’s product-moment correlations between sessions 1 and 2 as well as the standard error of measurement (SEM) or typical error and the coefficient of variation (CV) were computed for sake of completeness and to allow for comparability with previous psychometric studies on the test-retest reliability of the verbal fluency task. The
SEM 4 is calculated by dividing the standard deviation of the difference scores by √2
(Hopkins, 2000) and the CV is calculated by dividing the standard deviation of the difference scores by the grand mean over the two measurements (Atkinson & Nevill, 1998; Hopkins,
2000).
5.3.3.3 Analyses of the Influence of Cue Type and Item Difficulty on Reliability
Due to differences in their search strategies, semantic fluency is usually easier than phonological fluency (Lezak et al., 2004). Differences in difficulty level between items for the semantic and items for the phonological fluency condition have also been shown (Katzev et al., 2013; Lezak et al., 2004). Items for the semantic easy and the phonological hard fluency condition may lead to ceiling or floor effects and hence, a restriction of range and poor
4 Note that the standard error of measurement is not identical with the standard error of mean, although the latter is also commonly abbreviated by the acronym SEM.
152 Chapter 5 Third Study – Psychometric properties of the verbal fluency task discrimination power. Therefore, we expected an inverted U-shape distribution for the indices of the test-score and test-retest reliability. That is, we expected superior reliability for (i) hard over easy semantic items and (ii) easy over hard phonological items.
With regard to our hypothesis (see above), we tested for a significant two-way interaction between cue type (semantic vs. phonological) × item difficulty (easy vs. hard) using bootstrapping (Efron, 1979). Through resampling with replacement from the observed data, the bootstrap method provides a means of estimating the accuracy and precision of any estimate of a population parameter (Efron & Tibshirani, 1993; Markiewicz, Reader, &
Matthews, 2014). Permutations under the null hypothesis for the two-way interaction (H0:
[semantic easy – semantic hard] – [phonological easy – phonological hard] = 0) with randomly reassigning the groups of the 2 × 2 factorial design from the initial set of groups with 10000 bootstraps were performed. Results of the bootstrapping algorithm were then compared to the outcome of the null hypothesis for testing for a significant two-way interaction (two-tailed). The same bootstrapping approach as described above was used to test for significant main effects of cue type (semantic vs. phonological) and item difficulty (easy vs. hard). Again permutations under the null hypothesis for the main effect cue type (H0:
[semantic-phonological] = 0) and for the main effect item difficulty (H0: [easy-hard] = 0) with randomly reassigning the groups from the initial set of groups with 10000 bootstraps was performed and results were compared to the outcome of the null hypothesis of the respective main effect (two-tailed).
This was done separately for all reliability indices ( λ2 - λ4, glb, and ωt) of the three samples of participants as well as for both ICC indices (ICC(2,1) and ICC(3,1)) for the sample of normal young adults.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 153
5.3.3.4 Criterion-Related Concurrent Validity
Criterion-related validity refers to the relationship between test scores and a measured criterion. Concurrent validity refers to the relation of test scores to an individual’s standing on a current criterion (Sireci & Sukin, 2013). In the present study, the measured criterion was a group assignment (patient vs. normal control). The criterion-related concurrent validity of the verbal fluency task was hence assessed by using a quasi-experimental design and testing for significant differences in test performance between groups (Sireci & Sukin, 2013). For each normal old adult one stroke patient was matched for age, sex, and education . Matching was performed using an in-house toolbox written in Matlab (The Maths Works Inc., Natick, MA,
USA). That is, Mahalanobis distances of each possible pair of patient and normal control was computed for the matching variables age , sex , and education , and those pairs of stroke patients and normal controls with the least distance were chosen (Kaller et al., 2014;
Köstering et al., 2016).
A repeated-measures analyses of variance (RM-ANOVA) was then computed with the within- subject factors cue type (semantic and phonological) and item difficulty (easy and hard), group
(patients vs. normal control) as between subjects-factor, the matching factors age, sex, and education as covariates, and the total number of words produced as dependent variable.
5.4 Results
Descriptive statistics for the total number of words produced as the main outcome variable are reported in Tables 5.2 and 5.3 for the two testing sessions of the normal young adults and the single testing session in the samples of stroke patients and normal old adults, respectively.
The terminology for the classification of reliability outcomes in the present paper follows
Strauss et al. (2006; p. 13f.) with low reliability indicated by r < .6, marginal reliability by .6
≤ r < .7, adequate reliability by .7 ≤ r < .8, high reliability for .8 ≤ r < .9, and very high reliability by r ≥ .9.
154 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
Table 5.2. Descriptive Statistics of Performance on the Verbal Fluency Task for Session 1 and Session 2 in Normal Young Adults
Session1 Session 2 Difference (Session 2-Session1) Semantic Phonological Semantic Phonological Semantic Phonological Easy Hard Easy Hard Easy Hard Easy Hard Easy Hard Easy Hard
Mean 15.47 11.19 13.35 9.23 18.04 12.61 15.32 10.68 2.57 1.41 1.98 1.55 (SD) (2.61) (2.13) (2.92) (2.25) (2.72) (2.32) (2.85) (2.14) (1.48) (1.12) (1.92) (1.45) Min 9.25 7.5 7.75 5 11.5 7.25 9.5 6.5 -.75 -1.5 -2.5 -2.5 Max 21.5 16.75 19.5 15.5 22.75 17.5 22.25 15.75 6.5 4.75 6 5.25
Note. SD, standard deviation; Min, minimum; Max, maximum.
Table 5.3. Descriptive Statistics of Performance on the Verbal Fluency Task for Normal Old Adults and Stroke Patients
HC Stroke Semantic Phonological Semantic Phonological Easy Hard Easy Hard Easy Hard Easy Hard
Mean 15.03 10.81 12.34 8.86 10.16 7.61 7.93 5.42 (SD) (2.75) (2.16) (2.95) (2.66) (3.62) (2.7) (3.81) (2.83) Min 10.75 5 7.5 4.75 1.25 2 .75 .25 Max 21.25 16.5 20.75 15.25 21.5 14.25 18.25 14
Note. HC, healthy control; SD, standard deviation; Min, minimum; Max, maximum.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 155
5.4.1 Test Score Reliability, Split-Half Reliability, and Internal Consistency
The results for the indices of reliability and internal consistency can be found in Table 4 for all three samples. Indices are given separately for the four cells of the 2 × 2 factorial design of the verbal fluency task (semantic vs. phonological by easy vs. hard) and for all items of the semantic and the phonological fluency condition (Table 5.4).
The glb as the greatest estimate of the lower bound to test-score reliability ranged from .741 to .936 in the normal young adults, from .740 to .947 in the normal old adults, and from .813 to .961 in the stroke patients. Hence, for all samples and independent of the type of verbal fluency task (semantic vs. phonological and easy vs. hard) test-score reliability was found to be adequate to very high.
The split-half reliability in terms of λ4 ranged from .692 to .934 in the normal young adults, from .744 to .943 in the normal old adults, and from .814 to .960 in the stroke patients, hence indicating adequate to very high split-half reliability across all samples and types of verbal fluency task.
The internal consistency as measured with ωt ranged from .751 to .923 in the normal young adults, from .746 to .928 in the normal old adults, and from .815 to .961 in the stroke patients, again indicating adequate to very high estimates.
The values for the additional indices (i.e. λ2 and λ3) computed here ranged from .711 to .923
(see Table 5.4).
On a descriptive level it needs to be noted that the indices for the test-score reliability, split- half reliability, and the internal consistency where highest for the total of 8 semantic and 8 phonological items. Within the cells of the factorial design the easy items were superior to the hard items for both types of verbal fluency and for almost all indices computed for the three samples.
156 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
Table 5.4. Reliabaility Indices and Bootstrapping Results
Reliability Group Indices Semantic Phonological Statistics (p-value) Main Effect Interaction Effect All Easy Hard All Easy Hard Cue Type Item Difficulty
Test-score NC Young λ 2 .825 .716 .689 .864 .824 .746 .238 .003 .332 reliability Session 1 λ 3 ( α) .822 .711 .683 .860 .823 .742 .280 .002 .355 λ 4 .877 .764 .751 .908 .874 .783 .220 < .001 .112 ωt .876 .766 .751 .900 .874 .785 .352 .004 .152 glb .881 .762 .747 .923 .874 .778 .065 < .001 .088
NC Young λ 2 .836 .719 .677 .863 .795 .683 .336 .028 .168 Session 2 λ 3 ( α) .833 .716 .669 .861 .794 .677 .342 .028 .195 λ 4 .893 .742 .692 .891 .806 .756 .941 .007 .990 ωt .882 .748 N.E. .899 .807 .757 .457 .035 .102 glb .900 .741 .687 .900 .810 .757 .996 .002 .985
NC Young λ 2 .866 .772 .744 .904 .871 .797 .096 .005 .286 Session 1 and 2 λ 3 ( α) .864 .767 .739 .903 .870 .794 .099 .007 .273 λ 4 .909 .823 .811 .934 .905 .825 .190 .004 .106 ωt .901 .824 .812 .923 .905 .826 .242 .055 .237 glb .914 .815 .808 .936 .906 .820 .220 < .001 .043
NC Old λ 2 .852 .797 .724 .901 .836 .835 .153 .963 .278 λ 3 ( α) .848 .794 .722 .898 .835 .830 .173 .905 .319 λ 4 .913 .826 .744 .943 .853 .857 .209 .951 .180 ωt .894 .831 .746 .928 .853 .873 .165 .825 .262 glb .916 .835 .740 .947 .854 .878 .132 .721 .055
Stroke patients λ 2 .921 .904 .804 .953 .923 .904 < .001 < .001 < .001 λ 3 ( α) .920 .903 .803 .952 .923 .904 < .001 < .001 < .001 λ 4 .934 .909 .814 .960 .925 .912 < .001 .004 < .001 ωt .941 .909 .815 .961 .925 .912 .007 .002 .674 glb .943 .907 .813 .961 .926 .915 .012 .003 < .001
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 157
Test-retest HC Young ICC(2,1) .644 .568 .708 .674 .651 .638 .602 .234 .028 reliability ICC(3,1) .893 .831 .849 .857 .801 .775 .249 .459 .311 Pearson’s r .896 .832 .852 .857 .801 .776 N.A. N.A. N.A. SEM 5.989 1.096 .865 7.105 1.288 1.044 N.A. N.A. N.A. CV .037 .046 .051 .052 .064 .074 N.A. N.A. N.A.
Note. λ2-4 = split-half reliability indices according to Guttman (1945); α = Cronbach’s (1951) alpha, which is equivalent to Guttman’s λ3; ωt = total omega (McDonald, 1999; Revelle & Zinbarg, 2008); glb = greatest lower bound (Bentler & Woodward, 1980); SEM, standard error of measurement; CV, coefficient of variation; N.E., not estimable; N.A., not applicable.
158 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
5.4.2 Test-Retest Effects and Reliability
5.4.2.1 Repeated-Measures Analysis of Variance
Difference scores were computed by subtracting each participant’s individual score at session
1 from the individual score at session 2. Difference scores reflected a general improvement in terms of an increased total number of words produced for all combinations of the factorial design (Table 5.2).
An RM-ANOVA on the average number of words revealed significant main effects for
2 2 session (F 1,68 = 263.977, p < .001, ƞp = .795), cue type (F 1,68 = 67.917, p < .001, ƞp = .500),
2 and item difficulty (F 1,68 = 868.626, p < .001, ƞp = .927). Participants hence produced significantly more words at session 2 (estimated marginal mean ± standard error, 14.165 ±
.250) when compared to session 1 (12.311 ± .234). Furthermore, participants also produced significantly more words in the semantic (14.328 ± .267) when compared to the phonological condition (12.148 ± .274), and, finally, participants generated more words for easy items
(15.547 ± .279) in comparison to hard items (10.928 ± .214).
The RM-ANOVA further revealed a significant two-way interaction of session × item
2 difficulty (F 1,68 = 29.153, p < .001, ƞp = .300), a strong trend for the two-way interaction of
2 cue type × item difficulty (F 1,68 = 3.563, p = .063, ƞp = .050) and a strong trend for the three-
2 way interaction (F 1,68 = 3.736, p = .057, ƞp = .052). The two-way interaction of cue type × session was not significant (p = .131). For the two-way interaction of session × item difficulty , pairwise comparisons yielded a significant difference between easy and hard items for both testing sessions. However, this difference was slightly larger at session 2 (5.040 ± .162) in comparison to session 1 (4.197 ± .187). For the two-way interaction of cue type × item difficulty , on a descriptive level there was a difference between easy and hard items for both semantic and phonological fluency and this difference was slightly larger for the semantic condition (4.851 ± .192) in comparison to the phonological condition (4.386 ± .206).
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 159
For the three-way interaction, on a descriptive level the difference between session 2 versus session 1 was notably smaller for semantic hard (1.413 ± .268) when compared to semantic easy (2.573 ± .321) items. The difference between session 1 and session 2 for phonological easy items (1.979 ± .348) was slightly larger to phonological hard (1.393 ± .265) items (see
Table 5.2). Therefore, the total number of words produced remained relatively constant for semantic hard and all phonological items but showed a clear improvement for semantic easy items.
5.4.2.2 Analyses of Test-Retest Reliability
Intra-class correlation coefficients of absolute agreement, ICC(2,1), and relative consistency,
ICC(3,1), are separately reported in Table 5.4 for the for cells of the factorial design and for all semantic and all phonological items. Concurring with expectations, results revealed superior absolute and relative test-retest reliability for hard over easy semantic items and, vice versa, for easy over hard phonological items. The coefficients for the absolute agreement were however reduced in comparisons with the coefficients for relative consistency, which is in line with the significant improvement across sessions (see above). These differences between absolute and relative ICCs were particularly pronounced for semantic easy and phonological hard items, again highlighting the dependency of adequate test-retest reliability on appropriate levels of item difficulty.
For comparison with other studies (Table 5.1), additional measures such as Pearson’s r, CV, and SEM are reported in Table 5.4, again indicating sufficient test-retest reliability of the version of the verbal fluency task applied here. Furthermore, ICCs for the individual semantic and phonological items are reported in Supplementary Table S5.1, ranging from r = .430 to
.700 and r = .455 to .794 for coefficients of absolute agreement and relative consistency, respectively.
160 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
5.4.3 Effects of Item Difficulty and Cue Type
5.4.3.1 Performance Scores
The effect of cue type (semantic vs. phonological) and item difficulty (easy vs. hard) on the mean number of words produced are pres ented in Figure 5. 1A for both sessions combined for the sample of normal young adults, in Figure 5. 1B for normal old adults and Figure 5.1C for stroke patients. All samples produced most words in the semantic easy and least words in the phonological hard f luency condition. All samples produced more words in the semantic as compared to the phonological fluency condition and for the easy as compared to the hard items. This is in line with the significant main effects of cue type and item difficulty reported in section 5.4 .2.1. Furthermore, stroke patients produced fewer words as compared to both samples of normal controls irrespective of the type of verbal fluency, which is in line with the significant main effect of group reported in section 5.4.4.
Figure 5.1. Effect of cue type (semantic vs. phonological) and item difficulty (easy vs. hard) for (A) both sessions combined for the sample of normal young adults (n = 69), (B) the sample of normal old adults (n = 40), and (C) the sample of chronic stroke patie nts (n = 174). Error bars denote the standard error of the mean.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 161
5.4.3.2 Reliability Estimates
Results for all indices of the bootstrapping method to test for significant main effects of cue type and item difficulty as well as to test for a significant two-way interaction of cue type × item difficulty are reported in Table 5.4. In the following the results for the glb and the ICC indices will be presented.
For the glb, using 10000 bootstraps, the main effect of cue type was significant for the stroke patients (p = .012) and showed a trend for session 1 of normal young adults (p = .065). For all other samples the main effect of cue type was not significant (smallest p = .132) (Table 5.4).
The main effect of item difficulty reached significance for session 1 (p < .001), session 2 (p =
.002), for both sessions combined (p < .001) of the normal young controls, and for the stroke patients (p = .003). For the sample of normal old controls the main effect of item difficulty was not significant (p = .721) (Table 5.4). Except for the sample of normal old controls, all other indices reported in Table 5.4 concordantly show a significant main effect of item difficulty .
Results further revealed a significant two-way interaction of cue type × item difficulty for the sample of stroke patients (p < .001) and for both sessions combined for the normal young adults (p = .043). There was a strong trend for session 1 of the normal young adults (p = .088) and the normal old adults (.055), whereas the interaction was non-significant for session 2 of the normal young adults (p = .985) (Table 5.4).
The following results were obtained for the ICC indices: Solely the two-way interaction of cue type × item difficulty for the ICC (2,1) (p = .028) reached significance. All other effects were not significant (smallest p = .234) (Table 5.4).
162 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
5.4.4 Criterion-Related Concurrent Validity
5.4.4.1 Comparability of Matched Pairs
By using an in-house toolbox the matching procedure revealed 40 matched pairs of stroke patients and normal controls . In line with the high pairwise correlation of age, found during the matching procedure (see Fig. 5.2A; r = .905), the analysis of variance (ANOVA) for age with group (patients vs. normal controls) as between -subjects factor was not significant (F 1,79
= . 278, p =.599). The same applied for the matching variable of education (see Fig. 5.2B, r =
.925). Again, the ANOVA for education with group as between-subjects factor was not significant (F 1,79 = .335, p =.564). For the matching variable sex there were 38 equally matched pairs (m:m, N=16; f:f, N=22) and two pairs with a female normal c ontrol and a male patient. The pairwise difference in sex between matched pairs was not significant (Pearson
Exact Chi-Square test, χ2 = .205, p = .821). Thus, patients and normal controls were highly comparable regarding the variables age , sex , and education .
Figure 5.2. Pairwise correlation for the matching variable (A) age and (B) education.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 163
5.4.4.2 Group Differences in Verbal Fluency Performance
A RM-ANOVA on the average number of words revealed a significant main effect for group
2 2 (F 1,75 = 30.757, p < .001, ƞp = .291), cue type (F 1,75 = 6.027, p = .016, ƞp = .074), education
2 2 (F 1,75 = 17.839, p < .001, ƞp = .192), and sex (F 1,75 = 7.621, p = .007, ƞp = .092). Hence, normal controls produced significantly more words (estimated marginal mean ± standard error, 11.642 ± .362) as compared to the stroke patients (8.797 ± .362), both groups produced significantly more words in the semantic (11.245 ± .255) as compared to the phonological condition (9.194 ± .307), the average number of words increased with increasing level of education, and women produced slightly more words as compared to men. The RM-ANOVA further revealed a significant two-way interaction of group × item difficulty (F 1,75 = 10.829, p
2 2 = .002, ƞp = .126), cue type × age (F 1,75 = 4.481, p = .038, ƞp = .056), item difficulty ×
2 education (F 1,75 = 7.883, p = .006, ƞp = .095), and a significant three-way interaction of cue
2 type × item difficulty × sex (F 1,75 = 6.914, p = .010, ƞ = .084). All other interaction effects were not significant (smallest p = .174). For the two-way interaction group × item difficulty , pairwise comparisons showed that normal controls produced significantly more words as compared to stroke patients for both easy and hard items. However, this difference between patients and normal controls was slightly larger for the easy (4.637 ± .275) as compared to the hard (3.317 ± .191) items. For the two-way interaction cue type × age , the difference between semantic and phonological fluency (with more words produced in the semantic condition) decreased with increasing age. For the two-way interaction item difficulty × education , the difference between easy and hard items increased with higher level of education and was slightly larger for the easy as compared to the hard items. For the three-way interaction cue type × item difficulty × sex, participants said more words in the semantic as compared to the phonological condition as well as more words in the easy as compared to the hard items. The difference between easy and hard items was identical for both sexes, the difference between the semantic and phonological fluency condition was slightly larger for female as compared to
164 Chapter 5 Third Study – Psychometric properties of the verbal fluency task male participants. By revealing significant differences between stroke patients and their well- matched normal controls, results indicate adequate criterion-related concurrent validity.
Furthermore, differentiation between patients and normal controls was best for easy items.
5.5 Discussion
The present study investigated the psychometric properties of a German version of the verbal fluency task proposed by Katzev et al. (2013). For all three groups of participants investigated here indices for the test-score reliability, split-half reliability, and internal consistency for the
2×2 factorial manipulation of cue type (semantic vs. phonological fluency) and item difficulty
(easy vs. hard) indicated adequate to very high reliability (see Table 5.4). These findings are in line with previous studies that also found high reliability in terms of Cronbach’s alpha
(Schmand et al., 2008) as well as good construct and criterion validity (Cohen & Stanczak,
2000) for letter based fluency tasks. Furthermore, the present study expanded previous findings by showing adequate to high reliability not only for phonological but also for semantic verbal fluency tasks. Contrary to expectations and collectively for all samples, reliability indices for the easy items were superior to the hard items, irrespective of cue type.
A reason that we did not find the expected pattern (hard > easy for the semantic items and easy > hard for the phonological items) is that the significant two-way interaction of group × item difficulty indicated that a better discrimination between normal controls and stroke patients is based on the easy items. As reliability is an indicator for the discriminating power of a task (i.e. the ability to distinguish between two groups measured on the same task)
(Chapman & Chapman, 1973; Melinder et al., 2005) a better reliability for the easy over hard items corroborate these findings. Furthermore, findings are in line with the significant main effect of item difficulty found consistently for all reliability indices.
The influence of item difficulty on the test-retest reliability of the verbal fluency task as a potential cause for the inconsistency of previous findings was systematically investigated. To
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 165 this end, particular focus was placed on the absolute agreement between individual scores over repeated test application. Results revealed higher test-retest reliability coefficients in terms of the absolute agreement for the semantic hard over semantic easy items and for the phonological easy over the phonological hard items, respectively (Table 5.4), indicating adequate test-retest reliability for the semantic hard and a marginal test-retest reliability for the phonological easy items, respectively. This suggests that achieving sufficiently high test- retest reliability in the verbal fluency task depends substantially on the utilization of appropriate, i.e. moderate, degrees of item difficulty. The importance of adequate item difficulty is further corroborated by the distribution of performance scores across individual subjects of the group of normal young adults. Box plots in Figure 5.3 illustrate floor effects and a considerable restriction of range particularly for phonological hard items, corresponding to the lowest test-retest reliability of the four verbal fluency conditions.
166 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
Semantic Fluency, Easy Condition Semantic Fluency, Hard Condition ABat Sessions 1 and 2 at Sessions 1 and 2 25 25
20 20
15 15
10 10
5 5 Mean Performance (#) Session 2 Mean Performance (#) Session 2
0 0 0 5 10 15 20 25 0 5 10 15 20 25 Mean Performance (#) Session 1 Mean Performance (#) Session 1
CDPhonological Fluency, Easy Condition Phonological Fluency, Hard Condition at Sessions 1 and 2 at Sessions 1 and 2 25 25
20 20
15 15
10 10
5 5 Mean Performance (#) Session 2 Mean Performance (#) Session 2
0 0 0 5 10 15 20 25 0 5 10 15 20 25 Mean Performance (#) Session 1 Mean Performance (#) Session 1
Figure 5.3. Bivariate product-moment correlations (Pearson’s r, see also Table 5.4) between session 1 and 2 for (A) semantic fluency, easy condition, r = .832, (B) semantic fluency, hard condition, r = .852, (C) phonological fluency, easy condition, r = .801, and (D) phonological fluency, hard condition, r = .776. Dots denote performance scores of individual subjects; lines represent linear regressions and corresponding non-simultaneous 95% confidence bands. Distributions are illustrated by box plots, where the line within the box indicates the median, left or lower border of the box represent 25% and the right and upper borde represent 75% of the distribution for the horizontal and vertical boxplot, respectively. The whiskers extending vertically and horizontally from the boxplot represent the range between the maximum and minimum scores.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 167
Previous studies also found marginal to high test-retest reliabilities for semantic and phonological fluency, but reliability coefficients were based on measures of relative consistency (e.g. Harrison 2000, Ross 2007) and not on absolute agreement. However, relative consistency coefficients tend to overestimate reliability, because only the rank order of scores within the sample as a whole is considered. That is, as effects of repeated assessment on individual scores are not considered, reliability coefficients based on consistency measures are not sufficient for normative assessments (Calamia, Markon, &
Tranel, 2013; Duff, 2012). In this regard, effects of repeated test administrations in the present sample resulted in a general improvement in terms of an increased number of nouns produced from session 1 to session 2 for all four conditions. Furthermore, these differences in individual scores over repeated test administrations were subject to item difficulty as reflected in the strong trend for a significant three-way interaction. Considering that for the analysis of the test-score reliability (Table 5.4) the semantic easy are superior to the semantic hard items the question remains why this is not the case for the analysis of the test-retest reliability. All participants started with the semantic easy condition and due the tasks novelty a proper word retrieval strategy was not established at the start of the test. In the course of the test participants learned a strategy that was applied in the later trials. At the repeated testing session this strategy could be adopted from the start (i.e. also for the semantic easy items), which is in line with the largest improvement for the semantic easy items in session 2. This clearly pronounced improvement for the semantic easy items may have led to diminished test- retest reliability for these items.
As an exception (cf. Table 5.1), Lemay and colleagues (2004) reported reliability coefficients in terms of the absolute agreement, but only used one item each for semantic (‘tools’) and phonological (‘F’) fluency, yielding highly divergent test-retest reliabilities of r=.20 and r=.76, respectively. This underlines another important finding of the present study, namely that the test-retest reliability of the verbal fluency task significantly benefits from aggregation
168 Chapter 5 Third Study – Psychometric properties of the verbal fluency task across several items. Administration of only one item may lead to bias and an under- or overestimation of individual ability in verbal fluency due to effects of personal experience, interests, and life style. For instance, the semantic categories ‘tools’ and ‘clothing’ may be differentially favorable for male and female subjects. As another example, the category
‘vegetables’ was found to be an easy item in the 1980s (Mannhaupt, 1983) and turned into a hard item during testing a comparable sample of German students three decades later (Katzev et al., 2013). Moreover, categories may also be somewhat ill-defined resulting in individually different interpretations of subclasses included in a given category. As an example, some people tend to name only construction-type tools (hammer, saw), whereas others also include kitchen-related (spoon, knife) and garden-related (shovel, rake) tools (Lemay et al. 2004).
These and other potential biases resulting from individual items are expected to be considerably attenuated by administration of multiple items.
Nonetheless, many of the previous studies (e.g., Bird et al., 2004; Fliessbach et al., 2006;
Harrison, Buxton, Husain, & Wise, 2000; Hurks, 2012; Lemay et al., 2004; Woods, 2005;
Table 1) used only one single item to repeatedly assess inter-individual differences in semantic and phonological fluency for evaluating test-retest reliability. The resulting inconsistency of reported findings (see Table 5.1) is mirrored by the highly variable test-retest reliability measures of the individual semantic and phonological items in the present study that were in many cases low and distributed over a wide range (Supplementary Materials,
Table S5.1).
In addition to the number and difficulty of items, the duration of the retest interval and the characteristics of the investigated sample also contribute to differences in the resulting test- retest reliability (Calamia et al., 2013; Duff, 2012; Heaton et al., 2001). For example, carry- over effects due to the repeated measurement are more likely the shorter the retest interval length (Duff, 2012), while longer intervals do not necessarily suppress these effects
(Beglinger et al., 2005; Duff, 2012). As normal young adults were tested over a short period
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 169 of time (a one week retest interval) results of the present study are not necessarily generalizable to clinical populations and/or longer test-retest intervals (Beglinger et al., 2005;
Calamia et al., 2013; Duff, 2012; Heaton et al., 2001; McCaffrey, Ortega, & Haase, 1993).
A significant difference between patients and normal controls indicated good criterion-related concurrent validity. Therefore, even in a well matched but rather small sample (40 patients vs.
174 patients of the previous analysis) reliable differences between patients and controls could be found, revealing robust results with disease related differences of verbal fluency.
Furthermore, a significant two-way interaction of group × item difficulty indicated that the difference between patients and normal controls was larger for the easy as compared to the hard items. This might be due to a more pronounced decline of other cognitive processes required for both types of verbal fluency tasks such as processing speed, memory, or language skills (Cumming et al., 2013). Furthermore, it has been noted that slowed information processing as well as executive dysfunctions predominate in stroke patients and performance differences are often due to the timed component of cognitive tasks (Cumming et al., 2013).
Since it may not be practicable in clinical settings to administer such an extensive verbal fluency task as proposed by Katzev et al. (2013) which includes 16 items, we would like to propose an item selection based on our findings. Before we address this topic, it needs to be noted that indices for all semantic and all phonological items were superior when compared to the subdivision into easy and hard items (Table 5.4). A significant two-way interaction of group × item difficulty indicated that the easy items for both the semantic and phonological fluency differentiated best between patients and normal controls. Furthermore, indices for the test-score reliability for the easy were superior to the hard items. However, results of the test- retest reliability indicate that moderately difficult items (semantic hard, phonological easy) exhibited better reliability. At first, we would therefore suggest to use the 4 easy items of the phonological condition. For the semantic fluency condition we would also suggest to use the 4 easy items as the sample of normal young adults is not particularly representative for clinical
170 Chapter 5 Third Study – Psychometric properties of the verbal fluency task use. Furthermore, diminished measures of test-retest reliability for the semantic easy items might have been caused by the non-counterbalanced presentation order of semantic and phonological cues.
Current results have to be interpreted with caution as some limitations are present. As has been stated above, a first limitation constituted that the order in which the semantic and phonological cues were presented was not counterbalanced (e.g. order of semantic and phonological cues remained the same for each participant and session). In addition, although the level of item difficulty chosen in the current study was applicable to the present sample of normal adults it might not be applicable for various clinical populations as (objective and subjective) difficulty levels of items might change with respect to the affected brain areas of a particular clinical population. A difference in performance between normal controls and stroke patients was also present in our study. That is, patients performed worse as compared to normal controls on all verbal fluency conditions, but showed a similar distribution for the mean number of words produced for the four cells of the factorial design, with most words produced in the semantic easy and least words produced in the phonological hard fluency condition (see Fig. 5.1). While phonological fluency is generally sensitive to frontal brain damage (e.g., Milner, 1964; Benton, 1968; Reverberi et al., 2006; Robinson et al., 2012), detecting impairments in semantic fluency in frontal patients depends on the task-related demands on cognitive control that are exerted by the administered semantic categories (Baldo et al., 2006; Katzev et al., 2013). Results were obtained in a German speaking sample and allocation of hard and easy letters may not be generalizable to other languages.
Taken together, this is the first study extensively investigating the verbal fluency task’s reliability by taking both semantic and phonological items as well as item difficulty into account. Results indicate that a German version of the verbal fluency task proposed by Katzev et al. (2013) exhibits good test-score reliability, criterion-related construct validity, and test- retest reliability. When investigating verbal fluency it is of great importance to consider item
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 171 difficulty and test-retest reliability clearly benefits from aggregation across multiple items.
Future studies should aim at investigating the psychometric properties of the verbal fluency task in other clinical populations. It would be desirable to collect norm data for this task and further validate the clinical utility of the proposed selection of items in normal adults and clinical populations.
5.6 Acknowledgement
The authors report no conflicts of interest. The present research was supported by grants of the
BrainLinks-BrainTools Cluster of Excellence (projects #128 to C.P.K. and #36 to C.W., C.P.K., and
M.M.) funded by the German Research Foundation (DFG; grant # EXC 1086). C.S.M.S. and L.V.S. received scholarship funds from the State Law on Graduate Funding of the University of Freiburg,
Germany.
172 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
5.7 References
Abwender, D. A., Swan, J. G., Bowerman, J. T., & Connolly, S. W. (2001). Qualitative analysis of verbal fluency output: review and comparison of several scoring methods. Assessment, 8 (3), 323-338. Alvarez, J. A., & Emory, E. (2006). Executive function and the frontal lobes: a meta-analytic review. Neuropsychol Rev, 16 (1), 17-42. Atkinson, G., & Nevill, A. M. (1998). Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med, 26 (4), 217-238. Baldo, J. V., Schwartz, S., Wilkins, D., & Dronkers, N. F. (2006). Role of frontal versus temporal cortex in verbal fluency as revealed by voxel-based lesion symptom mapping. J Int Neuropsychol Soc, 12 (6), 896-900. Basso, M. R., Bornstein, R. A., & Lang, J. M. (1999). Practice Effects on Commonly Used Measures of Executive Function Across Twelve Months. The Clinical Neuropsychologist, 13 (3), 283-292. Bates, B. T., Zhang, S., Dufek, J. S., & Chen, F. C. (1996). The effects of sample size and variability on the correlation coefficient. Medicine and Science in Sports and Exercise , 28(3), 386-391. Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory-II. San Antonio , 78 (2), 490-8. Beglinger, L. J., Gaydos, B., Tangphao-Daniels, O., Duff, K., Kareken, D. A., Crawford, J., . . . Siemers, E. R. (2005). Practice effects and the use of alternate forms in serial neuropsychological testing. Arch Clin Neuropsychol, 20 (4), 517-529. Bentler, P. M., & Woodward, J. A. (1980). Inequalities among lower bounds to reliability: With applications to test construction and factor analysis. Psychometrika , 45 (2), 249- 267. Benton, A. L. (1968). Differential behavioral effects in frontal lobe disease. Neuropsychologia, 6 (1), 53-60. Bird, C. M., Papadopoulou, K., Ricciardelli, P., Rossor, M. N., & Cipolotti, L. (2004). Monitoring cognitive changes: psychometric properties of six cognitive tests. Br J Clin Psychol, 43 (Pt 2), 197-210. Borkowski, J. G., Benton, A. L., & Spreen, O. (1967). Word fluency and brain damage. Neuropsychologia, 5 (2), 135-140. Calamia, M., Markon, K., & Tranel, D. (2013). The robust reliability of neuropsychological measures: meta-analyses of test-retest correlations. Clin Neuropsychol, 27 (7), 1077- 1105. Chapman, L. J., & Chapman, J. P. (1973). Problems in the measurement of cognitive deficits. Psychological Bulletin , 79 (6), 380. Chiu, H. F., Chan, C. K., Lam, L. C., Ng, K. O., Li, S. W., Wong, M., & Chan, W. F. (1997). The modified Fuld Verbal Fluency Test: a validation study in Hong Kong. J Gerontol B Psychol Sci Soc Sci, 52 (5), P247-250. Cohen, M. J., & Stanczak, D. E. (2000). On the reliability, validity, and cognitive structure of the Thurstone Word Fluency Test. Arch Clin Neuropsychol, 15 (3), 267-279.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 173
Costafreda, S. G., Fu, C. H., Lee, L., Everitt, B., Brammer, M. J., & David, A. S. (2006). A systematic review and quantitative appraisal of fMRI studies of verbal fluency: role of the left inferior frontal gyrus. Hum Brain Mapp, 27 (10), 799-810. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. psychometrika , 16 (3), 297-334. Cumming, T. B., Marshall, R. S., & Lazar, R. M. (2013). Stroke, cognitive deficits, and rehabilitation: still an incomplete picture. International Journal of Stroke , 8(1), 38-45. Diesfeldt, H. F. (1985). [Test-retest reliability of category-bound word production in psychogeriatrics]. Tijdschrift voor Gerontologie en Geriatrie, 16 (1), 17-20. Dikmen, S. S., Heaton, R. K., Grant, I., & Temkin, N. R. (1999). Test–retest reliability and practice effects of Expanded Halstead–Reitan Neuropsychological Test Battery. Journal of the International Neuropsychological Society, 5 (04), 346-356. Duff, K. (2012). Evidence-based indicators of neuropsychological change in the individual patient: relevant concepts and methods. Archives of Clinical Neuropsychology , 27(3), 248-261. Duff, K. (2012). Evidence-based indicators of neuropsychological change in the individual patient: relevant concepts and methods. Arch Clin Neuropsychol, 27 (3), 248-261. Efron, B. (1979). Bootstrap methods: another look at the jackknife. Annals of Statistics 7, 1- 26. Efron, B. and Tibshirani, R.J. (1993). An Introduction to the Bootstrap, Chapman & Hall, New York. Fliessbach, K., Hoppe, C., Schlegel, U., Elger, C. E., & Helmstaedter, C. (2006). [NeuroCogFX--a computer-based neuropsychological assessment battery for the follow-up examination of neurological patients]. Fortschr Neurol Psychiatr, 74 (11), 643-650. Gruenewald, P. J., & Lockhead, G. R. (1980). The free recall of category examples. Journal of Experimental Psychology: Human Learning and Memory, 6 (3), 225-240. Guttman, L. (1945). A basis for analyzing test-retest reliability. Psychometrika , 10 (4), 255- 282. Harrison, J. E., Buxton, P., Husain, M., & Wise, R. (2000). Short test of semantic and phonological fluency: normal performance, validity and test-retest reliability. Br J Clin Psychol, 39 ( Pt 2) , 181-191. Heaton, R. K., Temkin, N., Dikmen, S., Avitable, N., Taylor, M. J., Marcotte, T. D., & Grant, I. (2001). Detecting change: A comparison of three neuropsychological methods, using normal and clinical samples. Arch Clin Neuropsychol, 16 (1), 75-91. Henry, J. D., & Crawford, J. R. (2004). A meta-analytic review of verbal fluency performance following focal cortical lesions. Neuropsychology, 18 (2), 284-295. Hopkins, W. G. (2000). Measures of reliability in sports medicine and science. Sports Med, 30 (1), 1-15. Hurks, P. P. (2012). Does instruction in semantic clustering and switching enhance verbal fluency in children? Clin Neuropsychol, 26 (6), 1019-1037. Katzev, M., Tuscher, O., Hennig, J., Weiller, C., & Kaller, C. P. (2013). Revisiting the functional specialization of left inferior frontal gyrus in phonological and semantic fluency: the crucial role of task demands and individual ability. J Neurosci, 33 (18), 7837-7845.
174 Chapter 5 Third Study – Psychometric properties of the verbal fluency task
Köstering, L., Nitschke, K., Schumacher, F. K., Weiller, C., & Kaller, C. P. (2015). Test- retest reliability of the Tower of London Planning Task (TOL-F). Psychol Assess, 27 (3), 925-931. Köstering, L., Schmidt, C. S., Egger, K., Amtage, F., Peter, J., Kloppel, S., . . . Kaller, C. P. (2015). Assessment of planning performance in clinical samples: Reliability and validity of the Tower of London task (TOL-F). Neuropsychologia, 75 , 646-655. Lemay, S., Bedard, M. A., Rouleau, I., & Tremblay, P. L. (2004). Practice effect and test- retest reliability of attentional and executive tests in middle-aged to elderly subjects. Clin Neuropsychol, 18 (2), 284-302. Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsychological Assessment . New York: Oxford University Press. Mannhaupt, H.-R. (1983). Produktionsnormen für verbale Reaktionen zu 40 geläufigen Kategorien. [German category norms for verbal items in 40 categories.]. Sprache & Kognition, 2 (4), 264-278. Markiewicz, P. J., Reader, A. J., & Matthews, J. C. (2014). Assessment of bootstrap resampling performance for PET data. Physics in Medicine & Biology , 60 (1), 279. Martin, A., & Fedio, P. (1983). Word production and comprehension in Alzheimer's diseáse: The breakdown of semantic knowledge. Brain and Language, 19 (1), 124-141. McCaffrey, R. J., Ortega, A., & Haase, R. F. (1993). Effects of repeated neuropsychological assessments. Arch Clin Neuropsychol, 8 (6), 519-524. Melinder, M. R., Barch, D. M., Heydebrand, G., & Csernansky, J. G. (2005). Easier tasks can have better discriminating power: the case of verbal fluency. Journal of Abnormal Psychology , 114 (3), 385. Metternich, B., Buschmann, F., Wagner, K., Schulze-Bonhage, A., & Kriston, L. (2014). Verbal Fluency in Focal Epilepsy: A Systematic Review and Meta-analysis. Neuropsychol Rev . Milner, B. (1964). Some effects of frontal lobectomy in man. The frontal granular cortex and behavior , 313-334. Muller, R., & Buttner, P. (1994). A critical discussion of intraclass correlation coefficients. Statistics in Medicine , 13(23-24), 2465-2476. Mummery, C. J., Patterson, K., Hodges, J. R., & Wise, R. J. (1996). Generating 'tiger' as an animal name or a word beginning with T: differences in brain activation. Proc Biol Sci, 263 (1373), 989-995. Rende, B., Ramsberger, G., & Miyake, A. (2002). Commonalities and differences in the working memory components underlying letter and category fluency tasks: a dual-task investigation. Neuropsychology, 16 (3), 309-321. Revelle, W. (2014). psych: Procedures for psychological, psychometric, and personality research. Northwestern University, Evanston, Illinois , 165 . Revelle, W., & Zinbarg, R. E. (2009). Coefficients alpha, beta, omega, and the glb: Comments on Sijtsma. Psychometrika , 74 (1), 145-154. Reverberi, C., Laiacona, M., & Capitani, E. (2006). Qualitative features of semantic fluency performance in mesial and lateral frontal patients. Neuropsychologia, 44 (3), 469-478. Roberts, P., & Dorze, G. L. (1994). Semantic verbal fluency in aphasia: A quantitative and qualitative study in test-retest conditions. Aphasiology , 8(6), 569-582.
Chapter 5 Third Study – Psychometric properties of the verbal fluency task 175
Robinson, G., Shallice, T., Bozzali, M., & Cipolotti, L. (2012). The differing roles of the frontal cortex in fluency tests. Brain, 135 (7), 2202-2214. Rosen, V. M., & Engle, R. W. (1997). The role of working memory capacity in retrieval. J Exp Psychol Gen, 126 (3), 211-227. Ross, T. P. (2003). The reliability of cluster and switch scores for the Controlled Oral Word Association Test. Arch Clin Neuropsychol, 18 (2), 153-164. Ross, T. P., Calhoun, E., Cox, T., Wenner, C., Kono, W., & Pleasant, M. (2007). The reliability and validity of qualitative scores for the Controlled Oral Word Association Test. Arch Clin Neuropsychol, 22 (4), 475-488. Sarkis, R. A., Busch, R. M., Floden, D., Chapin, J. S., Kalman Kenney, C., Jehi, L., . . . Najm, I. (2013). Predictors of decline in verbal fluency after frontal lobe epilepsy surgery. Epilepsy Behav, 27 (2), 326-329. Schmand, B., Groenink, S. C., & van den Dungen, M. (2008). [Letter fluency: psychometric properties and Dutch normative data]. Tijdschr Gerontol Geriatr, 39 (2), 64-76. Schmidt, C. S., Schumacher, L. V., Römer, P., Leonhart, R., Beume, L., Martin, M., Dressing, A., Weiller, C. & Kaller, C. P. (2017). Are semantic and phonological fluency based on the same or distinct sets of cognitive processes? Insights from factor analyses in healthy adults and stroke patients. Neuropsychologia , 99 , 148-155. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: uses in assessing rater reliability. Psychol Bull, 86 (2), 420-428. Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A Compendium of Neuropsychological Tests New York: Oxford University Press. Thompson-Schill, S. L., Swick, D., Farah, M. J., D’Esposito, M., Kan, I. P., & Knight, R. T. (1998). Verb generation in patients with focal frontal lesions: A neuropsychological test of neuroimaging findings. Proceedings of the National Academy of Sciences, 95 (26), 15855-15860. Tombaugh, T. N., Kozak, J., & Rees, L. (1999). Normative Data Stratified by Age and Education for Two Measures of Verbal Fluency: FAS and Animal Naming. Archives of Clinical Neuropsychology, 14 (2), 167-177. Troyer, A. K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two components of verbal fluency: Evidence from younger and older healthy adults. Neuropsychology, 11 (1), 138-146. Tukey, J. W. (1977). Exploratory data analysis . Reading, MA: Addison-Wesley. Vlaar, A. M., & Wade, D. T. (2003). Verbal fluency assessment of patients with multiple sclerosis: test-retest and inter-observer reliability. Clinical Rehabilitation , 17(7), 756- 764. Weir, J. P. (2005). Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res, 19 (1), 231-240. Woods, S. P., Scott, J. C., Sires, D. A., Grant, I., Heaton, R. K., Tröster, A. I., & Group, T. H. N. R. C. (2005). Action (verb) fluency: Test–retest reliability, normative standards, and construct validity. Journal of the International Neuropsychological Society, 11 (04), 408-415.
Chapter 6 General Discussion 177
“It always seems impossible until it’s done.” Nelson Mandela
Chapter 6
General Discussion
6.1 Summary of key findings
The main objective of this thesis was to investigate the underlying cognitive components and neural correlates of semantic and phonological verbal fluency in a large sample of healthy controls and stroke patients. Specifically, a German version of the verbal fluency task with a large number of semantic and phonological items that could be further classified as being of an easy or hard difficulty level was used. Open questions from the existing literature (see
Chapter 2) were addressed in three different studies. Their objectives and results will be summarized in the following.
Previous behavioral studies showed that different search strategies for word retrieval are engaged in semantic and phonological fluency (Chapter 2). Researchers have therefore postulated that semantic and phonological verbal fluency are subserved by distinct cognitive processes. Factor analyses, in which semantic and phonological fluency loaded on one factor, challenge these findings of distinct cognitive processes (Chapter 2 and 3). The aim of the first study was to resolve these inconsistencies between studies. By using a large item set of exclusively semantic and phonological fluency, results from factor analyses of the first study revealed a two-factor solution in both healthy controls and stroke patients. Confirmatory factor analyses confirm these results of a two-factor solution. In addition to these clearly distinct factors, the semantic and phonological factor also revealed a substantial portion of common variance. Results of the first study demonstrate that semantic and phonological fluency are based on clearly distinct but also common cognitive processes (Chapter 3;
Schmidt et al., 2017).
178 Chapter 6 General Discussion
In line with the proposition of distinct cognitive processes, lesion and functional neuroimaging studies suggest that semantic and phonological fluency are differentially subserved by distinct brain areas in left temporal and left frontal lobes, respectively (Milner,
1964; Baldo et al., 2010; Robinson et al., 2012). Although previous studies have described a strong association between semantic fluency and temporal brain areas and a strong association between phonological fluency and frontal brain areas, the implied double dissociation had not been confirmed in statistical terms thus far (Chapter 2; Baldo et al., 2006; 2010). The main objective of the second study was to confirm the often implied double dissociation in statistical terms. Both a region-based and a more specific voxel-wise approach were used to test for a significant interaction between type of verbal fluency (semantic vs. phonological) and lesion location (temporal vs. frontal) (Chapter 4). Results indeed revealed the aforementioned double dissociation. By using the more precise voxel-wise approach regions in frontal cortex that indicated worse performance for patients with a lesion, irrespective of the type of verbal fluency, were also found (Chapter 4). These regions were clearly distinct from those regions where the interaction (cue type × lesion location) was established. These findings of both shared and specific brain areas being involved for semantic and phonological fluency are in line with findings from the first study of this thesis (Chapter 3), indicating both distinct and common cognitive processes.
The third study investigated the psychometric properties of a German version of the verbal fluency task. Since in previous studies, different versions and variants with a differing number of items that were not controlled for differences in difficulty between items have been used, the impact of item difficulty was particularly examined (Chapter 5). Results of the third study indicate that easy items exhibit better reliability than hard items for both the semantic and phonological verbal fluency. Moreover, reliability clearly benefits from the aggregation across multiple items (Chapter 5).
Chapter 6 General Discussion 179
Now that the results of the three studies of this thesis have been summarized, the focus of the following section will be on the aspect of item difficulty. Limitations will be addressed and recommendations for improvements and further research will be given.
6.2 Item Difficulty
Most of the previous behavioral, lesion and functional neuroimaging studies have only investigated one of the two verbal fluency variants and used only few items (Chapter 2;
Tables 2.1 and 2.2). Moreover, results of previous studies are based on qualitative analyses that only describe associations between performance scores and the investigated brain areas
(Chapter 4). In order to establish the suggested double dissociation, quantitative measures are required and the samples need to be assessed on both semantic and phonological fluency
(Chapter 4). As semantic fluency is generally easier than phonological fluency (Lezak et al.,
2012) differential contributions of frontal and temporal brain areas may only reflect a difference in task difficulty (Chapter 4; Shallice, 1988). As previous studies did not control for differences in item difficulty, this resource artifact may prevent establishing the suggested double dissociation (Shallice, 1988; Davis, 2010). By using quantitative analyses as well as semantic and phonological items with a comparable level of item difficulty these limitations of previous studies were overcome (Chapter 5).
In previous studies that investigated the underlying cognitive processes, semantic and phonological fluency were compared together with other cognitive constructs (Chapter 3).
Since it is assumed that semantic and phonological fluency also engage common cognitive processes (e.g. working memory; Rosen & Engle, 1997; Troyer et al., 1998; Robinson et al.,
2012), a simultaneous investigation with other constructs may bias a direct comparison of semantic and phonological fluency (Chapter 3). Apart from that, again only few items for the semantic and phonological fluency that were not controlled for differences in item difficulty were used. That one category or letter is not the same as the other is especially relevant in the
180 Chapter 6 General Discussion clinical context when selecting appropriate items for repeated clinical assessments (Price et al., 2012) and caution is needed when using different forms of a test interchangeably (Barry,
Bates, & Labouvie, 2008). The educational background or gender may introduce a personal bias for some categories (Capitani, Laiacona, & Barbarotto, 1999; Loonstra, Tarlow, &
Sellers, 2001; Kim et al., 2011). Using several items with comparable levels of difficulty may hence overcome these differences. Furthermore, in the investigation of patients with neurodegenerative diseases (e.g. Parkinson’s disease) this might be especially important as individual semantic fluency tasks are differentially sensitive to the cognitive deficits in these patients (Azuma et al., 1997).
Besides item difficulty, differences in discrimination power have also been found for semantic and phonological fluency (Melinder et al., 2005). For establishing the magnitude or presence of differential cognitive deficits, considering both the impact of item difficulty and discrimination power should be the aim of future studies (Melinder et al., 2005).
The consideration of item difficulty, which has been specifically addressed in the studies of this thesis, is indispensable for future studies. Lesion and functional neuroimaging studies should further be complemented by using stimulation methods that can temporarily disrupt a specific network or elicit a specific lesion in the brain (Ziemann, 2010). Using semantic and phonological decision tasks Devlin and colleagues stimulated those brain areas in left inferior prefrontal cortex (LIPC) with TMS that were active during these tasks as investigated with a preceding fMRI study. They showed that a transient disruption by rTMS within anterior LIPC selectively inferred with semantic but not perceptual processing (Devlin et al., 2003).
However, Devlin and colleagues (2003) did not investigate the differential impact of rTMS on semantic vs. phonological decisions.
Findings of the three studies will be discussed with regard to their clinical applications.
Chapter 6 General Discussion 181
6.3 Clinical application
As stroke is associated with an increased risk of cognitive decline (Tang et al., 2018) findings from the three studies may aid with establishing therapeutic interventions in stroke rehabilitation. One approach would be to establish special exercises to aid the use of appropriate search strategies for the two fluency sub-tasks. This is especially important as deficits in letter fluency are correlated with the self-rating of having greater difficulties in everyday life (Martyr et al., 2012). Unless cognitive impairments after stroke have progressed to dementia, cognitive deficits are often neglected in the follow-up examination of stroke survivors (Tang et al., 2018). Subjective word-finding problems are further associated with reduced objective performance on verbal fluency tasks and a reduced engagement in social activities in patients with Alzheimer’s disease (Farrell et al., 2014).
Another issue relevant for the clinical assessment of verbal fluency is the task duration. Most commonly a time limit of 60s is chosen for each letter or category cue (Lezak et al., 2012).
However, this standard administration time of 60s may be too long and bothersome for individuals with a limited language performance such as aphasic patients. In their sample of aphasic patients, 60% stopped responding within the initial 30s (Kim et al., 2011).
Furthermore, sensitivity and specificity values for differentiating aphasic patients from non- aphasic patients were equivalent or even better for the 30s when compared to the 60s time limit (Kim et al., 2011). In our sample of stroke patients some did not complete or were unable to perform the task (i.e. task abortion at the request of the patient). This might have been due to the long administration time for each cue, which might have dejected some of the patients. Using shorter administration times it might have been possible to increase the sample of chronic stroke patients. However, Kim and colleagues (2011) only used one semantic category (animals) to assess verbal fluency and hence, results should be considered with caution as they may not be applicable for other semantic categories or phonological letters.
182 Chapter 6 General Discussion
For the differentiation between controls and patients with mild cognitive impairment or
Alzheimer’s disease a longer administration time was more accurate (Cunje, Molloy,
Standish, & Lewis, 2007). Here too, only one semantic category was used.
On the other hand, Harrison and colleagues (2000) found that no particular advantage was obtained from having a longer administration time of 90s. Future studies should not only establish an appropriate item selection for the investigation of verbal fluency but also establish appropriate administration times for different clinical populations.
Transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS) have shown promising results in stroke rehabilitation to improve cognition (Andrade et al.,
2015; Ehlis et al., 2016). Knowledge of the neural correlates subserving semantic and phonological fluency may aid with finding appropriate target locations for individual neurostimulation after stroke. Single-pulse TMS over Wernicke’s area has been shown to improve picture naming, indicating that lexical processes are facilitated by neurostimulation
(Mottaghy et al., 2006). Hence, a combination of individually targeted neurostimulation as well as neuropsychological therapy may promote rehabilitation after stroke.
6.4 Limitations and future directions
There are several limitations that will be addressed in the following. First of all, although a large patient group was investigated, lesion overlay was very sparse. More specific hypothesis such as the impact of item difficulty on the involvement of LIFG could hence not be tested.
This would have been desirable, as a coarse division of temporal and frontal brain areas subserving semantic and phonological verbal fluency, respectively, is an oversimplification
(Chapter 4). Nonetheless, by controlling for item difficulty the implied double dissociation between semantic and phonological verbal fluency and temporal and frontal brain areas could be confirmed. Since in previous lesion and neuroimaging studies, only one or two semantic and phonological items were used that also differed in item difficulty (Joanette & Goulet,
Chapter 6 General Discussion 183
1986; Robinson et al., 2012; Biesbroek et al., 2016) the rational for the task selection in verbal fluency should be considered in more detail in future studies.
Second, since neuropsychological assessment was only performed in patients in whom testing was feasible, the patient sample is biased towards the exclusion of patients with more severe stroke deficits (i.e. patients with larger lesions). The investigation of patients with larger lesions and/or greater deficits and also other clinical populations would be desirable in future studies. Third, items of the verbal fluency task were assessed in a German speaking sample and therefore, allocation of item difficulty might be different in other languages. This is particularly important as a precise differentiation between brain-damaged patients and healthy controls with different levels of intelligence also depends on item difficulty (Borkowski,
1967). Fourth, the main outcome variable assessed here was the number of correct words produced in the given time limit. Although scoring productivity in terms of the number of correct words is clinically relevant and useful, qualitative aspects of word production can offer additional insights into the cognitive processes and neural correlates that guide performance on verbal fluency tasks (Troyer et al., 1997; Reverberi et al., 2006; Ledoux et al.,
2014). This is particularly important as people usually tend to produce strings of words that share some underlying features, such as the same first sound, membership in a semantic subcategory (Bousfield & Sedgewick, 1944; Gruenewald & Lockhead, 1980; Ledoux et al.,
2014) or words with a strong associative relatedness based on the co-occurrence in everyday life (Ledoux et al., 2014). This is also an important aspect as clustering and switching are differentially correlated with the number of produced words in semantic and phonological fluency (Troyer et al., 1997). The best discrimination indices between patients with focal frontal- and temporal-lobe lesions were phonemic-fluency switching and semantic-fluency clustering, respectively (Troyer et al., 1998). This further suggests that clustering and switching are dissociable components and hence, future studies should aim at investigating the qualitative aspects of word production in verbal fluency tasks as well.
184 Chapter 6 General Discussion
With regard to the study investigating the psychometric properties of a German version of the verbal fluency task, it needs to be noted that the indices for reliability and internal consistency each reflect the properties for the verbal fluency task for the examined sample (Bortz &
Döring, 2006). For other samples these values should be re-examined and possibly normative data should be obtained to strengthen its use also in daily clinical neuropsychological assessments. As previous studies have found that age, sex, and level of education seem to play a role in the examination of verbal fluency (Strauss, Sherman, & Spreen, 2006), normative data are especially important. For example, older people usually tend to have a greater vocabulary and most often a greater linguistic experience as compared to younger people
(Baciu et al., 2016). The same holds for people with a better and longer education (Strauss,
Sherman & Spreen, 2006). Lastly, even though most of the previous studies have primarily found left hemisphere dominance for verbal fluency tasks, an involvement of the right hemisphere has also been found, particularly for non-verbal (e.g. design fluency) fluency tasks (Jones-Gotman & Milner, 1977; Baldo et al., 2006). Significant activations in the right hemisphere have also been reported for a phonological fluency task (Gourovitch et al., 2000).
As the sample of the present lesion-behavior study (Chapter 4) included exclusively patients with lesions of the left hemisphere, no prediction about the involvement of the right hemisphere can be made (Baldo et al., 2006). As distinct patterns of qualitative aspects of word retrieval in left and right frontal lobe patients have been found (Schwartz & Baldo,
2001), future studies should aim at investigating patients with both right and left hemisphere lesions.
6.5 Conclusion
In conclusion, the data and results presented in this thesis highlight the importance of using both semantic and phonological items and considering item difficulty for the investigation of the underlying cognitive processes and neural correlates in word production as measured with
Chapter 6 General Discussion 185 the verbal fluency task. A major advantage of the results of the presented studies is based on large sample sizes as well as the use of appropriate quantitative analyses to study interaction effects in lesion data. Still, for an appropriate use in daily neuropsychological assessments our understanding of the differences in the two verbal fluency variants needs further confirmation in other patient populations. In addition, future studies should aim at integrating lesion-based analyses with other neuroimaging modalities, such as fMRI or fNIRS, to further assess the complex interplay between different brain areas in the temporal and frontal lobes required for a successful word retrieval in the verbal fluency task. Nonetheless, future studies should aim at integrating these findings specifically for the development of therapeutic interventions in stroke rehabilitation. This is particularly important as in the neuroscience literature speech production is commonly equated with word production (e.g. words in response to other words, verb generation, words to semantic category or letter, names of depicted objects)
(Indefrey & Levelt, 2000) and communication in everyday life is largely based on speech.
Appendix Supplementary Material Chapter 3 187
Appendix
Supplementary Material Chapter 3
Supplementary Analyses : Confirmatory Factor Analysis (CFA) in Stroke Patients without
Aphasia.
To rule out any bias caused by patients with aphasia the confirmatory factor analysis (CFA) with model 1 and model 2 was repeated in a subsample of n=134 stroke patients without signs of aphasia in the acute phase (cf. 3.2). For model 1, all semantic and phonological items had significant loadings on their corresponding factor ranging from .599 to .904. The correlation between the two factors was .870, indicating an overlap of about 75.7% shared variance. The indices for goodness of fit (GFI) and for adjusted goodness of fit (AGFI) were well above the acceptable level of .85 (Table S1). Other model fit indices such as χ2 and χ2/df, the comparative fit index (CFI), and the root-mean-square error of approximation (RMSEA) also suggested a very good fit of the data to the model.
In order to test whether the two-factor model was superior to a one-factor solution, the correlation between factor 1 and factor 2 was set to 1 to constrain the two factors to be exactly the same. For model 2, the GFI and the AGFI were markedly lower as compared to model 1
(Table S1). Furthermore, all other model fit indices (i.e., χ2, χ2/df, CFI, and RMSEA) suggested that the null hypothesis of a good model fit to the data should be rejected. The comparison between both models further indicated by a change in χ2(1) of 75.767 (p < .001) that the two-factor solution in model 1 was significantly better than the one-factor solution in model 2. Taken together, the CFA confirmed that - despite their overlapping variance - semantic and phonological verbal fluency also measure distinct sets of cognitive processes.
188 Appendix Supplementary Material Chapter 3
Supplementary Tables
Supplementary Table S3.1. Items of the verbal fluency task
VF Type Difficulty Englisch German
Semantic easy vehicles Transportmittel quadrupeds Vierbeiner musical instruments Musikinstrumente
professions Berufe hard fluids Flüssigkeiten toys Spielzeuge
furniture Möbelstücke vegetables Gemüsearten
Phonological easy T T B B
S S K K hard V V
N N D D
F F
Supplementary Table S3.2. Results of the confirmatory factor analysis (CFA) for the clinical sample of stroke patients without aphasia
χ² DF p- χ²/DF TLI CFI GFI AGFI RMSEA rho2 value
Model 1 107.145 103 .370 1.040 .997 .997 .908 .879 .017
Model 2 182.912 104 <.001 1.759 .941 .949 .817 .760 .076
Note: DF = degrees of freedom; TLI = Tucker-Lewis index; CFI = comparative fit index; GFI = goodness-of-fit index; AGFFI = adjusted goodness-of-fit index; RMSEA = root mean square error of approximation.
Appendix Supplementary Material Chapter 3 189
Supplementary Figures
Supplementary Figure S3.1. Scree plot of the exploratory factor analysis in healthy young adults indicating that two factors should be retained.
Supplementary Figure S3.2. Output of the two-factor model for the CFA with all semantic items loading of factor 1 and all phonological items loading on factor 2. The two-headed arrow between factor 1 and factor 2 indicates that the two factors are correlated.
190 Appendix Supplementary Material Chapter 3
ABSemantic items Phonological items Factor 1 Factor 1 1 Factor 2 1 Factor 2
0.8 0.8
0.6 0.6
0.4 0.4 Factor Loading Factor Loading
0.2 0.2
0 0
vehiclesquadrupedsmusicalprofessions instruments vegetables fluids toys furniture T B S K V N D F
Supplementary Figure S3.3. Factor solutions of the exploratory factor analysis in stroke patients illustrating the pattern matrix (A) for the semantic items and (B) for the phonological items.
Appendix Supplementary Material Chapter 4 191
Supplementary Material Chapter 4
Supplementary Analysis
Patients with left parietal lesions as stroke controls
The multi-factorial extension of the non-parametric Brunner-Munzel rank-order test with the between-subjects factor lesion location (left temporal vs. left frontal vs. left parietal), the within-subjects factor cue type (semantic vs. phonological), and the total number of words produced as dependent variable revealed a significant two-way interaction of cue type × lesion location (F 2,52 = 5.541, p = .004) whereas the main effects of lesion location (F 2,52 =
2.299, p = .105) and cue type (F 2,52 = 1.173, p = .279) failed to reach significance.
For the significant interaction effect as illustrated in Supplementary Figure S4.2A, we first tested whether the formulated contrasts fulfilled the requirements for a classical double dissociation (cf. main manuscript): Patients with a left temporal lesion (red) as compared to patients with a left parietal lesion (gray) produced significantly fewer words in the semantic fluency condition (mean rank score ± standard error of the mean, M RS ± SEM, , 48.54 ± 8.14 vs. 66.35 ± 5.61, p = .048) but showed no significant differences in the phonological fluency condition (M RS ± SEM, 55.81 ± 9.67 vs. 67.50 ± 8.56, p = .118). The opposite pattern emerged for the patients with a left frontal lesion (blue) who – when compared to patients with a left parietal lesion (gray) - produced significantly fewer words in the phonological fluency condition (M RS ± SEM, 37.13 ± 6.61 vs. 67.50 ± 8.56, p = .006), but showed no significant difference in the semantic fluency condition (M RS ± SEM, 56.79 ± 5.06 vs. 66.35 ±
5.61, p = .139). In addition, we further computed the formulated contrasts for the within- patient comparison (cf. main manuscript) of the two fluency types separately for the left frontal and the left temporal patients: Patients with a left temporal lesion (red) showed a strong trend for producing fewer words in the semantic as compared to the phonological fluency condition (M RS ± SEM, 48.54 ± 8.14 vs. 55.81 ± 9.67, p = .054), whereas patients
192 Appendix Supplementary Material Chapter 4 with a left frontal lesion (blue) produced significantly fewer words in the phonological as compared to the semantic fluency condition (M RS ± SEM, 37.13 ± 6.61 vs. 56.79 ± 5.06, p <
.001). For patients with a left parietal lesion (gray), no significant difference was found between the semantic compared to the phonological fluency condition (M RS ± SEM, 66.35 ±
5.61 vs. 67.50 ± 8.56, p = .292) (see Supplementary Fig. S4.2A).
Thus, by using only patients with a left parietal lesion as stroke controls, results are essentially the same as compared to the results with the whole group of stroke controls. Again, results fully conform to the requirements for establishing a double dissociation both in the classical sense as well as based on significant within-patient comparisons. Lesions to the left temporal and left frontal lobe hence differentially affect performance in semantic and phonological fluency, respectively (Supplementary Fig. S4.2A). Notably, this double dissociation cannot be attributed to potential differences in task difficulty between the two types of verbal fluency as these were effectively controlled for (Supplementary Fig. S4.2A, left parietal).
Patients with left subcortical lesions as stroke controls
The multi-factorial extension of the non-parametric Brunner-Munzel rank-order test with the between-subjects factor lesion location (left temporal vs. left frontal vs. left subcortical), the within-subjects factor cue type (semantic vs. phonological), and the total number of words produced as dependent variable revealed a trend for the main effect of cue type (F 2,66 = 2.495, p = .055) and a significant two-way interaction of cue type × lesion location (F 2,66 = 5.996, p =
.003). The main effect of lesion location (F 2,66 = 1.335, p = .262) did not reach significance.
For the significant interaction effect as illustrated in Supplementary Figure S4.2B, we first tested whether the formulated contrasts fulfilled the requirements for a classical double dissociation (cf. main manuscript): Patients with a left temporal lesion (red) as compared to patients with left subcortical lesions (gray) showed a strong trend for producing fewer words in the semantic fluency condition (mean rank score ± standard error of the mean, M RS ± SEM,
Appendix Supplementary Material Chapter 4 193
, 60.96 ± 10.39 vs. 81.34 ± 5.57, p = .064) but showed no significant differences in the phonological fluency condition (M RS ± SEM, 70.69 ± 12.35 vs. 73.08 ± 7.91, p = .403). The opposite pattern emerged for the patients with a left frontal lesion (blue) who – when compared to patients with left subcortical lesions (gray) - produced significantly fewer words in the phonological fluency condition (M RS ± SEM, 46.10 ± 8.41 vs. 73.08 ± 7.91, p = .013), but showed no significant difference in the semantic fluency condition (M RS ± SEM, 71.29 ±
6.57 vs. 81.34 ± 5.57, p = .139). In addition, we further computed the formulated contrasts for the within-patient comparison (cf. main manuscript) of the two fluency types separately for the left frontal and the left temporal patients: Patients with a left temporal lesion (red) showed a strong trend for producing fewer words in the semantic as compared to the phonological fluency condition (M RS ± SEM, 60.96 ± 10.39 vs. 70.69 ± 12.35, p = .054), whereas patients with a left frontal lesion (blue) produced significantly fewer words in the phonological as compared to the semantic fluency condition (M RS ± SEM, 46.10 ± 8.41 vs. 71.29 ± 6.57, p <
.001). For patients with left subcortical lesions (gray), no significant difference was found between the semantic compared to the phonological fluency condition (M RS ± SEM, 81.34 ±
5.57 vs. 73.08 ± 7.91, p = .274) (see Supplementary Fig. S4.2B).
Thus, by using only patients with a left subcortical lesion as stroke controls, results are essentially the same as compared to the results with the whole group of stroke controls.
Again, results fully conform to the requirements for establishing a double dissociation both in the classical sense as well as based on significant within-patient comparisons. Lesions to the left temporal and left frontal lobe hence differentially affect performance in semantic and phonological fluency, respectively (Supplementary Fig. S4.2B). Notably, this double dissociation cannot be attributed to potential differences in task difficulty between the two types of verbal fluency as these were effectively controlled for (Supplementary Fig. S4.2B, left subcortical).
194 Appendix Supplementary Material Chapter 4
Supplementary Figures
Supplementary Figure S4. 1. Lesion overlays for the subsample of patients with lesions mainly (A) in left parietal cortex (n = 17, maximum overlap = 8) and (B) in left subcortical regions (n = 31, maximum overlap = 19) .
Supplementary Figure S4. 2. Illustrations of the double dissociation based on the significant two-way interaction cue type × lesion location for patients with lesions mainly (A) in left parietal cortex and (B) in left subcortical regions as stroke co ntrols. Mean rank scores +/ - SEM.
Appendix Supplementary Material Chapter 5 195
Supplementary Material Chapter 5
Supplementary Table S5.1. Descriptive statistics and test-retest reliability for individual items in healthy young adults
Descriptive statistics Absolute Agreement Relative Consistency Other Indices
Item Cue Type Item Mean SD Min Max ICC (2,1) CI ICC (3,1) CI Pearson’s r SEM CV Difficulty
Semantic Easy vehicles 14.848 3.021 8.0 20.5 .499 [.023, .743] .652 [.493, .770] .653 1.960 .093
quadruped 18.362 3.685 9.5 28.0 .486 [-.048, .754] .681 [.531, .790] .682 2.269 .087
musical 16.971 3.375 11.0 26.0 .558 [.186, .756] .657 [.498, .772] .666 2.173 .091 instruments professions 16.833 3.213 8.5 22.5 .624 [.282, .795] .706 [.564; .807] .706 1.888 .079
Hard fluids 10.739 2.751 4.5 21.0 .500 [.216, .688] .571 [.387, .710] .573 2.035 .134
toys 10.587 2.958 5.5 19.5 .700 [.263, .859] .794 [.686, .867] .795 1.419 .095
furniture 11.572 2.632 6.0 19.0 .617 [.401, .795] .657 [.449, .773] .665 1.692 .097
vegetables 14.717 3.088 9.0 26.0 .611 [.402, .752] .648 [.487, .766] .651 2.019 .103
Phonological Easy T 14.616 3.234 9.0 22.5 .467 [.116, .667] .546 [.357, .693] .550 2.477 .120
B 14.217 2.915 8.5 20.5 .504 [.251, .681] .560 [.374, .702] .560 2.191 .109
S 14.130 3.330 8.5 23.5 .548 [.225, .736] .630 [.464, .754} .631 2.242 .112
K 14.399 3.443 8.0 23.0 .547 [.238, .732] .625 [.427, .750] .626 2.340 .115
Hard V 8.551 2.858 2.5 16.5 .473 [.148, .680] .563 [.377, .705] .564 2.138 .177
N 9.725 2.464 4.5 15.0 .508 [.244, .687] .568 [.385, .709] .570 1.828 .133
D 9.928 2.622 4.5 14.5 .430 [.215, .605] .455 [.247, .624] .457 2.268 .162
F 11.616 2.621 6.0 18.5 .540 [.306, .703] .586 [.407, .722] .589 1.893 .115
Note . ICC2, intra-class correlation coefficient (ICC) of absolute agreement; ICC3, ICC of relative consistency; CI, 95% confidence interval; SEM, standard error of measurement; CV, coefficient of variation.
Appendix References 197
References
Abrahams, S., Goldstein, L. H., Simmons, A., Brammer, M. J., Williams, S. C., Giampietro, V. P., Andrew, C. M. & Leigh, P. N. (2003). Functional magnetic resonance imaging of verbal fluency and confrontation naming using compressed image acquisition to permit overt responses. Human brain mapping , 20 (1), 29-40. Acharya, A. B., & Dulebohn, S. C. (2017). Aphasia, Broca. Allen, M. D., & Fong, A. K. (2008). Clinical application of standardized cognitive assessment using fMRI. II. Verbal fluency. Behavioural Neurology , 20 (3-4), 141-152. Almairac, F., Herbet, G., Moritz-Gasser, S., de Champfleur, N. M., & Duffau, H. (2015). The left inferior fronto-occipital fasciculus subserves language semantics: a multilevel lesion study. Brain Structure and Function , 220 (4), 1983-1995. Alvarez, J. A., & Emory, E. (2006). Executive function and the frontal lobes: a meta-analytic review. Neuropsychol Rev, 16 (1), 17-42. Andrade, S. M., Fernández-Calvo, B., Boggio, P. S., de Oliveira, E. A., Gomes, L. F., Júnior, J. E. G. P., ... & Alves, N. T. (2015). Neurostimulation for cognitive rehabilitation in stroke (NeuroCog): study protocol for a randomized controlled trial. Trials , 16(1), 435. Ardila, A., Rosselli, M., & Bateman, J. R. (1994). Factorial structure of cognitive activity using a neuropsychological test battery. Behav Neurol, 7 (2), 49-58. Aschenbrenner, S., Tucha, O., & Lange, K. W. (2000). Regensburger Wortflüssigkeits-Test: RWT . Hogrefe, Verlag für Psychologie. Azuma, T. (2004). Working memory and perseveration in verbal fluency. Neuropsychology , 18 (1), 69. Azuma, T., Bayles, K. A., Cruz, R. F., Tomoeda, C. K., Wood, J. A., McGeagh, A., & Montgomery Jr, E. B. (1997). Comparing the difficulty of letter, semantic, and name fluency tasks for normal elderly and patients with Parkinson's disease. Neuropsychology , 11 (4), 488. Baciu, M., Boudiaf, N., Cousin, E., Perrone-Bertolotti, M., Pichat, C., Fournet, N., ... & Krainik, A. (2016). Functional MRI evidence for the decline of word retrieval and generation during normal aging. Age , 38 (1), 3. Baldo, J. V., & Shimamura, A. P. (1998). Letter and category fluency in patients with frontal lobe lesions. Neuropsychology , 12 (2), 259. Baldo, J. V., Shimamura, A. P., Delis, D. C., Kramer, J., & Kaplan, E. (2001). Verbal and design fluency in patients with frontal lobe lesions. Journal of the International Neuropsychological Society , 7(5), 586-596. Baldo, J. V., Schwartz, S., Wilkins, D., & Dronkers, N. F. (2006). Role of frontal versus temporal cortex in verbal fluency as revealed by voxel-based lesion symptom mapping. Journal of the International Neuropsychological Society , 12(06), 896-900. Baldo, J. V., Schwartz, S., Wilkins, D. P., & Dronkers, N. F. (2010). Double dissociation of letter and category fluency following left frontal and temporal lobe lesions. Aphasiology, 24 (12), 1593-1604. Barker-Collo, S., Starkey, N., Lawes, C. M., Feigin, V., Senior, H., & Parag, V. (2012). Neuropsychological profiles of 5-year ischemic stroke survivors by Oxfordshire stroke classification and hemisphere of lesion. Stroke, 43 (1), 50-55.
198 Appendix References
Barry, D., Bates, M. E., & Labouvie, E. (2008). FAS and CFL forms of verbal fluency differ in difficulty: a meta-analytic study. Applied neuropsychology , 15 (2), 97-106. Basho, S., Palmer, E. D., Rubio, M. A., Wulfeck, B., & Müller, R.-A. (2007). Effects of generation mode in fMRI adaptations of semantic fluency: Paced production and overt speech. Neuropsychologia, 45 (8), 1697-1706. Birn, R. M., Bandettini, P. A., Cox, R. W., Jesmanowicz, A., & Shaker, R. (1998). Magnetic field changes in the human brain due to swallowing or speaking. Magn Reson Med, 40 (1), 55-60. Birn, R. M., Kenworthy, L., Case, L., Caravella, R., Jones, T. B., Bandettini, P. A., & Martin, A. (2010). Neural systems supporting lexical search guided by letter and semantic category cues: a self-paced overt response fMRI study of verbal fluency. Neuroimage , 49 (1), 1099-1107. Biesbroek, J. M., van Zandvoort, M. J., Kappelle, L. J., Velthuis, B. K., Biessels, G. J., & Postma, A. (2016). Shared and distinct anatomical correlates of semantic and phonemic fluency revealed by lesion-symptom mapping in patients with ischemic stroke. Brain Structure and Function , 221 (4), 2123-2134. Birn, R. M., Kenworthy, L., Case, L., Caravella, R., Jones, T. B., Bandettini, P. A., & Martin, A. (2010). Neural systems supporting lexical search guided by letter and semantic category cues: a self-paced overt response fMRI study of verbal fluency. Neuroimage, 49 (1), 1099-1107. Bizzozero, I., Scotti, S., Clerici, F., Pomati, S., Laiacona, M., & Capitani, E. (2013). On which abilities are category fluency and letter fluency grounded? A confirmatory factor analysis of 53 Alzheimer's dementia patients. Dement Geriatr Cogn Dis Extra, 3(1), 179-191. Borkowski, J. G., Benton, A. L., & Spreen, O. (1967). Word fluency and brain damage. Neuropsychologia, 5 (2), 135-140. Bortz, J., & Döring, N. (2006). Quantitative Methoden der Datenerhebung. In: Forschungsmethoden und Evaluation. Springer-Lehrbuch. Springer, Berlin, Heidelberg. Bose, A., Wood, R., & Kiran, S. (2017). Semantic fluency in aphasia: clustering and switching in the course of 1 minute. International Journal of Language & Communication Disorders, 52 (3), 334-345. Bousfield, W. A., & Sedgewick, C. H. W. (1944). An analysis of sequences of restricted associative responses. The Journal of General Psychology , 30 (2), 149-165. Broca, P. (1861). Perte de la parole, ramollissement chronique et destruction partielle du lobe antérieur gauche du cerveau. Bull Soc Anthropol , 2(1), 235-8 Capitani, E., Laiacona, M., & Barbarotto, R. (1999). Gender affects word retrieval of certain categories in semantic fluency tasks. Cortex , 35 (2), 273-278. Channon, S., & Crawford, S. (2000). The effects of anterior lesions on performance on a story comprehension test: Left anterior impairment on a theory of mind-type task. Neuropsychologia , 38 (7), 1006-1017. Chapados, C., & Petrides, M. (2013). Impairment only on the fluency subtest of the Frontal Assessment Battery after prefrontal lesions. Brain , 136 (10), 2966-2978.
Appendix References 199
Chouiter, L., Holmberg, J., Manuel, A. L., Colombo, F., Clarke, S., Annoni, J. M., & Spierer, L. (2016). Partly segregated cortico-subcortical pathways support phonologic and semantic verbal fluency: A lesion study. Neuroscience, 329 , 275-283. Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological assessment , 6(4), 284. Cohen, A. (1966). Errors of speech and their implication for understanding the strategy of language users. IPO Annual Progress Report , 1, 44. Cohen, R. A., Kaplan, R. F., Zuffante, P., Moser, D. J., Jenkins, M. A., Salloway, S., & Wilkinson, H. (1999). Alteration of intention and self-initiated action associated with bilateral anterior cingulotomy. The Journal of neuropsychiatry and clinical neurosciences , 11 (4), 444-453. Costafreda, S. G., Fu, C. H., Lee, L., Everitt, B., Brammer, M. J., & David, A. S. (2006). A systematic review and quantitative appraisal of fMRI studies of verbal fluency: role of the left inferior frontal gyrus. Human Brain Mapping , 27 (10), 799-810. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. psychometrika , 16 (3), 297-334. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological bulletin , 52 (4), 281. Cunje, A., Molloy, D. W., Standish, T. I., & Lewis, D. L. (2007). Alternate forms of logical memory and verbal fluency tasks for repeated testing in early cognitive changes. International psychogeriatrics , 19 (1), 65-75. Curtis, V. A., Bullmore, E. T., Brammer, M. J., Wright, I. C., Williams, S. C., Morris, R. G., ... & McGuire, P. K. (1998). Attenuated frontal activation during a verbal fluency task in patients with schizophrenia. American Journal of Psychiatry , 155 (8), 1056-1063. Damasio, A. R., & Geschwind, N. (1984). The neural basis of language. Annual review of neuroscience , 7(1), 127-147. Davidson, P. S., Gao, F. Q., Mason, W. P., Winocur, G., & Anderson, N. D. (2008). Verbal fluency, Trail Making, and Wisconsin Card Sorting Test performance following right frontal lobe tumor resection. Journal of Clinical and Experimental Neuropsychology , 30 (1), 18-32. Demonet, J. F., Chollet, F., Ramsay, S., Cardebat, D., Nespoulous, J. L., Wise, R., . . . Frackowiak, R. (1992). The anatomy of phonological and semantic processing in normal subjects. Brain, 115 ( Pt 6) , 1753-1768. Devlin, J. T., Matthews, P. M., & Rushworth, M. F. (2003). Semantic processing in the left inferior prefrontal cortex: a combined functional magnetic resonance imaging and transcranial magnetic stimulation study. Journal of Cognitive Neuroscience , 15 (1), 71- 84. De Vreese, L. P., Neri, M., Rubichi, S., & Salvioli, G. (1996). Grammatical ambiguity resolution in right hemisphere-damaged patients: Evidence from an insertion task. Aphasiology , 10 (8), 801-814. Dronkers, N. F., Wilkins, D. P., Van Valin, R. D., Redfern, B. B., & Jaeger, J. J. (2004). Lesion analysis of the brain areas involved in language comprehension. Cognition , 92 (1), 145-177. Ehlis, A. C., Haeussinger, F. B., Gastel, A., Fallgatter, A. J., & Plewnia, C. (2016). Task-
200 Appendix References
dependent and polarity-specific effects of prefrontal transcranial direct current stimulation on cortical activation during word fluency. Neuroimage , 140 , 134-140. Farrell, M. T., Zahodne, L. B., Stern, Y., Dorrejo, J., Yeung, P., & Cosentino, S. (2014). Subjective Word ‐Finding Difficulty Reduces Engagement in Social Leisure Activities in Alzheimer's Disease. Journal of the American Geriatrics Society , 62 (6), 1056-1063. Ferraro, F. R., & Lowell, K. (2010). Boston Naming Test. Corsini Encyclopedia of Psychology . Friederici, A. D. (2011). The brain basis of language processing: from structure to function. Physiol Rev, 91 (4), 1357-1392. Gaillard, W. D., Hertz–Pannier, L., Mott, S. H., Barnett, A. S., LeBihan, D., & Theodore, W. H. (2000). Functional anatomy of cognitive development fMRI of verbal fluency in children and adults. Neurology , 54 (1), 180-180. Gazzaniga, M. S. (1989). Organization of the human brain. Science , 245 (4921), 947-952. Geisinger, K.F. (2013). Reliabilty. In Geisinger, K. F., Bracken, B. A., Carlson, J. F., Hansen, J.-I. C., Kuncel, N. R., Reise, S. P., & Rodriguez, M. C. (Eds.), APA handbook of testing and assessment in psychology, Vol. 1. Test theory and testing and assessment in industrial and organizational psychology (pp. 21 -42). Washington, DC, US: American Psychological Association. Gershberg, F. B., & Shimamura, A. P. (1995). Impaired use of organizational strategies in free recall following frontal lobe damage. Neuropsychologia , 33 (10), 1305-1333. Geschwind, N., Quadfasel, F. A., & Segarra, J. (1968). Isolation of the speech area. Neuropsychologia , 6(4), 327-340. Goldstein, L. B., Bushnell, C. D., Adams, R. J., Appel, L. J., Braun, L. T., Chaturvedi, S., . . . Pearson, T. A. (2011). Guidelines for the primary prevention of stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke, 42 (2), 517-584. Goldstein, B., Obrzut, J. E., John, C., Hunter, J. V., & Armstrong, C. L. (2004). The Impact of Low-Grade Brain Tumors on Verbal Fluency Performance. Journal of Clinical and Experimental Neuropsychology , 26 (6), 750-758. Golestanirad, L., Das, S., Schweizer, T. A., & Graham, S. J. (2015). A preliminary fMRI study of a novel self-paced written fluency task: observation of left-hemispheric activation, and increased frontal activation in late vs. early task phases. Front Hum Neurosci, 9 , 113. Goodglass, H. (1993). Understanding aphasia . Academic Press. Gourovitch, M. L., Kirkby, B. S., Goldberg, T. E., Weinberger, D. R., Gold, J. M., Esposito, G., . . . Berman, K. F. (2000). A comparison of rCBF patterns during letter and semantic fluency. Neuropsychology, 14 (3), 353-360. Gruenewald, P. J., & Lockhead, G. R. (1980). The free recall of category examples. Journal of Experimental Psychology: Human Learning and Memory , 6(3), 225. Heim, S. (2005). The structure and dynamics of normal language processing: insights from neuroimaging. Acta Neurobiol Exp (Wars), 65 (1), 95-116. Heim, S., Eickhoff, S. B., & Amunts, K. (2008). Specialisation in Broca's region for semantic, phonological, and syntactic fluency?. Neuroimage , 40 (3), 1362-1368.
Appendix References 201
Helmstaedter, C., Gleissner, U., Zentner, J., & Elger, C. E. (1998). Neuropsychological consequences of epilepsy surgery in frontal lobe epilepsy. Neuropsychologia , 36 (7), 681-689. Henry, J. D., & Crawford, J. R. (2004). A meta-analytic review of verbal fluency performance following focal cortical lesions. Neuropsychology, 18 (2), 284-295. Herrmann, M. J., Ehlis, A. C., & Fallgatter, A. J. (2003). Frontal activation during a verbal- fluency task as measured by near-infrared spectroscopy. Brain Research Bulletin , 61 (1), 51-56. Hirshorn, E. A., & Thompson-Schill, S. L. (2006). Role of the left inferior frontal gyrus in covert word retrieval: neural correlates of switching during verbal fluency. Neuropsychologia, 44 (12), 2547-2557. Huang, J., Carr, T. H., & Cao, Y. (2002). Comparing cortical activations for silent and overt speech using event-related fMRI. Hum Brain Mapp, 15 (1), 39-53. Hugdahl, K., Lundervold, A., Ersland, L., Smievoll, A. I., Sundbekg, H., Bakndon, R., & Roscher, B. E. (1999). Left frontal activation during a semantic categorization task: an fMRI-study. International Journal of Neuroscience , 99 (1-4), 49-58. Hughlings Jacksons, J. (1878). On affections of speech from disease of the brain. Brain ,1, 304-330. Indefrey, P. (2011). The spatial and temporal signatures of word production components: a critical update. Front Psychol, 2 , 255. Indefrey, P., & Levelt, W. J. (2000). The neural correlates of language production. In The new cognitive neurosciences; 2nd ed. (pp. 845-865). MIT press. Indefrey, P., & Levelt, W. J. (2004). The spatial and temporal signatures of word production components. Cognition, 92 (1-2), 101-144. Janowsky, J. S., Shimamura, A. P., Kritchevsky, M., & Squire, L. R. (1989). Cognitive impairment following frontal lobe damage and its relevance to human amnesia. Behavioral neuroscience , 103 (3), 548. Joanette, Y., & Goulet, P. (1986). Criterion-specific reduction of verbal fluency in right brain- damaged right-handers. Neuropsychologia , 24 (6), 875-879. Jones-Gotman, M., & Milner, B. (1977). Design fluency: The invention of nonsense drawings after focal cortical lesions. Neuropsychologia , 15 (4), 653-674. Jurado, M. A., Mataro, M., Verger, K., Bartumeus, F., & Junque, C. (2000). Phonemic and semantic fluencies in traumatic brain injury patients with focal frontal lesions. Brain Inj, 14 (9), 789-795. Jurado, M. B., & Rosselli, M. (2007). The elusive nature of executive functions: a review of our current understanding. Neuropsychol Rev, 17 (3), 213-233. Katzev, M., Tuscher, O., Hennig, J., Weiller, C., & Kaller, C. P. (2013). Revisiting the functional specialization of left inferior frontal gyrus in phonological and semantic fluency: the crucial role of task demands and individual ability. J Neurosci, 33 (18), 7837-7845. Kim, J., Fann, D. Y., Seet, R. C., Jo, D. G., Mattson, M. P., & Arumugam, T. V. (2016). Phytochemicals in Ischemic Stroke. Neuromolecular Med, 18 (3), 283-305. Kim, H., Kim, J., Kim, D. Y., & Heo, J. (2011). Differentiating between aphasic and nonaphasic stroke patients using semantic verbal fluency measures with administration time of 30 seconds. European neurology , 65 (2), 113-117.
202 Appendix References
Kolb B, Whishaw IQ (2008) Fundamentals of human neuropsychology (6th ed). New York, NY: Worth Publishers. Leclercq, M., Couillet, J., Azouvi, P., Marlier, N., Martin, Y., Strypstein, E., & Rousseaux, M. (2000). Dual task performance after severe diffuse traumatic brain injury or vascular prefrontal damage. Journal of Clinical and Experimental Neuropsychology , 22 (3), 339-350. Ledoux, K., Vannorsdall, T. D., Pickett, E. J., Bosley, L. V., Gordon, B., & Schretlen, D. J. (2014). Capturing additional information about the organization of entries in the lexicon from verbal fluency productions. Journal of Clinical and Experimental Neuropsychology , 36 (2), 205-220. Levelt, W. J. (2001). Spoken word production: a theory of lexical access. Proc Natl Acad Sci U S A, 98 (23), 13464-13471. Levelt, W. J., Praamstra, P., Meyer, A. S., Helenius, P., & Salmelin, R. (1998). An MEG study of picture naming. J Cogn Neurosci, 10 (5), 553-567. Levelt, W. J., Roelofs, A., & Meyer, A. S. (1999). A theory of lexical access in speech production. Behav Brain Sci, 22 (1), 1-38; discussion 38-75. Levy, D. A., Bayley, P. J., & Squire, L. R. (2004). The anatomy of semantic knowledge: medial vs. lateral temporal lobe. Proceedings of the National Academy of Sciences of the United States of America , 101 (17), 6710-6715. Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychological assessment (5th ed.). New York, NY, US: Oxford University Press. Li, M., Zhang, Y., Song, L., Huang, R., Ding, J., Fang, Y., ... & Han, Z. (2017). Structural connectivity subserving verbal fluency revealed by lesion-behavior mapping in stroke patients. Neuropsychologia , 101 , 85-96. Loonstra, A. S., Tarlow, A. R., & Sellers, A. H. (2001). COWAT metanorms across age, education, and gender. Applied neuropsychology , 8(3), 161-166. Luckhurst, L., & Lloyd-Jones, T. J. (2001). A selective deficit for living things after temporal lobectomy for relief of epileptic seizures. Brain and Language , 79 (2), 266-296. Martin, R. C., Loring, D. W., Meador, K. J., & Lee, G. P. (1990). The effects of lateralized temporal lobe dysfunction on normal and semantic word fluency. Neuropsychologia , 28 (8), 823-829. Martyr, A., Clare, L., Nelis, S. M., Marková, I. S., Roth, I., Woods, R. T., ... & Morris, R. G. (2012). Verbal fluency and awareness of functional deficits in early-stage dementia. The Clinical Neuropsychologist , 26 (3), 501-519. McGuire, P. K., Silbersweig, D. A., & Frith, C. D. (1996). Functional neuroanatomy of verbal self-monitoring. Brain, 119 ( Pt 3) , 907-917. Melinder, M. R., Barch, D. M., Heydebrand, G., & Csernansky, J. G. (2005). Easier tasks can have better discriminating power: the case of verbal fluency. Journal of Abnormal Psychology , 114 (3), 385. Milner, B. (1964). Some effects of frontal lobectomy in man. In: J.M. Warren & K. Akert (Eds.), The frontal granular cortex and behavior (pp. 313-334). New York: McGraw- Hill. Moscovitch, M. (1994). Cognitive resources and dual-task interference effects at retrieval in normal people: The role of the frontal lobes and medial temporal cortex. Neuropsychology , 8(4), 524.
Appendix References 203
Mottaghy, F. M., Sparing, R., & Töpper, R. (2006). Enhancing picture naming with transcranial magnetic stimulation. Behavioural neurology , 17 (3, 4), 177-186. Mummery, C. J., Patterson, K., Hodges, J. R., & Wise, R. J. (1996). Generating 'tiger' as an animal name or a word beginning with T: differences in brain activation. Proceedings in the Royal Society B: Biological Sciences , 263(1373), 989-995. Netsell, R., & Daniel, B. (1974). Neural and mechanical response time for speech production. Journal of Speech, Language, and Hearing Research , 17 (4), 608-618. Owen, A. M., Downes, J. J., Sahakian, B. J., Polkey, C. E., & Robbins, T. W. (1990). Planning and spatial working memory following frontal lobe lesions in man. Neuropsychologia , 28 (10), 1021-1034. Parks, R. W., Levine, D. S., Long, D. L., Crockett, D. J., Dalton, I. E., Weingartner, H., . . . et al. (1992). Parallel distributed processing and neuropsychology: a neural network model of Wisconsin Card Sorting and verbal fluency. Neuropsychol Rev, 3 (2), 213- 233. Patel, M. D., Coshall, C., Rudd, A. G., & Wolfe, C. D. (2002). Cognitive impairment after stroke: clinical determinants and its associations with long-term stroke outcomes. J Am Geriatr Soc, 50 , 700-706. Patterson, J. (2011). Verbal Fluency. In Encyclopedia of Clinical Neuropsychology (pp. 2603- 2606). Springer New York. Paulesu, E., Goldacre, B., Scifo, P., Cappa, S. F., Gilardi, M. C., Castiglioni, I., . . . Fazio, F. (1997). Functional heterogeneity of left inferior frontal cortex as revealed by fMRI. Neuroreport, 8 (8), 2011-2017. Perani, D., Cappa, S. F., Tettamanti, M., Rosa, M., Scifo, P., Miozzo, A., ... & Fazio, F. (2003). A fMRI study of word retrieval in aphasia. Brain and Language , 85 (3), 357- 368. Perret, E. (1974). The left frontal lobe of man and the suppression of habitual responses in verbal categorical behaviour. Neuropsychologia , 12 (3), 323-330. Peterburs, J., Bellebaum, C., Koch, B., Schwarz, M., & Daum, I. (2010). Working memory and verbal fluency deficits following cerebellar lesions: relation to interindividual differences in patient variables. The Cerebellum , 9(3), 375-383. Petrea, R. E., Beiser, A. S., Seshadri, S., Kelly-Hayes, M., Kase, C. S., & Wolf, P. A. (2009). Gender Differences in Stroke Incidence and Poststroke Disability in the Framingham Heart Study. Stroke, 40 (4), 1032-1037. Phelps, E. A., Hyder, F., Blamire, A. M., & Shulman, R. G. (1997). FMRI of the prefrontal cortex during overt verbal fluency. Neuroreport , 8(2), 561-565. Poldrack, R. A., Wagner, A. D., Prull, M. W., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. (1999). Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex. Neuroimage , 10 (1), 15-35. Poulin, V., Korner-Bitensky, N., & Dawson, D. R. (2013). Stroke-specific executive function assessment: a literature review of performance-based tools. Aust Occup Ther J, 60 (1), 3-19. Price, C. J. (2010). The anatomy of language: a review of 100 fMRI studies published in 2009. Annals of the New York Academy of Sciences , 1191 (1), 62-88.
204 Appendix References
Price, S. E., Kinsella, G. J., Ong, B., Storey, E., Mullaly, E., Phillips, M., ... & Perre, D. (2012). Semantic verbal fluency strategies in amnestic mild cognitive impairment. Neuropsychology , 26 (4), 490. Rende, B., Ramsberger, G., & Miyake, A. (2002). Commonalities and differences in the working memory components underlying letter and category fluency tasks: a dual-task investigation. Neuropsychology , 16(3), 309-321. Reverberi, C., Laiacona, M., & Capitani, E. (2006). Qualitative features of semantic fluency performance in mesial and lateral frontal patients. Neuropsychologia , 44(3), 469-478. Robinson, G., Shallice, T., Bozzali, M., & Cipolotti, L. (2012). The differing roles of the frontal cortex in fluency tests. Brain , 135 (7), 2202-2214. Roby-Brami, A., Hermsdorfer, J., Roy, A. C., & Jacobs, S. (2012). A neuropsychological perspective on the link between language and praxis in modern humans. Philos Trans R Soc Lond B Biol Sci, 367 (1585), 144-160. Rogers, R. D., Sahakian, B. J., Hodges, J. R., Polkey, C. E., Kennard, C., & Robbins, T. W. (1998). Dissociating executive mechanisms of task control following frontal lobe damage and Parkinson's disease. Brain: a journal of neurology , 121 (5), 815-842. Rosen, V. M., & Engle, R. W. (1997). The role of working memory capacity in retrieval. Journal of Experimental Psychology: General , 126(3), 211-227. Ruff, R. M., Light, R. H., Parker, S. B., & Levin, H. S. (1997). The psychological construct of word fluency. Brain and language , 57 (3), 394-405. Schecklmann, M., Ehlis, A. C., Plichta, M. M., & Fallgatter, A. J. (2008). Functional near- infrared spectroscopy: a long-term reliable tool for measuring brain activity during verbal fluency. Neuroimage , 43 (1), 147-155. Schlösser, R., Hutchinson, M., Joseffer, S., Rusinek, H., Saarimaki, A., Stevenson, J., ... & Brodie, J. D. (1998). Functional magnetic resonance imaging of human brain activity in a verbal fluency task. Journal of Neurology, Neurosurgery & Psychiatry , 64 (4), 492- 498. Schmidt, C. S., Schumacher, L. V., Römer, P., Leonhart, R., Beume, L., Martin, M., ... & Kaller, C. P. (2017). Are semantic and phonological fluency based on the same or distinct sets of cognitive processes? Insights from factor analyses in healthy adults and stroke patients. Neuropsychologia , 99 , 148-155. Schwartz, S., & Baldo, J. (2001). Distinct patterns of word retrieval in right and left frontal lobe patients: a multidimensional perspective. Neuropsychologia , 39 (11), 1209-1217. Schwartz, S., Baldo, J., Graves, R. E., & Brugger, P. (2003). Pervasive influence of semantics in letter and category fluency: A multidimensional approach. Brain and language , 87 (3), 400-411. Shao, Z., Janse, E., Visser, K., & Meyer, A. S. (2014). What do verbal fluency tasks measure? Predictors of verbal fluency performance in older adults. Front Psychol, 5 , 772. Simanova, I., Hagoort, P., Oostenveld, R., & Van Gerven, M. A. (2012). Modality- independent decoding of semantic information from the human brain. Cerebral Cortex , 24 (2), 426-434. Sireci, S.G. & Sukin, T. (2013). Test Validity. In Geisinger, K. F., Bracken, B. A., Carlson, J. F., Hansen, J.-I. C., Kuncel, N. R., Reise, S. P., & Rodriguez, M. C. (Eds.), APA
Appendix References 205
handbook of testing and assessment in psychology, Vol. 1. Test theory and testing and assessment in industrial and organizational psychology (pp. 61 -84). Washington, DC, US: American Psychological Association. Small, S. L., & Burton, M. W. (2002). Functional magnetic resonance imaging studies of language. Curr Neurol Neurosci Rep, 2 (6), 505-510. Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A Compendium of Neuropsychological Tests New York: Oxford University Press. Stanhope, N., Guinan, E., & Kopelman, M. D. (1998). Frequency judgements of abstract designs by patients with diencephalic, temporal lobe or frontal lobe lesions. Neuropsychologia , 36 (12), 1387-1396. Stefanova, E. D., Kostic, V. S., Ziropadja, L., Markovic, M., & Ocic, G. (2002). Serial position learning effects in patients with aneurysms of the anterior communicating artery. Journal of clinical and experimental neuropsychology , 24 (5), 687-694. Stuss, D. T., & Alexander, M. P. (2000). Executive functions and the frontal lobes: a conceptual view. Psychol Res, 63 (3-4), 289-298. Stuss, D. T., Alexander, M. P., Hamer, L., Palumbo, C., Dempster, R., Binns, M., ... & Izukawa, D. (1998). The effects of focal anterior and posterior brain lesions on verbal fluency. Journal of the International Neuropsychological Society , 4(3), 265-278. Stuss, D. T., Alexander, M. P., Palumbo, C. L., Buckle, L., Sayer, L., & Pogue, J. (1994). Organizational strategies with unilateral or bilateral frontal lobe injury in word learning tasks. Neuropsychology , 8(3), 355. Stuss, D. T., & Benson, D. F. (1986). The frontal lobes. New York: Raven. Stuss, D. T., Craik, F. I., Sayer, L., Franchi, D., & Alexander, M. P. (1996). Comparison of older people and patients with frontal lesions: Evidence from word list learning. Psychology and Aging , 11 (3), 387. Stuss, D. T., Toth, J. P., Franchi, D., Alexander, M. P., Tipper, S., & Craik, F. I. (1999). Dissociation of attentional processes in patients with focal frontal and posterior lesions. Neuropsychologia , 37 (9), 1005-1027. Sylvester, C. Y. C., & Shimamura, A. P. (2002). Evidence for intact semantic representations in patients with frontal lobe lesions. Neuropsychology , 16 (2), 197. Szatkowska, I., Grabowska, A., & Szymanska, O. (2000). Phonological and semantic fluencies are mediated by different regions of the prefrontal cortex. Acta Neurobiol Exp (Wars), 60 , 503-508. Tang, E. Y., Amiesimaka, O., Harrison, S. L., Green, E., Price, C., Robinson, L., ... & Stephan, B. C. (2018). Longitudinal Effect of Stroke on Cognition: A Systematic Review. Journal of the American Heart Association , 7(2), e006443. Thompson-Schill, S. L., Swick, D., Farah, M. J., D’Esposito, M., Kan, I. P., & Knight, R. T. (1998). Verb generation in patients with focal frontal lesions: A neuropsychological test of neuroimaging findings. Proceedings of the National Academy of Sciences, 95 (26), 15855-15860. Thurstone, L.L. (1938). Primary mental abilities. : University of Chicago Press: Chicago. Troyer, A. K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two components of verbal fluency: evidence from younger and older healthy adults. Neuropsychology , 11 (1), 138.
206 Appendix References
Troyer, A. K., Moscovitch, M., Winocur, G., Alexander, M. P., & Stuss, D. (1998). Clustering and switching on verbal fluency: the effects of focal frontal- and temporal- lobe lesions. Neuropsychologia, 36 (6), 499-504. Tucha, O., Smely, C., & Lange, K. W. (1999). Verbal and figural fluency in patients with mass lesions of the left or right frontal lobes. Journal of Clinical and Experimental Neuropsychology , 21 (2), 229-236. Tupak, S. V., Badewien, M., Dresler, T., Hahn, T., Ernst, L. H., Herrmann, M. J., . . . Ehlis, A. C. (2012). Differential prefrontal and frontotemporal oxygenation patterns during phonemic and semantic verbal fluency. Neuropsychologia, 50 (7), 1565-1569. Ullman, M. T. (2001). A neurocognitive perspective on language: The declarative/procedural model. Nature Reviews Neuroscience , 2(10), 717. Unsworth, N., Spillers, G. J., & Brewer, G. A. (2011). Variation in verbal fluency: a latent variable analysis of clustering, switching, and overall performance. Q J Exp Psychol (Hove), 64 (3), 447-466. Van Der Elst, W. I. M., Van Boxtel, M. P., Van Breukelen, G. J., & Jolles, J. (2006). Normative data for the Animal, Profession and Letter M Naming verbal fluency tests for Dutch speaking participants and the effects of age, education, and sex. Journal of the International Neuropsychological Society, 12 (1), 80-89. Wagner, S., Sebastian, A., Lieb, K., Tüscher, O., & Tadi ć, A. (2014). A coordinate-based ALE functional MRI meta-analysis of brain activation during verbal fluency tasks in healthy control subjects. BMC neuroscience , 15 (1), 19. Walter Johnson, Oyere Onuma, Mayowa Owolabi & Sonal Sachdev (2016). Stroke: a global response is needed. Bulletin of the World Health Organization ; 94:634-634A. Weiller, C., Isensee, C., Rijntjes, M., Huber, W., Müller, S., Bier, D., ... & Diener, H. C. (1995). Recovery from Wernicke's aphasia: a positron emission tomographic study. Annals of neurology , 37 (6), 723-732. Weiss, E. M., Siedentopf, C., Hofer, A., Deisenhammer, E. A., Hoptman, M. J., Kremser, C., ... & Delazer, M. (2003). Brain activation pattern during a verbal fluency test in healthy male and female volunteers: a functional magnetic resonance imaging study. Neuroscience Letters , 352 (3), 191-194. Wernicke, C. (1874). Der aphasische Symptomencomplex: eine psychologische Studie auf anatomischer Basis . Cohn. Whiteside, D. M., Kealey, T., Semla, M., Luu, H., Rice, L., Basso, M. R., & Roper, B. (2016). Verbal Fluency: Language or Executive Function Measure? Appl Neuropsychol Adult, 23 (1), 29-34. Wood, C. C., Goff, W. R., & Day, R. S. (1971). Auditory evoked potentials during speech perception. Science , 173 (4003), 1248-1251. World Health Organization. (2017). The top 10 causes of death. Retrieved from http://www.who.int/mediacentre/factsheets/fs310/en/ Xu, X. D., Ren, H. Y., Prakash, R., Vijayadas, & Kumar, R. (2013). Outcomes of neuropsychological interventions of stroke. Ann Indian Acad Neurol, 16 (3), 319-328. Ye, Z., & Zhou, X. (2009). Executive control in language processing. Neurosci Biobehav Rev, 33 (8), 1168-1177. Ziemann, U. (2010). TMS in cognitive neuroscience: virtual lesion and beyond. Cortex , 46 (1), 124-127.
Appendix Curriculum Vitae 207
Curriculum Vitae
Die Seiten 207-208 (Lebenslauf) enthalten persönliche Daten. Sie sind deshalb nicht Bestandteil der Online-Veröffentlichung.
208 Appendix Curriculum Vitae
Appendix List of Publications 209
List of Publications
Published Peer-Reviewed Articles
2015
Köstering, L., Schmidt, C. S. , Egger, K., Amtage, F., Peter, J., Kloppel, S., Beume L. A.,
Hoeren M., Weiller C., Kaller, C. P. (2015). Assessment of planning performance in
clinical samples: Reliability and validity of the Tower of London task (TOL-F).
Neuropsychologia, 75 , 646-655.
Schmidt, C. S. , Lassonde, M., Gagnon, L., Sauerwein, C. H., Carmant, L., Major, P.,
Paquette N., Lepore F., Gallagher, A. (2015). Neuropsychological functioning in
children with temporal lobe epilepsy and hippocampal atrophy without mesial
temporal sclerosis: a distinct clinical entity? Epilepsy Behav, 44 , 17-22.
2016
Dressing, A., Nitschke, K., Kümmerer, D., Bormann, T., Beume, L., Schmidt, C. S. , Ludwig,
M. V., Mader, I., Willmes, K., Rijntjes, M., Kaller, C. P., Weiller, C. & Markus
Martin (2016). Distinct Contributions of Dorsal and Ventral Streams to Imitation of
Tool-Use and Communicative Gestures. Cerebral Cortex .
Köstering, L., Schmidt, C. S. , Weiller, C., & Kaller, C. P. (2016). Analyses of Rule Breaks
and Errors During Planning in Computerized Tower Tasks: Insights From
Neurological Patients. Arch Clin Neuropsychol .
Martin, M., Beume, L., Kummerer, D., Schmidt, C. S. , Bormann, T., Dressing, A., Ludwig
V. M., Umarova R. M., Mader I., Rijntjes M., Kaller C. P., Weiller, C. (2016).
Differential Roles of Ventral and Dorsal Streams for Conceptual and Production-
Related Components of Tool Use in Acute Stroke Patients. Cereb Cortex, 26 (9), 3754-
3771.
210 Appendix List of Publications
Martin, M., Dressing, A., Bormann, T., Schmidt, C. S. , Kummerer, D., Beume, L., Sauer D.,
Mader I., Rijntjes M., Kaller C. P., Weiller, C. (2016). Componential Network for the
Recognition of Tool-Associated Actions: Evidence from Voxel-based Lesion-
Symptom Mapping in Acute Stroke Patients. Cerebral Cortex .
2017
Beume, L. A., Martin, M., Kaller, C. P., Kloppel, S., Schmidt, C. S. , Urbach, H., Egger K.,
Rijntjes M., Weiller C., Umarova, R. M. (2017). Visual neglect after left-hemispheric
lesions: a voxel-based lesion-symptom mapping study in 121 acute stroke patients.
Exp Brain Res .
*Schmidt, C. S. , Schumacher, L. V., Römer, P., Leonhart, R., Beume, L., Martin, M.,
Dressing, A., Weiller, C. & Kaller, C. P. (2017). Are semantic and phonological
fluency based on the same or distinct sets of cognitive processes? Insights from factor
analyses in healthy adults and stroke patients. Neuropsychologia , 99 , 148-155.
* Studies of this thesis.
Appendix List of Publications 211
Oral Presentations
2017
Schmidt CS & Kaller CP (2017). Semantische und phonologische Wortflüssigkeit: Neue
Daten zu alten Hypothesen. Oral presentation at the annual meeting of the German
Society of Neuropsychology (GNP), Sep 21st-23rd 2017, Konstanz, Germany.
Poster Presentations
2013
Schmidt CS , Carmant L, Sauerwein C, Major P, Lepore F, Gallagher A (2013). The
relationship between memory problems and hippocampal asymmetrie in children with
temporal lobe epilepsy. Poster presented at the 30th International Epilepsy Congress,
June 23rd-27th, Montreal, Canada.
2014
Schmidt CS , Köstering L, Nitschke K, Schumacher FK, Weiller C, Kaller CP (2014).
Effects of item difficulty on the test-retest reliability of the verbal fluency task. Poster
presented at the annual meeting of the German Society of Neurology, Sep 15-19th
2014, Munich, Germany.
Schmidt CSM , Köstering L, Nitschke K, Schumacher FK, Weiller C, Kaller CP (2014). The
retest-reliability of the verbal fluency task and the impact of item difficulty. Poster
presented at the 2014 Meeting "Psychologie und Gehirn", June 19th-21st, Lübeck,
Germany.
2015
Schmidt CS , Köstering L, Graebner K, Luzay L, Urbach H, Weiller C, Reis J, Kaller CP
(2015). Predicting inter-individual differences in planning performance from
transcallosal signal transduction between left and right mid-dlPFC – a TMS-EEG
212 Appendix List of Publications
study. Poster presented at the 2015 meeting “Psychologie und Gehirn”, June 4-6th,
Frankfurt, Germany.
2016
Schmidt CS , Köstering L, Reisert M, Luzay L, Graebner K, Urbach H, Weiller C, Reis J,
Kaller CP (2016). Interhemispheric signal propagation in complex cognition:
Combined TMS-EEG over bilateral mid-dlPFC. Poster presented at the
annual meeting of the Organization for Human Brain Mapping, June 26-30th, Geneva,
Switzerland.