Journal of Alzheimer’s Disease 19 (2010) 781–793 781 DOI 10.3233/JAD-2010-1275 IOS Press Review Article Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease. A Critical Review

Liam D. Kaufmana,b,∗, Jay Prattc, Brian Levinec,d and Sandra E. Blacka,b,d aLC Campbell Cognitive Research Unit, Division of Neurology, Deparatment of Medicine, Sunnybrook Research Institute, Sunnybrook Health Sciences Centre, University of Toronto Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada bInstitute of Medical Science, University of Toronto, Toronto, Ontario, Canada cDepartment of Psychology University of Toronto, Toronto, Ontario, Canada dThe Rotman Research Institute, Toronto, Ontario, Canada

Accepted 17 September 2009

Abstract. The number of people living with Alzheimer’s disease (AD), the major cause of dementia, is projected to increase dramatically over the next few decades, making the search for treatments and tools to measure the progression of AD increasingly urgent. The antisaccade task, a hands- and language-free measure of inhibitory control, has been utilized in AD as a potential diagnostic test. While antisaccades do not appear to differentiate AD from healthy aging better than measures of episodic memory, they may still be beneficial. Specifically, antisaccades may provide not only a functional index of the Dorsolateral Prefrontal Cortex (DLPFC), which is damaged in the later stages of AD, but also a tool for monitoring the progression of AD. Further work is required to: 1) strengthen the link between antisaccade errors, in AD, with the DLPFC; 2) insure that antisaccade errors do not result from memory, visuospatial, or other deficits associated with AD; and 3) further validate the clinical analogue of the antisaccade task.

Keywords: Alzheimer’s disease, antisaccades, dementia, dorsolateral prefrontal cortex

INTRODUCTION rope, while in Asia it will nearly quadruple [1]. As new pharmaceuticals are developed to treat and possi- Alzheimer’s disease (AD), characterized by gradual, bly prevent AD, early diagnosis and disease treatment progressive loss of episodic memory, is the most com- monitoring will become increasingly important. Cur- mon single cause of dementia affecting four million rently used NINCDS-ADRDA diagnostic criteria and Americans and is quickly becoming one of the “most the newly proposed criteria [2] both include deficits burdensome health conditions worldwide” [1]. In the in episodic memory as the core diagnostic feature of next two decades, the number of individuals diagnosed AD. Although decline in episodic memory is central with AD will nearly double in North America and Eu- to typical AD, an understanding of additional deficits associated with AD may aid in tracking the progression ∗Correspondence to: Liam D. Kaufman, LC Campbell Cognitive of AD and monitoring the effectiveness of treatments. Research Unit, Division of Neurology, Dept. of Medicine, Sunny- Once such deficit that has been noted in patients with brook Research Institute, Sunnybrook Health Sciences Centre, Uni- AD is difficulty exerting flexible control over prepotent versity of Toronto, Room A421- 2075 Bayview Avenue, Toronto, Ontario, M4N 3M5, Canada. Tel.: +1 416 480 4551; Fax: +1 416 saccades during the antisaccade task (Table 1) [3–10]. 480 4552; E-mail: [email protected]. Results from antisaccade studies indicate that the task

ISSN 1387-2877/10/$27.50  2010 – IOS Press and the authors. All rights reserved 782 L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease may have potential for monitoring progression, specif- direction than adults [13,14]. Between the ages of 10 ically the emergence of dorsofrontal functional deficits and 15, antisaccade performance improves dramatical- in AD, as well as monitoring new treatments. ly and continues to improve into early adulthood. The reduction in antisaccade error rates appears to close- ly mirror structural changes that occur from childhood THE ANTISACCADE TASK to adulthood. For example, volumetric imaging has shown that between the ages of 8 and 22, total white In the antisaccade task, a participant is told to inhibit matter volume increases linearly with age [15]. In the a reflexive (or automatic), visually-guided saccade to frontal lobes, grey matter shows a non-linear increase a peripheral target, and to make an antisaccade in the over time, with a peak volume around the age of 12, opposite direction to a non-existent target (Fig. 1) [11]. followed by a gradual decline presumed to relate to Thus, the antisaccade task probes one’s ability to exert synaptic extermination [15]. In addition to volumet- flexible control by overcoming the prepotent reflexive ric indices of structure, diffusion tensor imaging (DTI) saccade response. If the participant fails to inhibit the has revealed that fractional anisotropy, a measure of reflexive saccade and makes a saccade towards the pe- white matter microstructural integrity and myelination, ripheral target, this constitutes an antisaccade (i.e., in- increases in the frontal lobes from childhood to adult- hibition) error. The task has been widely adopted in hood [16]. some clinical disorders because of several advantages During older adulthood, aging is associated with a over other cognitive tests: it does not require a verbal general decline in grey matter volume [17] and a reduc- or manual response and is well tolerated in patients tion in white matter fractional anisotropy [18]: a grad- with dementia, including AD, and frontotemporal de- ual reversal of developmental changes. Studies of an- generation (FTD) [8]. Furthermore, patients are often tisaccade performance from young adulthood onwards unaware of their mistakes and rarely, if ever, become have reported either a non-significant upward trend in frustrated. Although the task relies on making a par- error rates [4,7,12,19] or a significant increase in error simonious response – a saccade –, multiple and easily rates with aging [13,20,21]. Furthermore, fMRI has quantifiable metrics, such as corrected versus uncor- revealed a compensatory shift in antisaccade related ac- rected errors, saccade amplitudes, and latencies, can be tivation between young adults and older adults indica- derived from the task. tive of functional differences between young and older Recently, the neural correlates of the task have be- adults [19]. come better understood [12]. An absence of verbal or manual responses enables the antisaccade task to be used in neuroimaging environments, such as mag- FUNCTIONAL IMAGING STUDIES netoencephelography (MEG) and functional magnet- ic resonance imaging (fMRI), which do not tolerate Developmental and aging studies indicate a relation- movement well. Consequently non-human primates ship between frontal lobe function and inhibitory fail- can be studied using the antisaccade task, providing a ure in the antisaccade task. Functional neuroimag- model for understanding its neural correlates [12]. The ing techniques such as positron emission tomography relative simplicity of the antisaccade task has enabled (PET) and fMRI have provided more specific infor- children, adolescents, adults, and the elderly to com- mation on neural substrates, implicating specific re- plete the task, which has also provided developmental gions within the frontal lobes for successful antisac- data [13]. cade generation. In the simplest imaging experiments using either PET or fMRI, blocks of antisaccades were compared to prosaccades, and revealed greater activa- DEVELOPMENTAL CHANGES tion for antisaccades in the frontal eye fields and the superior parietal lobule, when compared with prosac- The frontal lobes undergo rapid changes from child- cades [22,23]. Early functional imaging studies led to hood to adolescence, followed by gradual changes dur- conflicting views on the involvement of the dorsolat- ing later adulthood as the later evolved structures such eral prefrontal cortex (DLPFC): some found activation as the frontal lobes gradually become fully myelinated. in the right DLPFC [24,25], while others did not [22, Children under 10 years of age have great difficulty 26]. However, block designs have many shortcomings performing the antisaccade task, making more errors in for analyzing antisaccade related activity. Block de- L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease 783 signs do not allow temporal differentiation between the by Stuss and Alexander [31], task setting and task mon- components of an antisaccade such as: 1) inhibiting a itoring. The antisaccade task requires greater task-set prosaccade; 2) generating an antisaccade; and 3) mak- demands than the prosaccade task, which may explain ing a retrosaccade back to the central fixation. There is the greater left DLPFC activation for antisaccades rel- the potential that activation could either be nullified or ative to prosaccades. In contrast to task setting, involv- enhanced by a negative or a positive response respec- ing the left DLPFC, Stuss and Alexander [31] report- tively, from one of the other components. Furthermore ed an association between damage of the right lateral it is difficult to tease apart the effect of antisaccade di- dorsalfrontal regions and impairments in task monitor- rectional errors, reduced latencies, and hypometric re- ing. Greater right DLPFC activation for correct an- sponses on the blood oxygen level dependent (BOLD) tisaccades, relative to incorrect antisaccades, may re- signal. Event-related designs, in which each antisac- flect increased in task monitoring related to successful cade event is temporally spaced, and later averaged, antisacccade performance. have been used to overcome these problems. Using an event-related design, the BOLD response associated with either preparation or with the motor FOCAL LESION STUDIES phase of an antisaccade can be compared. When the BOLD response during these two phases was examined, The involvement of the DLPFC in antisaccades can it was discovered that increases in BOLD signal in the be inferred indirectly from functional neuroimaging right DLPFC and bilateral frontal eye fields, associated studies, but data from patients with focal lesions pro- with antisaccades, were actually due to preparation and vide more direct evidence that its integrity is necessary not action [24]. These findings were confirmed in a for making correct antisaccades. Lesions affecting ei- combined electroencephalography (EEG)/MEG study ther the left or right mid-DLPFC are consistently asso- which also found more activity in the medial aspects ciated with increased error rates [32–34]. For example, of the frontal eye fields, supplementary eye fields, and Pierrot-Deseilligny and colleagues examined patients prefrontal cortex during the planning phase of an an- with lesions that affected the posterior parietal cortex, tisaccade [27]. Finally, during the retrosaccade, when the frontal eye fields, supplementary motor area, or the the participant makes a saccade back to the center, there DLPFC [33]. They found that patients with lesions in is actually a negative BOLD response [28], suggesting the right or left DLPFC made more directional errors a mechanism by which retrosaccade activity may have than controls or patients with lesions in the frontal eye cancelled out positive DLPFC BOLD signal in some fields or supplementary motor area. In a recent study, block designs. patients with lesions resulting from infarcts were sep- Comparison between successful antisaccades and arated into two groups: those whose antisaccade per- antisaccade errors has also revealed some important formance was in the same range as normal controls details. For instance, fMRI and EEG have demon- and those who were outside that range [34]. Lesion strated increased BOLD signal and increased negative analysis showed that damage in the high error group potentials, respectively, in the DLPFC during correct primarily involved areas in either the right or left dor- antisaccades when compared with incorrect [29,30]. sofrontal cortex that have efferent connections through When Ford and colleagues compared correct antisac- the anterior limb of the internal capsule with the supe- cades with correct prosaccades, they found greater ac- rior colliculi. Specifically the mid-DLPFC, which has tivity in the anterior cingulate cortex, left DLPFC, bi- efferent connections through the anterior limb of the lateral frontal eye fields, pre-supplementary eye fields, anterior capsule into the superior colliculi, was the only and parietal areas [30]. When they compared correct area that was damaged in each patient in the high error to incorrect antisaccades, they found a greater BOLD group. Although Ploner and colleagues did not con- response in the right DLPFC, anterior cingulate cor- duct significance tests on latency differences between tex, and pre-supplementary eye fields. Thus, the left the two groups, the high error group showed longer DLPFC shows greater activity for antisaccades com- latencies than the low error group. Amplitude was not pared with prosaccades, and the right DLPFC exhibits reported [34]. greater activity for correct than incorrect antisaccades. Gaymard and colleagues observed that a patient with The differences between right and left DLPFC activa- a small focal lesion affecting the connections between tion reported by Ford and colleagues parallel the local- the DLPFC and superior colliculi made more direc- ization of two of the three frontal processes described tional errors than controls, but had normal latency and 784 L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease

Table 1 Studies investigating antisaccade performance in Alzheimer’s disease N Control age N Patient age Diagnostic MMSE PL PA AL AE criteria Fletcher and Sharpe (1986) [3] 11 69 (6.5) 13 69 (6.1) “Clinical 18.3 (4.1) ND* Hypo NA 74% GT Diagnosis” Currie et al. (1991) [4] 180 41 (18) 30 67 (8) NINCDS- NA NA NA NA > 30% GT ADRDA Abel et al. (2002) [5] 11 67.4 (5.4) 11 73.1 (9.4) NINCDS- 20.6 (7.6) ND NA NA 75.6% GT ADRDA Shafiq-Antonacci et al. (2003) [6] 245 62.8 (8.6) 35 70.9 (9.4) NINCDS- 17.1 (7.4) GT Hypo GT 55.9% (32.7) ND ADRDA Crawford et al. (2005) [7] 18 75.2 (3.8) 18 77.8 (4.8) NINCDS + 20.9 (4.3) ND ND ND 53.4% (23.6) GT DSM IV Mosimann et al. (2005) [8] 24 75.3 (5.8) 22 78.1 (6.8) NINCDS- 17.9 (4.7) ND ND ND 80% (42) GT ADRDA Boxer et al. (2006) [9] 20 64.4 (7.2) 18 58.4 (7.2) NINCDS- 18.7 (8.5) ND ND ND ∼ 60% GT ADRDA Garbutt et al. (2008) [10] 27 65 (1.5) 28 59.8 (1.4) NINCDS- 19.5 (5.3) GT ND GT ∼ 75% GT ADRDA MMSE = Mini-Mental Status Exam, PL = Prosaccade Latency, PA = Prosaccade Amplitude, AL = Antisaccade Latency, few studies included antisaccade amplitude, thus it was omitted from the present table, AE = Antisaccade Errors, ND = No difference, GT = Greater than, Hypo = Hypometric, NA = not applicable. amplitude [35]. A subsequent study that examined 30 of any impairment in visually-guided saccades. De- patients with subcortical lesions found that only those spite the fact that AD patients make significantly more with damage to the anterior limb of the internal capsule, antisacccade errors than age-matched controls, anti- the genu or the most anterior portion of the posterior saccades have limited diagnostic potential in differen- limb of the internal capsule had high error rates. In con- tiating patients from age-matched controls. Shafiq- trast, those not considered impaired on the task had no Antonacci and colleagues [6] found that although ro- damage to those regions; rather, their damage involved bust differences between patients with AD and controls the posterior limb of the internal capsule and parts of exist, the antisaccade task had only a modest sensitivi- the thalamus and basal ganglia [36]. Both groups of ty and specificity and concluded that “antisaccade per- patients showed similar latencies and amplitude during formance cannot identify AD in individual cases”. In the prosaccade task, but latency and amplitude were support of this conclusion, results from the most opti- not reported for the antisaccade task. mistic studies indicate that the task can only differen- The fact that there is some understanding about the tiate 40% of patients with AD from age-matched con- neural correlates of the antisaccade task coupled with its trols (Table 2) [6–8]. Moreover, groups of patients with other advantages make the antisaccade task an attractive AD show a much larger variance in the percentage of measure of inhibitory control and executive function. antisaccade errors relative to age-matched controls [3, Given that lesion and functional imaging evidence both 5,6], emphasizing that some patients are either not sig- support a critical role of the DLPFC in the antisaccade nificantly impaired or are not impaired at all. Relative task, the task may also provide a useful index of DLPFC to tests of episodic memory, antisaccades offer little integrity in the dementias, such as AD. utility for the detection of AD; however, they may be an index of DLPFC involvement and thus useful for monitoring emergence of DLPFC deficits, progression DIFFERENTIATING AD FROM HEALTHY over time and possibly response to treatment in AD. AGING

Patients with AD make significantly more antisac- ANTISACCADE ERRORS AND cade errors than controls (Table 1) and also leave many DORSOLATERAL PREFRONTAL more of their errors uncorrected [9]. During the prosac- PATHOLOGY IN AD cade task (Fig. 1), patients with AD perform mostly within normal range (Table 1), emphasizing that er- Declines in hippocampal volume and episodic mem- rors in the antisaccade task are made in the absence ory appear to be the earliest brain-behavior correla- L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease 785

Table 2 Power analysis and differentiation capabilities of the antisaccade task in Alzheimer’s disease Control errors (SD) AD errors (SD) Mean difference Cohen’s d* Shafiq-Antonacci et al. [6] 31.44 (15.36) 55.88 (32.74) 24.44 0.96 Crawford et al. [7] 18.4 (13.4) 53.4 (23.6) 35 1.83 Mosimann et al. [8] 25 (38) 80 (42) 55 1.37 Only studies which provided both mean and standard deviation (SD) values were included in this table. 2 2 *Cohen’s d = (MeanAD – MeanCon)/SQRT((SDAD +SDCON )/2).

DLPFC, critical for antisaccades, is thought to be rela- tively free of tangle pathology during the early phases of the disease, though paradoxically, amyloid-β PET labeling suggests presence of amyloid-β even in pre- clinical stages of the disease [38]. Therefore patients with early AD, who may have little or no dorsolater- al frontal pathology, may show little to no impairment on the antisaccade task. The Mini-Mental Status Ex- am (MMSE), a general measure of cognition, shows a strong correlation with antisaccade error rates: as MMSE scores decline and dementia worsens, patients make more errors [4–6,9]. However, Boxer and col- leagues [9] examined a subgroup of patients with mild AD, who had MMSE scores greater than 22 out of 30, and found that they did not make significantly more er- rors than controls, suggesting antisaccade impairments may not arise in the early stages of the disease. An absence of antisaccade impairments in patients with mild AD, followed by a gradual increase in errors with dementia severity, as indexed by the MMSE, suggests that the task may provide information on not only the progression of the disease but also the progression and magnitude of DLPFC functional impairment.

BARRIERS IN ADOPTING THE ANTISACCADE TASK AS A PROBE OF DLPFC FUNCTION

The validity of using the antisaccade task as a mea- Fig. 1. Prosaccade And Antisaccade Tasks. The arrows denotes sure of executive function and inhibitory control in pa- where participant is suppose to be looking. During the Prosaccade tients with AD is supported by its correlation with other task they fixate on the cross, then make a saccade to the peripheral measures of executive function such as the Stroop task target. During the Antisaccade task they fixate on the cross but then and the Trials A and B tasks [7,9]. However, before look away from the peripheral target. the antisaccade task can be used as a clinical index of DLPFC function in AD, there are several issues that tions in AD, mirroring the development of AD pathol- require further clarification. First, the relationship be- ogy [2]. Progression and topographical distribution of tween dementia severity and antisaccade performance neuropathology in AD has been categorized into six is based on cross-sectional between-subject data, not stages, in which AD neurofibrillary tangles first appear longitudinal within-subject data. Second, the proposed in the limbic system initially in the medial temporal relationship between DLPFC pathology and increased lobes, then gradually move into temporal-parietal asso- error rates in AD is largely based on indirect inference ciation cortices and finally the frontal regions [37]. The and would benefit further from confirmatory structural 786 L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease and functional neuroimaging. Third, a successful anti- difficult to infer whether directions were even under- saccade consists of several underlying processes, such stood or remembered. An emphasis on mild to moder- as inhibition of a prepotent saccade and generation of ate patients may reduce the likelihood of including pa- a voluntary saccade, and impairment in any of the sub- tients who fail to understand the task and could enhance processes may result in increased error rates, obfus- the understanding of the relationship between antisac- cating potential brain behavior relationships. Further- cade performance and severity of dementia. Second- more, despite the potential testing convenience provid- ly, patients in the earlier stages of AD show greater ed by clinical versions of the antisaccade task, which heterogeneity in neuropsychological impairment. For will be discussed shortly, typical saccade measurement instance, some AD patients have been categorized as relies on costly eye-monitoring equipment not avail- “frontal variant AD” as they not only perform signifi- able in most clinical environments, potentially deter- cantly worse on frontal tasks, than average AD patients, ring widespread clinical adoption. but also show greater frontal atrophy [38]. Inclusion Finally, each study in Table 1 has utilized slightly dif- of only a small number of mild AD patients may have ferent temporal (gap, step, or overlap) and spatial vari- failed to capture variations in frontal deficits potentially ants (angle of peripheral target offset) of the antisac- underestimating error rates in mild AD. cade task, which complicates direct comparison. The three main temporal variants of the antisaccade task, the gap, step, and overlap refer to the temporal char- NEUROIMAGING AND DEMENTIA acteristics between the central fixation point and the peripheral target. The gap variant, used by Boxer and The relationship between AD neurofibrillary tangle colleagues [7], typically includes a temporal gap of 200 pathology distribution and antisaccade deficits is large- ms between the disappearance of the central fixation ly based on an assumption that those who are less de- point and the appearance of the peripheral target. In mented, and make fewer errors, may have less pathol- the overlap condition, the fixation target overlaps tem- ogy in the DLPFC. Boxer and colleagues [9] used porally with the peripheral target, while the step variant voxel-based morphometry to examine the association is a compromise between the overlap and gap variants, between brain atrophy and antisaccade performance in in which the central fixation point disappears simulta- patients with AD and those with FTD. They found that neously with the appearance of the peripheral target. the volume of an area located ventrally to the right Additionally, many studies fail to explicitly report the frontal eye fields was correlated with correct responses, parameters they have used, making direct comparison while supplementary eye fields volume was correlated difficult. Despite the variation and opacity in reported with antisaccade latency. Although these findings re- antisaccade parameters, the overall findings from the inforce the relationship between the dorsofrontal cor- previously mentioned studies are consistent with each tex and antisaccade performance, they failed to find other: patients with AD make significantly more errors a relationship between antisaccade performance with than controls. either DLPFC or frontal eye field volumes, which, as stated earlier, are regions critical for inhibition of prosaccades and generation of antisaccades respective- THE ANTISACCADE TASK AS AN INDEX OF ly. By including both FTD and AD patients in a single SEVERITY group, despite the heterogeneity in the distribution and type of pathology within FTD and AD [40,41], Boxer The relationship between MMSE scores and antisac- and colleagues may have prevented detection of cor- cade error rates supports the notion that error rates may relations between structure and performance [9]. For mirror disease progression, but are 1) based on cross- instance, high variability in antisaccade performance, sectional comparisons and 2) under-represent those in coupled with the low sensitivity of antisaccades [6], the earliest stages of AD. Most existing studies have indicates many patients are unimpaired and may lack focused on patients in the moderate to severe stages of sufficient DLPFC pathology and atrophy, thus elim- AD (see Table 1), which may exaggerate differences inating any significant correlations between structure between AD and controls and may strengthen the cor- and function. Furthermore, Boxer and colleagues cor- relation between MMSE scores and error rates. For rected for MMSE scores, which are strongly correlated instance, patients on the lowest end of the MMSE spec- with error rates in AD; controlling for MMSE scores trum tend to make 100% uncorrected errors making it may have inadvertently eliminated any relationship be- L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease 787 tween error rates and DLPFC or frontal eye field vol- in the anterior cingulate and DLPFC in schizophrenia, umes. Utilizing methodology from antisaccade focal a psychopathology associated with increased antisac- lesion studies [34], future investigations should divide cade errors [45]. Thus, antisaccade imaging studies patients with AD into two groups: 1) those whose per- conducted in healthy adults and other patient groups, centage of errors fall within normal range and 2) those such as schizophrenia, could provide an experimental who make significantly more errors than normal. Such framework that could be repeated in AD. Functional a division would be of interest in the middle stage of imaging results coupled with structural neuroimaging, disease, as error rates would be common in moderate in AD, could provide more direct in vivo evidence for to severe-stage disease. On the milder cases, however, a correlation between DLPFC changes and increased grouping patients by the differences in error rates could error rates. reveal whether this correlates with regional differences in atrophy, perfusion or white matter integrity in ar- eas believed to be critical for antisaccade performance. FRACTIONATION OF PROCESSES IN THE Furthermore, could differences between high and low ANTISACCADE TASK error AD patients could be correlated with neuropsy- chological deficits in executive function, more specifi- The antisaccade task is comprised of multiple sub- cally with inhibitory control? Ultimately such a design processes that contribute to a successful execution [12]. would strengthen brain behavior correlations by reduc- A participant must be able to fixate on the central fix- ing the heterogeneity of performance and presumably ation point, then inhibit a reflexive saccade, invert the pathology distribution within AD groups. saccade vector to a non-existent target, and make a vol- Diffusion tensor imaging, a technique for examining untary saccade in the direction of the updated vector. white matter tissue microstructure, has revealed differ- In addition to the processes directly related to the task, ences in patients with AD relative to controls. Specif- there are secondary processes that are also critical for ically, the superior longitudinal fasciculus, a bundle of successful antisaccades. For instance, a patient must fibers that connects posterior and frontal regions of the be able to understand and remember the task’s direc- brain and also includes connections to the DLPFC, was tions and vigilantly attend to the task. According to the reported to show a decrease in the fractional anisotropy accumulator model, the process that reaches threshold (FA) [41]. FA is a measure of white matter cohesion first, either the antisaccade or the erroneous prosaccade, and integrity, so a decrease is indicative of underlying determines which behavior is initiated [12]. For in- structural aberrations that may elucidate brain-behavior stance, if the antisaccade process reaches threshold be- correlations. Correlating error rates and diffusion ten- fore an erroneous prosaccade, an antisaccade is carried sor metrics such as FA would provide insight into the out. Deficits in any of the above-mentioned processes potential role white matter injury may have on antisac- could result in a slowing of the antisaccade process, cade performance in AD. increasing the chance that an erroneous prosaccade is Functional imaging techniques such as single photon generated. In determining whether antisaccade errors emission computerized tomography (SPECT), a semi- in AD result from inhibitory deficits and ultimately quantitative measure of regional cerebral blood flow, DLPFC dysfunction, it will be critical to fractionate the and fMRI have provided important insights into the re- underlying sub-processes to insure other components lationship between memory and brain function in AD impaired in AD, such as memory, are not contributing and may provide further insight into antisaccade errors. significantly to antisaccade errors. Using SPECT, Garrido and colleagues found that pa- tients with AD showed decreased cerebral blood flow Fractioning the antisaccade task: inhibition control in the left medial temporal lobes, relative to controls during a verbal recognition memory task [43]. Like- Fractionation of inhibition from the other processes wise Grady and colleagues used fMRI to study com- can be accomplished through additional saccade task pensatory frontal network activity during memory tasks manipulations. Two tasks which focus on inhibition in mild AD [44]. The relationship found between func- are: 1) the no-go task, in which the participant main- tional aberrations and poor memory might be mirrored tains fixation while peripheral targets appear, and 2) in a relationship between increased antisaccade errors the go no-go task, in which the participant must fix- and abnormal DLPFC activation patterns. For instance, ate during some trials (no-go trials) and make saccades fMRI and PET have revealed decreases in activation during other trials (go trials). No-go and go no-go tasks 788 L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease are easier than the antisaccade tasks because they do Fractioning the antisaccade task: voluntary saccades not require a vector inversion or volitional generation of a saccade: they simply require inhibition in a pro- A voluntary saccade to a non-existent target differs portion of the trials. Crawford and co-workers found both neurologically and in difficulty from the reflexive that patients with AD made significantly more inhibi- saccades generated in the prosaccade task. While the tion errors on both tasks, when compared with controls, frontal eye fields are involved in generating voluntary suggesting that inhibitory deficits, not deficits in vec- saccades, a region in the parietal lobes is implicated tor inversion or volitional control, may be the greatest in generating reflexive visually guided saccades [12]. contributor to antisaccade errors in AD [7]. When patients with frontal eye field focal lesions make antisaccades, they typically show longer antisaccade Fractioning the antisaccade task: fixation latency and depending on the study, error rates remain normal [51] or are elevated [52]. In contrast, patients Typical antisaccade coding schemes require a partic- with AD do not consistently show longer latencies, but ipant to fixate centrally for at least 200 ms prior to a do consistently make more errors. Furthermore, the successful antisaccade; trials that fail to meet this re- frontal eye fields are one of the last regions to be affect- quirement are eliminated from analysis [6]. Reported ed by AD tangle pathology making it further unlikely fixation deficits in patients with AD [3] may result in that voluntary saccade generation is impaired or is the greater numbers of eliminated trials in the AD group, cause of increased error rates in AD. resulting in fewer analyzable trials. Presenting a static image with several targets and asking the participant to maintain fixation on a single target for a specified time MEMORY, UNDERSTANDING, AND provides a parsimonious method for testing fixation. ATTENTION The task could be repeated using different static images and different fixation points. A failure to make correct antisaccades could result from deficits in memory, understanding, and/or atten- Fractioning the antisaccade task: vector inversion tion. Despite significant impairments in each of those domains in AD, studies have posited that those impair- Successful antisaccades require a vector inversion: ments do not contribute to increased errors because of the prosaccade vector must be inverted to a saccade in two factors. First, when performance on an antisaccade the opposite direction towards a non-existent target, a block is divided into two halves, performance is sta- visuospatial process which imaging and lesion studies ble between the two halves; indicating patients are not have suggested is mediated by the posterior parietal cor- progressively forgetting task instructions [7]. Second, tex [46,47]. In the early stages of the clinical presenta- patients usually generate at least one correct antisac- tion of AD, neurofibrillary tangle pathology is found in cade, or one corrected antisaccade, a reaction not seen the posterior parietal lobe, suggesting a possible associ- in prosaccades. This suggests that patients remem- ation between visuospatial/vector inversion deficits and bered the task well enough to make at least one antisac- brain pathology. Although not typically the present- cade [3]. However, just as inhibitory control is consid- ing symptom, spatial and visuomotor deficits are often ered impaired in patients with DLPFC lesions, despite detectable in the mild to moderate stages of AD [47– generating a few correct responses, patients with AD 50]. Correlations of performance of antisaccade tasks may still make errors due to poor short-term memory. have been reported with two visuospatial tasks: figure Neuropsychological indices of memory, such as verbal copying with correct antisaccades [9] and spatial span episodic memory [9] and memory quotient [4], cor- with uncorrected errors [7]. Despite these correlations relate with antisaccade performance, suggesting that of visuospatial deficits with antisaccade errors, it is un- impairments in memory may partly contribute to anti- clear if they are contributing factors to the increased saccade errors. Perhaps the most pertinent question is error rates present in AD. For instance, error rates are whether patients not only remember and understand the significantly elevated during the no-go task, which on- task’s instructions immediately before the task begins, ly includes the inhibitory component of the antisac- but also once it has ended. A patient’s understanding cade task, not the visuospatial component. Identifying of the task is often determined by having the patient whether visual spatial deficits contribute to increased either point to where they are suppose to look, or verbal error rates in AD requires further investigation. repetition of the instructions [7]. Repeating these steps L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease 789 after the task is completed would not only test memory ror rates than the laboratory version, scores were highly but would also test a patient’s understanding of the task, correlated with those generated in the laboratory ver- but this has yet to be reported in any published study. sion (r = 0.921). The clinical version is also a compo- Although deficits in attention have been documented nent of the HIV Dementia scale that has been validated in patients with AD [53], their effect on antisaccade in both patients with HIV [57] and patients with subcor- performance has not been explored. If impairments in tical vascular cognitive impairment [58]. The clinical memory, understanding, or attention lead to increased variant of the antisaccade task provides a parsimonious errors, there would be little utility in using the anti- method for testing and overcomes the shortcomings of saccade task as a measure of disease progression as the laboratory version. However, it is unclear how dif- memory tasks would suffice. If either impairments in ficult it is for a clinician to remember a sequence of 10 memory, understanding, or attention contribute to in- to 20 trials while keeping track of how many trials have creased error rates, a link between the DLPFC, inhibito- been administered and how many errors a patient has ry control and error rates in AD would be difficult to made. The clinical version, in its present form, limits infer. Thus, excluding these processes and behavioral the clinician to record only errors in direction, neglect- domains as major contributors to antisaccade error rates ing other metrics such as uncorrected errors, fixation should remain a priority for future studies. errors, errors of omission, latency, and amplitude partly because of the cognitive load of administering the task, but also because latency and amplitude can only be reli- TASK SEQUENCE ably recording with eye-tracking equipment. Although the reliability of the clinical version, between differ- Statistical analysis of antisaccade data has primarily ent clinical centers, has not been tested, the laboratory relied on univariate models that assume performance version has been tested in patients with schizophrenia, on a single trial is independent from other trials. For revealing that the task can be carried out reliably in example, these models assume that performance on a different centers [59]. trial (n+ 1) is in no way affected by the previous trial (n). However, this assumption appears false, as data from normal controls indicates performance on a cur- NON-ALZHEIMER’S DEMENTIA rent trial is influenced by the direction of the previous trial, relative to the direction of the current trial [54– 56]. For instance, if a peripheral target appears on the The majority of studies examining antisaccade per- left side for two trials in a row, the second trial would formance in patients with dementia have focused on be categorized as “same” and would be associated with AD. However, antisaccade performance has been ex- fewer errors than a “different” trial (a trial preceding by amined in other types of dementia (Table 3) [9,10,61– a trial of a different direction, i.e., left then right) [54]. 63]. A pattern of inhibitory impairment, as measured However, this effect may not be consistent across age by antisaccade errors, seems to mirror the distribu- and patient groups, potentially confoundingdifferences tion of pathology in frontotemporal dementia, seman- between patients with AD and elderly controls. tic dementia and progressive non-fluent aphasia. Fron- totemporal dementia and progressive non-fluent apha- sia, characterized by deficits in behavioral regulation CLINICAL ADAPTATION and non-fluent speech respectively are both associated with pathology in the frontal lobes [60]. In contrast, the Although clinical variations of the antisaccade task core diagnostic feature of semantic dementia is a deficit are easy to administer, typical antisaccade experiments in word comprehension and it is associated with pathol- use sophisticated eye-tracking labs that are often costly ogy in the anterior temporal lobes. The difference in to establish, lack portability, and use techniques that the presence of frontal pathology between these groups require calibration, making the task clinically less ap- is mirrored by their performance on the antisaccade pealing. In an effort to avoid these shortcomings, Cur- task: frontotemporal dementia and progressive aphasia rie and colleagues developed a clinical version of the are associated with high antisaccade error rates, while antisaccade task that uses the clinician’s nose as the semantic dementia is not [9]. Interestingly, all three central fixation and fingers as the peripheral targets [4]. of these groups correct more errors than patients with Although the clinical variant yielded slightly lower er- AD [9]. 790 L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease

Table 3 Studies investigating antisaccade performance in non-Alzheimer’s dementia N Control age N Patient age Diagnostic criteria MMSE PL PA AL AE FTD Meyniel et al. (2005) [61] 10 68 (9) 23 67 (9) Lund-Manchester 26.1 (2.8) GT NA GT 63% GT Boxer et al. (2006) [9] 20 64.4 (7.2) 14 59.9 (5.6) Neary et al. (1998) 25.7 (3.7) ND ND ND 60% GT Garbutt et al. (2008) [10] 27 65 (1.5) 24 57.4 (1.7) Neary et al. (1998) 23.5 (7.5) NA Hypo ND ∼ 55% GT PNFA Boxer et al. (2006) [9] 20 64.4 (7.2) 7 65.7 (7.9) Neary et al. (1998) 23.1 (5.3) ND ND ND ∼ 70% GT Garbutt et al. (2008) [10] 27 65 (1.5) 6 64.5 (3.0) Neary et al. (1998) 25.2 (3.5) ND ND ND ∼ 52% GT SD Boxer et al. (2006) [9] 20 64.4 (7.2) 10 60.3 Neary et al. (1998) 20.1 (7.6) ND ND ND ∼ 25% ND Garbutt et al. (2008) [10] 27 65 (1.5) 19 60.3 (1.3) Neary et al. (1998) 21.7 (7.3) ND ND ND ∼ 28% ND PSP Vidailhet et al. (1994) [62] 12 63.9 (8.3) 10 62.5 (5.5) NA NA ND Hypo NA ∼ 73.5 GT Meyniel et al. (2005) [61] 10 68 (9) 14 70 (6) Litvan’s 28.6 (2.0) ND NA GT 70% GT Rivaud-Pechoux et al. (2007) [63] 10 64 (9) 12 66 (9.9) Litvan’s NA ND NA GT GT Garbutt et al. (2008) [10] 27 65 (1.5) 10 65.5 (1.3) Litvan et al. (1996) 26.8 (2.6) GT NA NA ∼ 92% GT CBD Vidailhet et al. (1994) [62] 12 63.9 (8.3) 10 66.5 (6.8) NA NA GT ND NA ∼ 37.5 ND Rivaud-Pechoux et al. (2007) [63] 10 64 (9) 8 76 (5.4) Litvan’s 1997 NA GT NA GT ND Garbutt et al. (2008) [10] 27 65 (1.5) 15 62.7 (2.0) Several methods 19.8 (7.7) GT Hypo NA ∼ 75% GT MMSE = Mini-Mental Status Exam, PL = Prosaccade Latency, PA = Prosaccade Amplitude, AL = Antisaccade Latency, few studies included antisaccade amplitude, thus it was omitted from the present table, AE = Antisaccade Errors, ND = No difference, GT = Greater than, Hypo = Hypometric, NA = not applicable.

Although corticobasal syndrome and progressive stages, when the drugs will be the most effective, it is supranuclear palsy are primarily characterized by mo- unlikely that the antisaccade task will provide greater tor impairments such as asymmetric extrapyramidal diagnostic utility than existing tests for AD. However, signs and vertical gaze palsy respectively, they are al- in addition to its potential role as a probe of dorsolateral so associated with frontal deficits [64]. The clinical function and as a test to monitor treatment response, it criteria for diagnosis of progressive supranuclear palsy may aid in differentiating other forms of dementia. is highly predictive of autopsy findings, whereas cor- ticobasal syndrome is less specific and can be associ- ated with either corticobasal degeneration, progressive CONCLUSION supranuclear palsy pathology, both overlapping [40] or even the tau negative ubiquitin positive pathology [65]. The neural correlates of antisaccades continue to be Patients with corticobasal syndrome and controls made mapped and reported using a variety of neuroimag- an equal number of errors on the antisaccade task, while ing techniques providing further insights into brain- patients with progressive supranuclear palsy made sig- behavior correlations of this simple task. Antisaccades nificantly more errors than controls [61,62]. How- provide a well tolerated, language-free and hands-free ever, when patients with corticobasal syndrome com- neuropsychological probe that may not only help those pleted a mixed block of prosaccades and antisaccades, with AD, but could be especially helpful in testing they made many more errors than controls compared patients with expressive language problems or motor with patients with progressive supranuclear palsy who deficits, such as those with progressive aphasia and showed no difference between mixed and non-mixed amyotrophic lateral sclerosis. blocks [63]. Differences in antisaccade performance Despite having limited utility in differentiating indi- between the two groups may provide information to viduals with AD from normal aging, the available ev- help distinguish these diseases. idence indicates that the task may provide insight into frontal lobe function and an index of DLPFC pathology in AD. The potential utility of the antisaccade task as DISCUSSION a neuropsychological probe of DLPFC and ultimately for progression and for treatment monitoring appears Pressure to increase the diagnostic accuracy for de- promising but requires further investigation. The re- mentia, and specifically AD, is mounting due to the lationship between the DLPFC, inhibitory control and availability of new potential treatments. In the early errors rates in AD requires exclusion of other potential L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease 791 contributors, such as memory impairments, and would [12] Munoz DP, Everling S (2004). Look away: the anti-saccade benefit from converging evidence from multiple neu- task and the voluntary control of eye movement. Nat Rev Neurosci 5, 218-228. roimaging modalities. This overview has outlined fu- [13] Klein C, Foerster F, Hartnegg K, Fischer B (2005) Lifespan ture avenues of research for testing the link between development of pro- and anti-saccades: multiple regression inhibitory deficits and DLPFC changes in patients with models for point estimates. Brain Res Dev Brain Res 160, AD. 113-123. [14] Munoz DP, Broughton JR, Goldring JE, Armstrong IT (1998) Age-related performance of human subjects on saccadic eye movement tasks. Exp Brain Res 121, 391-400. ACKNOWLEDGMENTS [15] Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A, Paus T, Evans AC, Rapoport JL (1999) Brain de- velopment during childhood and adolescence: a longitudinal LDK’s funding provided by Ontario Graduate Schol- MRI study. Nat Neurosci 2, 861-863. arship and Scace Graduate Fellowship in Alzheimer’s [16] Barnea-Goraly N, Menon V, Eckert M, Tamm L, Bammer Research (OSOTF). R, Karchemskiy A, Dant CC, Reiss AL (2005) White mat- ter development during childhood and adolescence: a cross- Authors’ disclosures available online (http://www.j- sectional diffusion tensor imaging study. Cereb Cortex 15, alz.com/disclosures/view.php?id=161). 1848-1854. [17] Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, Frackowiak RS (2001) A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 14, REFERENCES 21-36. [18] Pfefferbaum A, Sullivan EV (2003) Increased brain white mat- [1] Ferri CP, Prince M, Brayne C, Brodaty H, Fratiglioni L, Gan- ter diffusivity in normal adult aging: relationship to anisotropy guli M (2005) Global prevalence of dementia: a Delphi con- and partial voluming. Magn Reson Med 49, 953-961. sensus study. Lancet 366, 2112-2117. [19] Raemaekers M, Vink M, van den Heuvel MP, Kahn RS, Ram- [2] Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger- sey NF (2006) Effects of aging on BOLD fMRI during prosac- Gateau P, Cummings J, Delacourte A, Galasko D, Gauthier S, cades and antisaccades. J Cogn Neurosci 18, 594-603. Jicha G, Meguro K, O’brien J, Pasquier F, Robert P, Rossor M, [20] Abel LA, Douglas J (2007) Effects of age on latency and error Salloway S, Stern Y, Visser PJ, Scheltens P (2007) Research generation in internally mediated saccades. Neurobiol Aging criteria for the diagnosis of Alzheimer’s disease: revising the 28, 627-637. NINCDS-ADRDA criteria. Lancet Neurol 6, 734-746. [21] Olincy A, Ross RG, Youngd DA, Freedman R (1997) Age [3] Fletcher WA, Sharpe JA (1986) Saccadic eye movement dys- diminishes performance on an antisaccade eye movement task. function in Alzheimer’s disease. Ann Neurol 20, 464-471. Neurobiol Aging 18, 483-489. [4] Currie J, Ramsden B, McArthur C, Maruff P (1991) Validation [22] Kimmig H, Greenlee MW, Gondan M, Schira M, Kassubek J, of a clinical antisaccadic eye movement test in the assessment Mergner T (2001) Relationship between saccadic eye move- of dementia. Arch Neurol 48, 644-648. ments and cortical activity as measured by fMRI: quantitative [5] Abel LA, Unverzagt F, Yee RD (2002) Effects of stimulus pre- and qualitative aspects. Exp Brain Res 141, 184-194. dictability and interstimulus gap on saccades in Alzheimer’s [23] O’Driscoll GA, Alpert NM, Matthysse SW, Levy DL, Rauch disease. Dement Geriatr Cogn Disord 13, 235-243. SL, Holzman PS (1995) Functional neuroanatomy of anti- [6] Shafiq-Antonacci R, Maruff P, Masters C, Currie J (2003) saccade eye movements investigated with positron emission Spectrum of saccade system function in Alzheimer disease. tomography. Proc Natl Acad Sci U S A 92, 925-929. Arch Neurol 60, 1272-1278. [24] DeSouza JF, Menon RS, Everling S (2003) Preparatory set [7] Crawford TJ, Higham S, Renvoize T, Patel J, Dale M, Suriya associated with pro-saccades and anti-saccades in humans in- A, Tetley S (2005) Inhibitory control of saccadic eye move- vestigated with event-related FMRI. J Neurophysiol 89, 1016- ments and cognitive impairment in Alzheimer’s disease. Biol 1023. Psychiatry 57, 1052-1060. [25] McDowell JE, Clementz BA (2001) Behavioral and brain [8] Mosimann UP, Muri RM, Burn DJ, Felblinger J, O’Brien JT, imaging studies of saccadic performance in schizophrenia. Bi- McKeith IG (2005) Saccadic eye movement changes in Parkin- ol Psychol 57, 5-22. son’s disease dementia and dementia with Lewy bodies. Brain [26] Connolly JD, Goodale MA, Desouza JF, Menon RS, Vilis T 128, 1267-1276. (2000) A comparison of frontoparietal fMRI activation dur- [9] Boxer AL, Garbutt S, Rankin KP, Hellmuth J, Neuhaus J, ing anti-saccades and anti-pointing. J Neurophysiol 84, 1645- Miller BL, Lisberger SG (2006) Medial versus lateral frontal 1655. lobe contributions to voluntary saccade control as revealed by [27] McDowell JE, Kissler JM, Berg P, Dyckman KA, Gao the study of patients with frontal lobe degeneration. J Neurosci Y, Rockstroh B, Clementz BA (2005) Electroencephalogra- 26, 6354-6363. phy/magnetoencephalography study of cortical activities pre- [10] Garbutt S, Matlin A, Hellmuth J, Schenk AK, Johnson JK, ceding prosaccades and antisaccades. Neuroreport 16, 663- Rosen H, Dean D, Kramer J, Neuhaus J, Miller BL, Lisberger 668. SG, Boxer AL (2008) Oculomotor function in frontotemporal [28] Raemaekers M, Vink M, van den Heuvel MP, Kahn RS, Ram- lobar degeneration, related disorders and Alzheimer’s disease. sey NF (2005) Brain activation related to retrosaccades in sac- Brain 131, 1268-1281. cade experiments. Neuroreport 16, 1043-1047. [11] Hallett PE (1978) Primary and secondary saccades to goals [29] Everling S, Spantekow A, Krappmann P, Flohr H (1998) defined by instructions. Vision Res 18, 1279-1296. Event-related potentials associated with correct and incorrect 792 L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease

responses in a cued antisaccade task. Exp Brain Res 118, 27- [47] Nyffeler T, Rivaud-Pechoux S, Pierrot-Deseilligny C, Diallo 34. R, Gaymard B (2007) Visual vector inversion in the posterior [30] Ford KA, Goltz HC, Brown MR, Everling S (2005) Neural parietal cortex. Neuroreport 18, 917-920. processes associated with antisaccade task performance inves- [48] Buck BH, Black SE, Behrmann M, Caldwell C, Bronskill tigated with event-related FMRI. J Neurophysiol 94, 429-440. MJ (1997) Spatial- and object-based attentional deficits in [31] Stuss DT, Alexander MP (2007) Is there a dysexecutive syn- Alzheimer’s disease. Relationship to HMPAO-SPECT mea- drome? Philos Trans R Soc Lond B Biol Sci 362, 901-915. sures of parietal perfusion. Brain 120, 1229-44. [32] Guitton D, Buchtel HA, Douglas RM. (1985) Frontal lobe le- [49] Cummings JL, Cole G (2002) Alzheimer disease. JAMA 287, sions in man cause difficulties in suppressing reflexive glances 2335-2338. and in generating goal-directed saccades. Exp Brain Res 58, [50] Tippett WJ, Sergio LE (2006) Visuomotor integration is im- 455-472. paired in early stage Alzheimer’s disease. Brain Res 1102, [33] Pierrot-Deseilligny C, Rivaud S, Gaymard B, Agid Y (1991) 92-102. Cortical control of reflexive visually-guided saccades. Brain [51] Gaymard B, Ploner CJ, Rivaud-Pechoux S, Pierrot-Deseilligny 114, 1473-1485. C (1999) The frontal eye field is involved in spatial short-term [34] Ploner CJ, Gaymard BM, Rivaud-Pechoux S, Pierrot- memory but not in reflexive saccade inhibition. Exp Brain Res Deseilligny C (2005) The prefrontal substrate of reflexive sac- 129, 288-301. cade inhibition in humans. Biol Psychiatry 57, 1159-1165. [52] Machado L, Rafal RD (2004) Control of fixation and saccades [35] Gaymard B, Francois C, Ploner CJ, Condy C, Rivaud-Pechoux during an anti-saccade task: an investigation in humans with S (2003) A direct prefrontotectal tract against distractibility in chronic lesions of oculomotor cortex. Exp Brain Res 156, 55- the human brain. Ann Neurol 53, 542-545. 63. [36] Condy C, Rivaud-Pechoux S, Ostendorf F, Ploner CJ, Gay- [53] Perry RJ, Hodges JR (1999) Attention and executive deficits mard B (2004) Neural substrate of antisaccades: role of sub- in Alzheimer’s disease. A critical review. Brain 122, 383-404. cortical structures. Neurology 63, 1571-1578. [54] Barton JJ, Goff DC, Manoach DS (2006) The inter-trial effects [37] Braak H, Braak E (1991) Neuropathological stageing of of stimulus and saccadic direction on prosaccades and antisac- Alzheimer-related changes. Acta Neuropathol (Berl) 82, 239- cades, in controls and schizophrenia patients. Exp Brain Res 259. 174, 487-1498. [38] Pike KE, Savage G, Villemagne VL, Ng S, Moss SA, Maruff [55] Reuter B, Philipp AM, Kathmann N (2006) Effects of switch- P, Mathis CA, Klunk WE, Masters CL, Rowe CC (2007) Beta- ing between leftward and rightward pro- and antisaccades. amyloid imaging and memory in non-demented individuals: Biol Psychol 72, 88-95 evidence for preclinical Alzheimer’s disease. Brain 130, 2837- [56] Tatler BW, Hutton SB (2007) Trial by trial effects in the anti- 2844. saccade task. Exp brain res 179, 387-96. [39] Johnson JK, Head E, Kim R, Starr A, Cotman CW (1999) [57] Power C, Selnes OA, Grim JA, McArthur JC (1995) HIV Clinical and pathological evidence for a frontal variant of Dementia Scale: a rapid screening test. J Acquir Immune Defic Alzheimer disease. Arch Neurol 56, 1233-1239. Syndr Hum Retrovirol 8, 273-278. [40] Josephs KA, Petersen RC, Knopman DS, Boeve BF, Whitwell [58] van Harten B, Courant MN, Scheltens P, Weinstein HC (2004) JL, Duffy JR, Parisi JE, Dickson DW (2006) Clinicopathologic Validation of the HIV Dementia Scale in an elderly cohort analysis of frontotemporal and corticobasal degenerations and of patients with subcortical cognitive impairment caused by PSP. Neurology 66, 41-48. subcortical ischaemic vascular disease or a normal pressure [41] Snowden J, Neary D, Mann D (2007) Frontotemporal lobar hydrocephalus. Dement Geriatr Cogn Disord 18, 109-114. degeneration: clinical and pathological relationships. Acta [59] Radant AD, Dobie DJ, Calkins ME, Olincy A, Braff DL, Ca- Neuropathol (Berl) 114, 31-38. denhead KS, Freedman R, Green MF, Greenwood TA, Gur RE, [42] Medina D, DeToledo-Morrell L, Urresta F, Gabrieli JD, Mose- Light GA, Meichle SP, Mintz J, Nuechterlein KH, Schork NJ, ley M, Fleischman D, Bennett DA, Leurgans S, Turner DA, Seidman LJ, Siever LJ, Silverman JM, Stone WS, Swerdlow Stebbins GT (2006) White matter changes in mild cognitive NR, Tsuang MT, Turetsky BI, Tsuang DW (2007) Successful impairment and AD: A diffusion tensor imaging study. Neu- multi-site measurement of antisaccade performance deficits in robiol Aging 27, 663-672. schizophrenia. Schizophr Res 89, 320-329. [43] Garrido GE, Furuie SS, Buchpiguel CA, Bottino CM, Almeida [60] Neary D, Snowden JS, Gustafson L, Passant U, Stuss D, Black OP,Cid CG, Camargo CH, Castro CC, Glabus MF, Busatto GF S, Freedman M, Kertesz A, Robert PH, Albert M, Boone K, (2002) Relation between medial temporal atrophy and func- Miller BL, Cummings J, Benson DF (1998) Frontotemporal tional brain activity during memory processing in Alzheimer’s lobar degeneration: a consensus on clinical diagnostic criteria. disease: a combined MRI and SPECT study. J Neurol Neuro- Neurology 51, 1546-1554. surg Psychiatry 73, 508-516. [61] Meyniel C, Rivaud-Pechoux S, Damier P, Gaymard B (2005) [44] Grady CL, McIntosh AR, Beig S, Keightley ML, Burian H, Saccade impairments in patients with fronto-temporal demen- Black SE (2003) Evidence from functional neuroimaging of tia. J Neurol Neurosurg Psychiatry 76, 1581-1584. a compensatory prefrontal network in Alzheimer’s disease. J [62] Vidailhet M, Rivaud S, Gouider-Khouja N, Pillon B, Bonnet Neurosci 23, 986-993. AM, Gaymard B, Agid Y, Pierrot-Deseilligny C (1994) Eye [45] Hutton SB, Ettinger U (2006) The antisaccade task as a re- movements in parkinsonian syndromes. Ann Neurol 35, 420- search tool in psychopathology: a critical review. Psychophys- 426. iology 43, 302-313. [63] Rivaud-Pechoux S, Vidailhet M, Brandel JP, Gaymard B [46] Medendorp WP, Goltz HC, Vilis T (2005) Remapping the re- (2007) Mixing pro- and antisaccades in patients with parkin- membered target location for anti-saccades in human posterior sonian syndromes. Brain 130, 256-264. parietal cortex. J Neurophysiol 94, 734-740. [64] Kertesz A, Martinez-Lage P, Davidson W, Munoz DG (2000) The corticobasal degeneration syndrome overlaps progressive L.D. Kaufman et al. / Antisaccades: A Probe into the Dorsolateral Prefrontal Cortex in Alzheimer’s Disease 793

aphasia and frontotemporal dementia. Neurology 55, 1368- S, Rogaeva E (2006) Novel splicing mutation in the progran- 1375. ulin gene causing familial corticobasal syndrome. Brain 129, [65] Masellis M, Momeni P, Meschino W, Heffner R Jr, Elder J, 3115-3123. Sato C, Liang Y,St George-Hyslop P, Hardy J, Bilbao J, Black