Perception and the Medial : The Role of the Perirhinal Cortex

Lyssa Manning

Integrated Program in Neuroscience

McGill University, Montreal

August, 2017

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Master of Science

© Lyssa Manning, August 2017 and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 2

Table of Contents Abstract (English)...... p. 4 Résumé...... p. 4 Acknowledgements...... p. 6 Contribution of Authors...... p. 6 1. Introduction...... p. 7 1.1. Medial temporal lobe involvement in declarative ...... p. 8 1.2. The region-specific role of perirhinal cortex...... p. 11 1.3. Objectives……………………………………………………………………………………………………………………p. 16 1.4. Hypotheses………………………………………………………………………………………..…...………………….p. 17 1.5. Specific Aims……………………………………………………………………………………………………………….p. 17 2. Methods...... p. 17 2.1. Initial participant recruitment...... p. 17 2.2. Behavioural session...... p. 18 2.2.1 Neuropsychological Assessment…………………………………………………………………..p. 19 2.3. MRI data acquisition...... p. 20 2.4. Behavioural analysis…………………………………………………………………………………………………...p. 22 2.5. Image processing and analysis...... p. 23 2.6. fMRI data analysis……………………………………………………………………………………………………...p. 23 3. Results...... p. 27 3.1. Behavioural results...... p. 27 3.2. fMRI results...... p. 27 3.2.1 Task PLS results...... p. 27

3.2.1.1 ROI-based analysis of PRc regions identified in PLS analysis…………p. 28 3.2.2 Seed PLS results……………………………………………………………………………………………p. 28

3.2.2.1 Salient ROIs in Seed PLS………………………………………………………………..p. 29 4. Discussion...... p. 29 4.1. Task dependent activation patterns…………………………………………..……………………………………..…...p. 30 4.2. Square discrimination network……………………………………………………………………………………………..…...p. 30 4.3. Blob discrimination network……………………………………………….…………………………………….………………..p. 31 4.4. The PRc and non-visual conjunctive representations……………………………………………………………….p.33 4.5 Strengths and limitations...... p. 34 4.6 Conclusions…………………………………………………………………………………………………..………………………….…p. 34 5. References...... p. 36 6. Tables...... p. 45 Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 3

Table 1. Participant means for age, education, and neuropsychological tests………………………………..….p.45 Table 2. Mean accuracy (Acc) and reaction time (RT) in scanned tasks...... p. 45 Table 3. Local maxima for TPLS LV 1: blobs versus squares main effect...... p. 46 Table 4. Local maxima for Seed BPLS LV 1: Square accuracy network correlations with PRc…………...…p. 46 Table 5. Local maxima for Seed BPLS LV 2: Blob accuracy network correlations with PRc…………….…...p. 47 7. Figures...... p. 47 Figure 1. Sample task stimuli...... p. 47 Figure 2. Paired samples t-test comparing mean blobs and squares accuracy...... p. 48 Figure 3. Task specific blob (blue) vs squares (red) network….…………………………………………………………p. 49 Figure 4. Bilateral PRc activation during blob trials...... p. 50 Figure 5. Square accuracy network (Seed PLS LV1)………………………………………………………………………..p. 51 Figure 6. Blob accuracy network (Seed PLS LV2)…………………………………………………………………………….p. 52 Figure 7. Simplified network schematics……………………………………………………………...... p. 53

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 4

Abstract

Substructures of the medial temporal lobe (MTL), including the perirhinal cortex (PRc), have long been considered to contribute exclusively to learning and declarative memory. However, recent research has found evidence of PRc involvement in complex perceptual discrimination tasks requiring a feature conjunctive approach. Few studies have directly tested this in healthy young adults and only one study has attempted to asses PRc activation and it’s relationship to performance of these tasks in a small sample (N = 11). The current functional magnetic resonance imaging (fMRI) study uses a larger sample (N=24) of healthy young adults to investigate brain activity during performance of a complex visual discrimination task and uses seed partial least squares (PLS) analysis to relate activation in PRc to performance and activity in the rest of the brain. Task-based PLS identified distinct sets of brain regions involved in a complex feature conjunctive discrimination task vs a simple size comparison control task, with PRc ROIs showing significantly more activation for feature conjunctive discriminations than for control discriminations. Additionally, seed PLS using the PRc

ROIs identified in our task PLS analysis identified two key networks. The first network supported accuracy on the control task and involved bilateral PRc and posterior visual regions (BA 17, BA 18).

The second network supported accuracy on the conjunctive discrimination task and involved left PRc and frontal regions (BA 9, BA 47). This study provides evidence for distributed brain network involvement in this task and identifies specific networks involving right and left PRc both jointly and independently that support accuracy on this perceptual task.

Résumé Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 5

Les sous-structures du lobe temporal médian (MTL), y compris le cortex périrhinal (PRc), ont longtemps été considérées comme contribuant exclusivement à l'apprentissage et à la mémoire déclarative. Cependant, des recherches récentes ont révélé des signes d'implication du PRc dans des tâches complexes de discrimination perceptuelle nécessitant une approche conjoncturelle. Peu d'études ont directement testé ceci chez les jeunes adultes en bonne santé et une seule étude a essayé d'évaluer l'activation du PRc et sa relation avec la performance avec ces tâches dans un petit échantillon (N = 11).

L'étude actuelle d'imagerie par résonance magnétique fonctionnelle (IRMF) utilise un échantillon plus important (N = 24) chez des jeunes adultes en bonne santé pour enquêter sur l'activité cérébrale lors de la réalisation d'une tâche complexe de discrimination visuelle et utilise l'analyse « Seed - Partial Least

Squares » (PLS) pour relier l'activation dans le PRc à la performance et l'activité cérébrale dans le reste du cerveau. L’analyse « Task-Based – PLS » a identifié des ensembles distincts de régions cérébrales impliqués lors d’une tâche de discrimination conjoncturelle de caractéristique complexe par rapport à une tâche de contrôle de comparaison de taille simple, dont les régions d’intérêts (ROI) du PRc montrant une activation significativement plus grande pour les discriminations conjonctives de caractéristiques que pour les discriminations de contrôle. De plus, le « Seed – PLS », utilisant les ROI du PRc identifiés lors de notre analyse « Task-based – PLS », ont identifié deux réseaux clés. Le premier réseau a soutenu la performance à la tâche de contrôle et implique des régions bilatérales du PRc et visuelles postérieures

(BA 17, BA 18). Le deuxième réseau a supporté la performance à la tâche de discrimination conjoncturelle et implique les régions du PRc et frontales gauche (BA 9, BA 47). Cette étude fournit des preuves de l'implication répartie du réseau cérébral dans cette tâche et identifie des réseaux spécifiques impliquant le PRc droit et gauche, conjointement et indépendamment, qui prennent en charge la performance de cette tâche perceptuelle.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 6

Acknowledgements

First, I would like to thank Dr. Natasha Rajah for funding and supervising all of my thesis work.

Additionally, I would like to thank Stamatoula Pasvanis for her willingness to help me in each step from ethics preparation, to task presentation, subject recruitment, and technical troubleshooting. I am also grateful for the help of Holly Newbold-Fox and her availability to scan all participants and

Sonja Chu for her help in participant intake and scanning. Thank you also to my lab mate Elizabeth

Ankudowich for always discussing and challenging my ideas, approaches, and interpretations, and for providing feedback at every stage. Importantly, I would like to thank Dr. Morgan Barense who shared her task stimuli with us in order to expand on this field of research. Finally, I am grateful to my committee members, Dr. Janine Mendola and Dr. Pedro Rosa-Neto for their guidance in the preparation of this thesis.

Contribution of Authors

As first author, I, Lyssa Manning lead all aspects of conducting this research and preparing this thesis. This work included preparing and obtaining ethics approval from both MRRC and McGill

Faculty of Medicine, programming the task in E-prime, recruiting participants, administering screening questionnaires and obtaining demographic information from participants, collecting fMRI data, preprocessing functional and structural images, performing analyses and interpreting results, and writing all sections of this thesis. Stamatoula Pasvanis provided support in editing and translating ethics documents, designing the E-prime task, and testing participants. Sonja Chu also provided help in testing participants. Elizabeth Ankudowich aided in E-prime troubleshooting and technical design.

Dr. Morgan Barense developed the task stimuli and experimental paradigm. Finally, Dr. Natasha

Rajah established research design and hypotheses to be tested, developed the imaging protocol with the aid of Dr. Jamie Near and Ms. Holly Fox, provided guidance and support on all research conducted, and edited the entire thesis. Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 7

1. Introduction

For decades, research has supported the theory of the medial temporal lobe (MTL) as the

seat of long-term declarative memory, with the (HC) acting as a memory hub,

and the surrounding cortical areas involved in memory formation and storage. Since the

work of Scoville and Milner(1957) describing severe amnesia after bilateral MTL resection,

there has been extensive research into the role of MTL in memory.

Recently, however, researchers have begun to investigate the contributions of the MTL to

non-memory processes, such as perception, and have posited a perceptual-mnemonic

framework for MTL function (Ranganath, 2010). Bussey and Saksida (2007) propose the

perirhinal cortex (PRc) as key in MTL perceptual processing. By highlighting the connections

between PRc and anterior regions of the ventral visual stream, Bussey and Saksida present

the ventral stream as a visual “representational continuum”. They propose that the PRc, at

the apex of the continuum, contains complex feature conjunctive representations of stimuli.

Though many studies have shown perceptual deficits following PRc lesions in animals

(Bartko, Winters, Cowell, Saksida, & Bussey, 2007a, 2007b; Buckley MJ, Booth MC, Rolls ET,

& Gaffan D, 2001; T. j. Bussey & Saksida, 2007; T. J. Bussey, Saksida, & Murray, 2003;

Meunier, Bachevalier, Mishkin, & Murray, 1993; Saksida, Bussey, Buckmaster, & Murray,

2006), studies have varied in task design and analysis. To date, many studies in humans

(Barense et al., 2005; Barense, Henson, Lee, & Graham, 2010; Devlin & Price, 2007; Lee, Scahill,

& Graham, 2008; O’Neil, Cate, & Köhler, 2009) present the same challenges. Inconsistent use

of stimuli and experimental designs (for example, those with learning and memory demands

compared to those with purely perceptual demands) has made it difficult to interpret the Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 8 role of the PRc in perception and identify common processes across studies. Furthermore, the majority of the current literature does not directly investigate the relationship between

PRc functionality and performance on perceptual tasks. This study will use high resolution whole brain structural and functional MRI to investigate the role of PRc in higher order perceptual processing using a task designed by Barense et al (Barense et al., 2012;

Newsome, Duarte, & Barense, 2012; Ryan et al., 2012) that has been shown to activate the

PRc. Data will be analyzed using partial least squares (PLS) analysis to examine activation in the PRc in relation to performance and whole brain patterns of task-related activity. With this, we aim to strengthen the argument for PRc involvement in perceptual processing and provide support for the perceptual-mnemonic view of MTL function.

1.1 Medial temporal lobe involvement in declarative memory

The MTL is comprised of the HC, parahippocampal cortex (PHc), PRc, and (ERc). Since the case of HM, an epileptic patient who’s bilateral MTL resection

(targeting the HC and PRc) left him severely amnesic, these regions have been considered crucial for memory formation and retention(Scoville & Milner, 1957). However, research investigating extra-hippocampal MTL regions, has shown that they too are involved in memory processing. For example, Winters and Bussey (2005a, 2005b) found that transient blockage of PRc glutamate receptors NMDA and AMPA, necessary for memory encoding, consolidation, and retrieval, disrupted in rats. In this experiment, rats were familiarized to a specific object in one arm of a Y maze and time spent exploring the object was recorded. In the experimental phase, a novel object was placed in one maze arm and the familiar object was placed in the other. Injections of NMDA and AMPA antagonists Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 9

(AP-5 and CNQX, respectively) were given at pre and post sample exposure, and at pre- choice exposure (intending to target encoding, consolidation, and retrieval, respectively). In all conditions, rats explored familiar and novel objects equally, suggesting a failure to recognize previously encountered objects. By disabling the PRc and disrupting the ability to encode, consolidate, and retrieve the target, the authors highlighted the critical role of this region in object memory in rodents.

Extending this work to primates, investigators have used lesion studies to test the hypothesis that the MTL is a unitary memory system. Murray and Mishkin (1986) trained macaque monkeys pre-operatively on a delayed non matching to sample task where animals chose the novel object of a pair to receive a food reward. Monkeys were then broken into control, HC+rhinal (HC+Rh), and +rhinal (A+RH) lesion groups. HC+Rh monkeys performed significantly worse than controls, but A+Rh monkeys performed the worst, at a level no better than chance. This work was the first to show such a severe memory deficit in subjects with intact HC, suggesting that HC is not enough to support memory alone (see also

Meunier, Bachevalier, Mishkin, & Murray, 1993; Suzuki, Zola-Morgan, Squire, & Amaral,

1993; Zola-Morgan, Squire, Amaral, & Suzuki, 1989).

In humans, support for the MTL memory system comes from structural MRI analysis of amnesic patient HM (Corkin, Amaral, González, Johnson, & Hyman, 1997). It was found that in addition to HC ablation, HM also suffered from extensive damage to surrounding cortices, including much of the . In addition, human neuropsychological studies of patients with MTL damage have shown that patients exhibit impoverished memory as measured by a word list learning paradigm (Kopelman et al., 2007). Specifically, patients with Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 10 atrophy extending into the PHc recalled fewer words than patients with HC specific atrophy on a task of free recall. They also identified fewer correct words in a forced choice recognition test between a novel and familiar word. This indicates a role for extra-hippocampal involvement in memory and further supports the MTL memory system hypothesis.

Additional support comes from fMRI literature. In an associative memory paradigm, healthy participants learned pairings of abstract geometric stimuli with one of four button presses(Law et al., 2005). Baseline activity was recorded during a non-mnemonic button press task. FMRI analysis indicated increased BOLD signal during pairing of the stimuli-response associations in bilateral HC and PHc, and right PRc, compared to baseline. Moreover, BOLD signal increased in conjunction with accuracy on the task.

In summary, the above studies support the theory of a MTL memory system which maintains that the MTL is reserved for memory processing, with MTL structures contributing to support of declarative memory. Indeed, several models have been put forth regarding region- specific contributions of the MTL to declarative memory. One such model is the Binding of

Items and Contexts model (Eichenbaum, Yonelinas, & Ranganath, 2007) which proposes that HC and PHc are critical for recollection, the ability to remember in specific detail, and PRc is responsible for familiarity, a vague sense of having experienced a stimulus before but without rich contextual detail of the occurrence. However, recent updates to this model have expanded the role of PRc to include non-mnemonic perceptual capacity (Ranganath, 2010). While HC and

PHc are largely agreed to primarily support memory processes, the role of PRc is still debated.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 11

1.2 The region-specific role of perirhinal cortex

As mentioned above, it is generally agreed that distinct regions of the MTL contribute distinct functions to declarative memory encoding and retrieval. However, there remains debate whether PRc’s contribution to declarative memory is memory-specific or related to a higher order perceptual function. Animal studies laid the foundation for research into the perceptual role of PRc by showing that macaque monkeys with PRc lesions performed more poorly than controls on a task of concurrent stimulus discrimination(T. J. Bussey & Saksida,

2002). In this study, monkeys were trained to discriminate between pairs of stimuli with varying levels of feature ambiguity. Feature ambiguity refers to the overlap in features between two stimuli. For example, low/minimum feature ambiguity refers to pairs of stimuli in which rewarded and non-rewarded stimuli share no features. High/maximum feature ambiguity refers to pairs of stimuli in which rewarded and non-rewarded stimuli share all of the same features. Intermediate feature ambiguity refers to pairs of stimuli in which rewarded and non-rewarded stimuli share some, but not all, of the same features.

Discriminating high feature ambiguity pairs, therefore, requires the integration of multiple features at once. This integration of features is referred to as feature conjunction. A distinction between high ambiguity stimuli cannot be made by comparing single features in isolation, but rather relies on the conjunction of multiple features. Therefore, feature conjunctive processing is required to discriminate stimuli with high feature ambiguity. In the study by Bussey et al. (2002), monkeys with bilateral PRc lesions were significantly impaired on high and intermediate feature ambiguity discriminations. However, they performed normally on low ambiguity discriminations, supporting a role of PRc in feature conjunctive Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 12 processing specifically. A complementary study (Buckley MJ et al., 2001) investigated the ability of monkeys with PRc lesions to make oddity judgments from arrays of six stimuli.

Some judgments (size, colour, shape) could be made on the basis of one feature, while other judgments (faces) required more complex conjunctive processing. Buckley et al. found that monkeys with PRc lesions were significantly impaired when the stimulus category was more complex, and conjunctive processing was required. Together, these studies provided evidence for the PRc’s involvement in visual perception, specifically in the representation of complex stimuli that require the integration multiple features simultaneously.

Human behavioural studies have also explored the role of PRc in perception. Barense et al

(2005) tested patients with focal MTL lesions on a rapidly learned, forced choice discrimination task similar to that of Bussey et al (2002). Patients had damage either restricted to the HC (HC-), or extending into the PRc (HC+PRc). Stimuli were presented on screen in pairs, of which one item was arbitrarily coded as “correct”. Task difficulty was manipulated in a similar fashion to Bussey et al (2002), so that pairs varied in number of shared features. Participants chose one stimulus in the pair and were given performance feedback. The round was complete once the participant had chosen correctly eight times in a row. Barense et al (2005) found that HC- patients performed normally across all conditions, while HC+PRc patients performed normally when the level of feature ambiguity was low, but were significantly impaired when it was high. This behavioural result highlights the integral role of the PRc in resolving feature ambiguity specifically when stimuli share overlapping features. However, one of the biggest criticisms of PRc in perception is that much research involves memory dependent tasks, such as Barense et al. (2005) which required participants to Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 13 learn and remember correct stimuli across trials. This confound makes it difficult to claim a purely perceptual role of the PRc.

To address this criticism, Lee, Bussey et al. (2005) tested the individuals who participated in the Barense et al (2005) study on a non-mnemonic visual discrimination task.

Test stimuli from five different categories (faces, objects, scenes, art, and colour) were used.

Participants were required to choose which of two stimuli most closely resembled the target and levels of feature ambiguity were manipulated by morphing stimuli along a continuum.

HC+PRc patients showed impairment in discriminating high ambiguity objects and faces, consistent with a role for PRc in conjunctive feature processing. In another study, Lee,

Buckley et al.(2005) used the same subjects in another non-mnemonic perceptual task. In this study, patients were asked to identify the odd one out from an array of either four scenes or four faces in different orientations. All patients (HC- and HC+PRc) performed poorly on scene discrimination, however HC+PRc were additionally impaired on face discrimination. While the deficit seen in scene discrimination across both groups suggests a role for the HC in spatial processing, the deficit in face discrimination in HC+PRc patients suggests a role for PRc specifically in feature conjunctive processing.

To assert that the resolution of feature ambiguity is critical for successfully discrimination, Barense, Gaffan & Graham (2007) designed an experiment in which the same participants as Barense et al. (2005) and Bussey et al. (2005) performed two novel, trial unique object oddity tasks. For the first task, patients were asked to identify the odd one out from an array of seven “fribbles”. Fribbles (Williams & Simons, 2000) share the same main body and have four appendages, each with 3 possible values. This allows for manipulation of Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 14 feature ambiguity across pairs. In the second task, patients were asked to identify the odd one out from an array of four objects in four different orientations. Objects in an array were always from the same category, but low ambiguity objects could be distinguished by a single feature, whereas high ambiguity objects shared several overlapping features. HC+ PRc patients were significantly impaired on both tasks only when feature ambiguity was high.

HC- patients showed no impairment on either task, regardless of level of ambiguity.

Together, these studies highlight the non-mnemonic role of PRc in perception, specifically for discriminating between stimuli with highly ambiguous or overlapping features.

These behavioural findings have also been supported by fMRI research using the same task from Lee, Buckley, et al.(2005) in healthy participants (Lee et al., 2008). Scene oddity discriminations correlated with increased BOLD signal in the posterior HC and PHc, while face oddity discriminations correlated with increased BOLD activation in the left PRc. These results run parallel with the behavioural results seen in patients (Lee, Buckley, et al., 2005) in which high ambiguity face discrimination was preserved in patients with preserved PRc.

In another fMRI study, healthy participants performed a discrimination task in which they identified the odd one out from an array of four three-dimensional object or animal stimuli(Devlin & Price, 2007). Easy conditions consisted of easily distinguishable targets (i.e. black drills vs white mug) presented from the same angle, and difficult conditions consisted of visually similar targets (i.e. black drills vs black tape dispenser) presented from various angles.

Therefore, the ability to discriminate difficult trials required feature conjunctive processing, integrating visual features to perceive the object at an abstract, view-invariant level. The authors found that participants engaged bilateral PRc cortex for difficult, but not easy, trials. Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 15

Extending Devlin and Price’s findings, Barense et al. (2010) tested healthy young adults on a similar task. Participants identified the odd one out from an array of either three novel faces, objects (“greebles” (Gauthier & Tarr, 1997)), or 3D scenes. Stimuli were presented all from the same orientation (easy) or all from different orientations (difficult). The authors found increased BOLD activation in the right PRc when viewing difficult arrays of faces and objects.

Since all stimuli were trial-unique, this result provides further evidence for PRc in non- mnemonic perceptual tasks requiring feature conjunctive processing.

In another study using face stimuli, participants identified which face in a pair looked least like a simultaneously presented target face (O’Neil et al., 2009). To maximize difficulty, faces were morphed along a continuum. As in Barense et al. (2010), BOLD activity was observed in right PRc during oddity judgments. Additionally, the authors examined PRc accuracy effects and found that PRc activation successfully distinguished between correct and incorrect trials. These results were the first to investigate and link PRc activity to performance on a non-mnemonic perceptual task.

As face stimuli have been shown to involve unique networks of brain regions not similarly involved in perception of other stimuli (Rossion, Hanseeuw, & Dricot, 2012), Barense et al

(2012), tested participants on a trial unique paired “blob” discrimination task. High feature ambiguity blob pairs shared two of three features, whereas low feature ambiguity pairs shared no features. A matched difficulty square size discrimination was used as a control task. Barense et al. found that bilateral PRc BOLD activation was specifically and significantly increased when participants were viewing high ambiguity blobs relative to low ambiguity blobs, but not high Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 16 difficult squares vs easy squares. This evidence suggests that PRc activation is not related to difficulty but rather is specifically recruited for conjunctive feature processing.

Using the same “blob” paradigm, Ryan et al. (2012) compared task related PRc activation in younger and older adults. They found that both groups engaged bilateral PRc when performing the high ambiguity discrimination task, consistent with previous results (Barense et al., 2012).

Additionally, they strove to associate PRc activation with performance, a measure that would support the relationship found by O’Neil et al. (2009). They found that in older adults, but not younger adults, PRc activation was positively related to task performance in the left anterior

PRc only.

As evidenced from the above research, PRc is consistently involved in feature conjunctive perceptual processing. What is not apparent, however, is the lateralization of PRc recruitment, as studies have reported left (Lee et al., 2008), right (Barense et al., 2010; O’Neil et al., 2009), and bilateral (Barense et al., 2012; Devlin & Price, 2007; Ryan et al., 2012) PRc results, as well as left(Ryan et al., 2012) and right (O’Neil et al., 2009) PRc performance effects.

1.3 Objectives FMRI investigation of the PRc in perception is a relatively novel field, and as such, methods of analysing task-related activity as well as its relationship to performance vary across studies. Prior studies have analysed PRc contribution to object perception in isolation from the rest of the brain. However, there is growing consensus that perception and cognition are mediated by larger neural networks(Park & Friston, 2013). To address this, in the current study we focus on PRc perceptual processing within the framework of large-scale activation patterns. To do this, we apply multivariate partial least squares (PLS) analysis to examine task- Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 17 dependent PRc activation both singularly and in relation to co-activated brain regions. This will allow us to explore how activity in the PRc and other brain regions relate to performance on the perceptual task developed by Barense et al(2012). In addition, we will use high-resolution fMRI to get precise localization and measurement of PRc during this task.

1.4 Hypotheses

We expect to find:

1) Whole brain patterns of activation that distinguish between pairs of blobs with high

feature ambiguity and single feature square stimuli. In particular, we expect that

activation in PRc will be more associated with blobs than squares.

2) PRc BOLD activation, in conjunction with additional whole brain network regions will be

positively related to performance on the blob discrimination task.

1.5 Specific Aims

1) Use high resolution MRI scans to better visualize PRc.

2) Evaluate task-related BOLD signal in whole brain patterns during performance of a

perceptual discrimination task. Additionally, we will assess task-related differences in

PRc activation in particular.

3) Assess how task related activation relates to task performance.

2. Methods

2.1 Initial participant recruitment

Undergraduate participants were recruited from a university wide online ad. Interested participants were contacted by email and required to fill out a medical screening questionnaire.

This form collected data on current and lifetime history of any physical, neurological and/or Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 18 psychological illnesses. Additionally, it collected data on education, visual acuity, general physical health, risk factors for cardiovascular disease, and history of smoking, drug and alcohol use. Based on the medical questionnaire, participants were excluded if they were left-handed as left handed participants have been shown to demonstrate altered lateralization of function

(Knecht et al., 2000). Additionally, in order to ensure that all stimuli were seen clearly, participants were excluded if they had untreated cataracts or glaucoma. Finally, participants with a current diagnosis of high cholesterol levels and/or high blood pressure left untreated in past 6 months were excluded due to potential for irregularities in the BOLD fMRI signal due to irregular blood flow.

Participants who met initial inclusion criteria were invited to come to the laboratory at the

Douglas Institute for a behavioural testing session. Of the recruited respondents, 28 young adults (mean age: 20.9 yrs, mean education 14.46, 13 female) were scheduled for a behavioural testing session. Of the 28 young adults who completed a behavioural testing session, three (3) were excluded for below average age & education norms on a memory screening test (CVLT – described below). One (1) participant was excluded from analysis due to a corrupted fMRI data file. The final subject pool was N=24, ages 18-28 (mean 20.79, SD= 2.65), education 13-18 yrs

(mean 14.5 yrs, SD 1.53 yrs), 12 female (Table 1).

2.2 Behavioral session

Participants were administered: Edinburgh Inventory (Oldfield, 1971) for handedness, Family history of AD screening form - Cache County questionnaire (Hayden et al., 2009), and the

Language and Social Background Questionnaire (LSBQ) (Bialystok et al., 2005; Bialystok, Craik, & Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 19

Freedman, 2007) to evaluate level of bilingualism. Because the study took place in Montreal, a predominantly bilingual city, assessment of bilingualism is important as previous research has shown that bilingualism may increase executive control ability (Bialystok et al., 2005, 2007).

Should there be group differences in the number of uni- ,bi-or multi-lingual subjects, we will include LSBQ score as a covariate in analyses.

Participants then underwent a battery of neuropsychological tests (listed below) administered by a trained researcher. The neuropsychological battery was administered to ensure that all participants fell within the range of normal cognitive function and met criteria for healthy cognition as per inclusion criteria. Participants who met the inclusion and exclusion criteria listed below were scheduled for an fMRI session. The length of the Behavioral Session was approximately 2 hrs.

2.2.1. Neuropsychological Evaluation

All included subjects met the cut-off criteria for the following neuropsychological tests to participate in the study. These tests were administered in order to ensure healthy cognition in included participants.

a. Mini-International Neuropsychiatric Interview (M.I.N.I.)(Sheehan et al., 1998) – cut-off

=/< 2 affirmative responses in each module

b. Holmes and Rahe Schedule of Recent Events (Sands, 1981) for stress – cut-off < 150;

c. Beck Depression Inventory (BDI)(Beck & Steer, 1993) - cut-off < 15

d. California Verbal Learning (CVLT)(Norman, Evans, Miller, & Heaton, 2000; Paolo, Troster,

& Ryan, 1997)– cut-off determined per case using age & education; Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 20

e. National Adult Reading Test (NART) (Strauss, Sherman, & Spreen, 2006) – cut-off =/< 2.5

standard deviation from normative data at given age and education per subject.

2.3 MRI data acquisition

MRIs were performed on a 3T Siemens Trio scanner, at the Douglas Brain Imaging Centre.

Subjects were asked to lie in a supine position in the MRI scanner while wearing a standard head coil. FMRI task images were back projected on a screen in the scanner bore that subjects viewed via a mirror mounted on the head coil. E-Prime presentation software (Psychology

Software Tools, PA) was used to present all fMRI tasks and to collect behavioral data. One 4- button fibre optic response box was used to collect behavioral responses. Plastic optical corrective glasses were worn by participants requiring correction for visual acuity.

Scans were obtained in the following Scan Order:

1. High resolution T1-weighted structural MRIs acquired for fMRI image registration, and

structural analysis (TR 14 ms, TE 4.92 ms, flip angle 25o, 176 1mm thick transverse

slices, 1 x 1 x 1 mm voxels, FOV = 256 mm2) (5 min)

2. High-resolution BOLD fMRI scans acquired using single-shot single shot T2*-weighted

gradient echo EPI pulse sequence (TR = 2000 msec, TE= 35 msec, FOV = 240 mm 2,2x2x2

mm voxels, 42 slices) while performing an Object Discrimination Tasks - 20 min:

Subjects performed 2 runs of an Object Discrimination task in the scanner in which

they decided if two simultaneously presented stimuli were identical (a match) or

different (a non-match). Stimuli for the study were developed and tested by Ryan et

al (2012)(Figure 1). See also Newsome et al (2012) and Barense et al(2012). The blob

stimuli are composed of three features: two curved blob-like shapes, one inside the Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 21

other, with a pattern separating them. For non-matched blobs, pairs overlap in two

of the three features, so that only one feature can be used to discriminate objects

pairs (which of the three features that differed was fully counterbalanced across

trials). For matched blobs, pairs are identical to one another on all three features. To

make the task more demanding and to discourage a simple feature-matching

strategy, blobs were rotated with respect to each other.

As a perceptual control task, pairs of squares of varying sizes were used. Side length

of non-matched square pairs differed by a minimum of 2 mm to a maximum of 4 mm.

Matched square pairs were identical to one another. The squares were rotated with

respect to each other to ensure sufficient difficulty on the task, and to match the

rotation component included in the blob discrimination task.

Thus, a total of four conditions were included in the experiment: non-matched blobs,

matched blobs, non-matched squares and matched squares. The total ratio of non-

matched to matched pairs in the experiment was 2:1. Pairs of stimuli were presented

in the scanner, one at a time, using E-Prime presentation software (Psychology

Software Tools, PA). Participants were instructed to judge whether or not the two

stimuli in the pair (either blobs or squares) were identical to one another.

Participants were informed that the stimuli were rotated relative to one another, and

that the degree of rotation in each trial was random. Each pair was presented for 5.5

seconds allowing sufficient time for response, with variable ITI of 2.2-8.8 sec (mean

ITI 4.8 sec). Trials were presented in mini-blocks of three pairs of blobs or three pairs

of squares. The mini-blocks were then presented in pseudo-random order in two Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 22

separate runs, each lasting approximately 10 minutes. Each scan included an equal

number of the four conditions described earlier.

Participant responses were gathered using an optical fibre response box. The

response box was placed in the dominant (right) hand of the participant, and they

were instructed to press the index finger to indicate a "matching pair" and the middle

finger to indicate a "nonmatching pair."

3. High resolution T2-weighted scan for volumetric analysis of medial temporal lobe sub-

regions (TR 2500ms, TE 198 ms, 320 slices of 0.60 mm thickness, 0.6x 0.6 x 0.6 mm

voxels, FOV = 206) (13 min)

The rationale for collecting high-resolution fMRI BOLD images (2x2x2) vs. the traditional 4x4x4 resolution images is to better identify regions of interest (ROIs) in the medial temporal region for individual subjects. Each whole brain acquisition consisted of twenty-four oblique slices of

2.0 mm thickness and in-plane resolution of 2X2 mm, acquired along the anterior-posterior commissural plane.

2.4 Behavioural Analysis

Mean accuracy scores for blobs and squares was calculated separately for each task for each participant as a percentage of correct responses. A paired samples t-test of mean task accuracy scores was used to confirm equivalency of difficulty between tasks. Additionally, d-prime discrimination scores were calculated for blobs and squares for each participant. D-prime scores are calculated as the z score of hits (correctly discriminating between two non-matched stimuli) minus the z score of false alarms (incorrectly responding that two matched stimuli are non-matched). D-prime scores are used to assess accuracy while taking into account response Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 23 bias. A large number of false alarms may be the result of a subject’s bias to respond that stimuli pairs are non-matched, resulting in an overall higher accuracy for correct trials. In the event that some subjects have significantly higher false alarm rates, d-prime scores will be used for analyses in place of raw accuracy scores to account for this bias. Finally, the Descriptives >

Explore function of SPSS was used to identify outliers on any of these measures.

2.5 Image processing

Statistical Parametric Mapping(SPM8) software, a commonly used and validated program for processing and analysis of fMRI data developed by Friston et al (2002; 1994; 1994; 1999), was used to spatially realign fMRI images and correct for movement artefact, using a 6 parameter rigid body spatial transform and a least squares approach. Functional images were co- registered to each individual’s own functional EPI-template (Rajah, Languay, & Valiquette,

2010) and transformed into stereotaxic space, a standard coordinate space based on the

Talairach atlas (Talairach & Tournoux, 1988) (normalization), to facilitate group analysis by controlling for variability in brain size (Collins DL, Neelin P, Peters TM, & Evans AC, 1994). The rationale for using individual EPI images (resolution of 2x2x2), as opposed to the standard MNI

EPI image (resolution of 4x4x4) is to preserve the higher resolution of our functional images.

Volumes were resampled into 2-mm cubic voxels and smoothed using an 8 mm full-width half maximum (FWHM) isotropic Gaussian kernel, to minimize inter-participant anatomic variability and facilitate group analysis (Collins DL et al., 1994).

2.6 fMRI data analysis

Multivariate spatio-temporal partial least squares (PLS)(McIntosh, Chau, & Protzner, 2004) was used to analyze fMRI data. PLS is a multivariate analysis that allows for the detection of spatially Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 24 and temporally distributed brain activation across experimental conditions. This type of analysis was used in order to facilitate the investigation of large scale cortical networks involved in this task. In this study, mean centered task PLS (T-PLS) was used to investigate brain activity during correct blob and correct square discrimination tasks. Seed behaviour PLS (B-PLS) was used to investigate brain activity directly correlated with accuracy on these tasks and with our seed regions (PRc ROIs).

For both T-PLS and B-PLS, each subject’s fMRI data is included in a group data matrix. PLS analysis of event-related fMRI “flattens” the temporal dimension (t), so that the time series of each voxel (m) is stacked side-by-side across the columns of the data matrix (column dimension

=m*t). This converts the three-dimensional event-related fMRI data into a two-dimensional matrix. The rows of this two-dimensional matrix for one individual subject represent activity during each event stacked within experimental condition (row dimension = events*conditions).

Event-related data for each event-type (BlobCorrect, BlobIncorrect, SquareCorrect,

SquareIncorrect) are averaged within subject, and subject fMRI data is stacked. Because there is only one group under investigation in this study (young adults), stacking of multiple groups is not required. The stacked data matrix contains the fMRI data for each event onset (time lag=0) as well as the subsequent seven TRs/time lags (TR=2 sec * 7 = 14 sec) following event onset for correctly answered blob discrimination trials, and correctly answered square discrimination trials. All included subjects had a minimum of 17 correct events per event type (mean # of correct events for blobs = 37, mean # of correct events for squares = 36).

For the T-PLS analysis, the fMRI data matrix described above undergoes voxel-wise grand- mean scaling. The matrix is mean centred, column-wise, within each event type. Singular value Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 25 decomposition (SVD) is performed on the group data matrix, to yield orthogonal paired latent variables (LVs). Each LV in the mean-centered T-PLS reflects i) a singular value that reflects the amount of covariance accounted for by the LV, ii) design saliences that represent the contrast effect represented by that LV, and iii) a singular image representing a pattern of event-related whole brain activity for the contrast effect identified in the design salience plot. The singular image contains numerical weights assigned at each voxel for each TR/ time lag in the data matrix. These weights are called brain saliences and they can be positive or negative. Regions with positive saliences are positively related to the contrast effect identified in the design salience plot of the LV of interest. Regions with negative saliences are negatively related to this effect. Therefore, the pattern of whole brain activity, as identified in the singular image, is symmetrically related to the contrast effect in the design salience plot.

For B-PLS, the group fMRI data matrix is correlated with a behavioural vector. We used a specific type of B-PLS called a Seed PLS. For this analysis, the behavioural vectors used are the mean activation of specific regions of interest, or “seeds” (bilateral PRc), and the mean accuracy scores for blob and square trials (% correctly discriminated pairs), stacked in the same order as the data matrix (by subject). SVD of this matrix is conducted to yield a series of LVs.

These LVs are similar to those output by a T-PLS, with a few key differences. Instead of design saliences, a B-PLS analysis creates i) a singular value, ii) a singular image of positive and negative brain saliences, as described above, iii) a correlation profile depicting how subjects’ accuracy scores relate to the pattern of brain regions identified in the singular image. The correlation profile in combination with brain saliences reflected in the singular image, represent a symmetrical pairing of brain-behaviour correlation patterns to pattern of brain activity. For B- Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 26

PLS, as with T-PLS, brain saliences can be positive or negative, and reflect whether activity in a given voxel is negatively or positively associated with the correlation profile. This analysis reveals regions whose activity correlates with the activity of the seed region and also with accuracy on the task.

LVs from both T-PLS and Seed PLS were tested using 500 permutation tests. For each permutation, a PLS is recalculated, and the probability that the permuted singular value exceeds the observed singular value for a specific LV is used to assess significance. For this study, significance is assessed at p<0.05. In order to determine standard error of brain saliences for each LV, 100 bootstraps were conducted by sampling subjects with replacement while maintaining experimental event/condition order for all subjects. A bootstrap ratio (BSR), similar to a Z-score, is calculated for each voxel-based brain salience. The BSR of a significant voxel salience reflects the stability of its activation. A threshold of BSR ≥ ±3.28, p < 0.001, with a minimum spatial extent = 10 voxels was set to report maximal stable brain saliences from each significant LV.

For each task in a significant LV, temporal brain scores are computed. These scores represent the degree to which each subject within the group displays the pattern of brain activity shown in the singular image in relation to design salience (T-PLS)/ correlation profile (B-

PLS) pairing at each time lag. This score is used to identify time lags at which the effect identified in the LV is maximally differentiated. Only reliable brain saliences (BSR ≥ ±3.28) from those time lags are reported. The resulting peak coordinates for reliable brain saliences are converted from MNI space to Talairach space using the icbm2tal transform (Lancaster et al., Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 27

2007) as implemented in GingerAle 2.3 (Eickhoff et al., 2009). Brodmann areas corresponding to significant peaks were identified using the Talairach and Tournoux atlas (1988).

To address specific aim 2, we conducted T-PLS analyses of fMRI data obtained during the object discrimination tasks to observe activation during conjunctive processing (high feature ambiguity) vs. non-conjunctive (low feature ambiguity) processing.

To address specific aim 3, we conducted a Seed PLS, to identify brain saliences that correlate with activity of bilateral PRc peak saliences and also with accuracy scores.

3. Results

3.1 Behavioural Results

Mean accuracy for blob trials and mean accuracy for square trials was calculated for all participants individually. Equivalency of difficulty was confirmed using a paired samples t test, t(23) = .327, p>.05 (Figure 2). This is consistent with previous literature using this task

(Barense et al., 2012; Ryan et al., 2012). SPSS > Descriptives > Explore function found no participants with outlying false alarm scores. Therefore, we proceeded with standardized accuracy scores as a measure of performance for all participants.

3.2 fMRI Results

We conducted two multivariate analyses to identify 1) whole brain patterns of activation in task-related activity during a perceptual discrimination task (T-PLS), and 2) correlations between whole brain patterns of task-related activity and task performance (Seed PLS).

3.2.1 Task PLS Results Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 28

A mean-centered T-PLS contrasting blob and square trials revealed one significant LV. The significant LV identified task-related differences between blob trials and square trials (Figure

3B). T-PLS LV1 accounted for 84.12% of the cross-block covariance and identified brain regions that were differentially activated during blob discrimination trials vs square discrimination trials. Positive salience brain regions were more active during blob discrimination trials than square discrimination trials. Negative salience brain regions were more active during square discrimination trials than blob discrimination trials (Figure 3A)

(Table 3).

3.2.1.1 ROI-based analysis of PRc regions identified in PLS analysis Because our hypotheses were concerned specifically with PRc, we investigated only salient peaks that were identified as PRc. T-PLS LV1 identified two peaks in PRc, Left BA 36/20 (MNI:

-30, -2, -40) and Right BA 36/20 (MNI: 34,-8,-38). The extent of the peaks straddled the border between BA 36 and BA 20 so both areas are identified in our label. We extracted the mean activity from a 4 mm cubic region around both peaks using the multiple voxel extraction (MVE) tool in PLS. We calculated the mean activity for lags 1-5 for each PRc ROI and compared activation between blob trials and square trials. Paired samples t tests revealed that bilateral PRc was significantly more active during blobs trials than squares trials t(23), p< 0.05 (Figure 4).

3.2.2 Seed PLS Results

Seed PLS was conducted using the bilateral PRc ROIs identified in the T-PLS (MNI: -30,-2,-40 ;

34, -8, -38) and participant raw accuracy scores. This PLS identified two significant LVs. LV1 accounted for 56.85% of the cross-block covariance. This LV identified a set of brain regions Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 29 in which activity was positively correlated with activity in bilateral PRc and with accuracy during the squares trials. LV2 accounted for 20.46% of the cross-block covariance. This LV identified a set of positive salience brain regions in which activity was negatively correlated with activity in the left PRc seed and positively correlated with accuracy during the blobs task. Together, these LVs represent two distinct sets of brain regions that differentially support accuracy during the square vs. blob tasks. As stated above, because no participant had outlying false alarm scores, raw accuracy scores were used for all analyses.

3.2.2.1 Salient ROIs in Seed PLS LV1 of the Seed PLS identified positive correlation between bilateral PRc and accuracy during the squares task (Figure 5B). Salient regions in LV1 of the Seed PLS are shown in Figure 5A and those that were most salient are defined in Table 4. Peak regions identified in bilateral

BA 18 (Right: 18, -56, 6 Left: -30, -86, -4) and right BA 17 (16, -94, 10) indicate a role for posterior visual areas, in conjunction with bilateral PRc in successful discrimination of square trials.

LV2 of the Seed PLS identified negative correlation between left PRc activity and accuracy during the blobs task (Figure 6B). Salient regions in LV2 of the Seed PLS are shown in Figure

6A, and those that were most salient are defined in Table 5. Peaks regions identified in bilateral BA9 (Right: 26, 66, 22 Left: -44, 22, 28) and right BA47 (24, 10, -26) indicate a role for frontal regions in successful discrimination of blobs trials.

4. Discussion

This study aimed to identify PRc activity during an ambiguous perceptual discrimination task, and to investigate its relationship with accuracy on the task. We found that, in comparison Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 30 with control size discrimination trials, PRc was significantly more active during ambiguous discrimination trials. We then identified two key networks, one supporting accuracy for control size discriminations, and the other supporting accuracy for ambiguous discrimination trials. A simplified view of these two networks is depicted in Figures 7A and 7B. We will discuss the findings, as well as caveats, of this study in the following sections.

4.1 Task dependent activation patterns

Our fMRI results demonstrated that there were significant differences in task-related patterns of activation between blob trials and square trials. These task differences, identified in LV1 of the T-PLS, indicate that participants recruited different sets of brain regions when performing blob discriminations than when performing square discriminations. Additionally, we identified peaks in bilateral PRc that showed significantly greater activation for blobs trials than squares trials. This is consistent with previous research suggesting that PRc is critically involved in perceptual discriminations requiring feature conjunctive processes in order to distinguish between highly ambiguous visual stimuli (Barense et al., 2012; Ryan et al., 2012). Importantly, these results are the first to identify distinct patterns of brain activity recruited for each task.

4.2 Square discrimination network

Using the PRc ROIs identified in the T-PLS described above as seed regions, Seed PLS revealed that bilateral PRc was correlated with a network of regions supporting accuracy for square discrimination trials. Previous studies using this task have only investigated the involvement of the PRc in blob trials and have used strictly ROI based approaches(Barense et al., 2012; Ryan et al., 2012). In the current study we identified a set of regions that, in correlation with bilateral Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 31

PRc activation, supported accuracy specifically for square discrimination trials. The regions implicated in this network (bilateral BA 18 and R BA 17) are primary and visual association areas, consistently implicated in basic visual processes, such as shape and edge detection & orientation (Hubel & Wiesel, 1962, 1968; Orban, Dupont, Vogels, Bormans, & Mortelmans,

1997). This result could be due to the fact that stimuli were rotated, highly contrasted white squares on a black background requiring the explicit recruitment of these faculties.

The correlation between bilateral PRc and posterior visual areas found in this result strengthens the argument that the PRc maintains connections with posterior visual areas, a finding that has been seen structurally, as well as functionally (T. j. Bussey & Saksida, 2007). This finding also lends support to the perceptual-mnemonic hypothesis of MTL function (Ranganath, 2010), stressing the continuum of perceptual processes from the primary visual and visual association areas all the way to the PRc, which is historically defined as part of the MTL.

4.3 Blob discrimination network

The second network of regions identified in the Seed PLS revealed that the left PRc was correlated with a network of regions supporting accuracy for blobs trials. Interestingly, key regions identified are prefrontal areas (bilateral BA 9, right BA 47) historically involved in working memory(Ranganath, 2010; Zhang, Leung, & Johnson, 2003), and deductive reasoning

(Goel, Gold, Kapur, & Houle, 1998) , both necessary to resolve feature ambiguity between complex stimuli. Additionally, the nature of this task demands a sustained attentional effort, a function that BA 9 has been consistently implicated in (Cabeza & Nyberg, 2000). Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 32

This result builds upon the previous understanding of PRc in complex object discrimination by illuminating the network within which the PRc is implicated. Additionally, implication of prefrontal areas BA 9 and BA 47 coincides with Badre et al’s (2008) hierarchical organization of

PFC function. This hypothesis asserts that cognitive control is modulated along the rostro- caudal axis of the PFC in a hierarchical fashion. The authors propose that more abstract representations (ex: discriminations between stimuli with overlapping features (i.e. texture, shape, size, orientation) made in response to a specific cue) recruit anterior PFC regions and simpler, more concrete representations (ex: choosing a colour coded response) are resolved in posterior PFC regions. Badre et al.’s model classifies levels of processing as varying in control demand based on the following levels of abstraction:

• first-order abstraction (“response”) ex: choosing a stimulus based on learned mapping • second-order abstraction (“feature”) ex: identify specific features of a stimulus in response to a colour cue • third-order abstraction (“dimension”) ex: determine if two objects match along one of multiple dimensions in response to a cue • fourth-order abstraction (“context”) ex: determine if two objects match along more than one dimension in response to a cue As per Badre et al.’s model, the task in our current study falls at the dimension level, due to the necessity of feature integration between stimuli. At this level of abstraction, this model hypothesizes activation in the anterior DLPFC. Fittingly, the results we observed in BA 9 and BA

47 fall within the model’s expected regions and support this view of PFC function (Figure 7C).

It is important to note that this hypothesis is distinct from Koechelin et al.’s (2003) cascade hypothesis of PFC hierarchical function. Koechelin et al. hypothesize that the rostro-caudal axis operates on a gradient of control signal divided into sensory (motor response selection from Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 33 direct sensory input), contextual (action selection based on contextual clues), episodic (action selection based on ongoing contextual information or rules) and branching (action selection based on rules and information on a more temporally removed context) control. Our task, in terms of Koechelin et al.’s cascade hypothesis, would fall under the contextual level of processing and be expected to engage posterior prefrontal regions. However, as previously discussed, peak regions of activation were identified in more anterior areas (BA 9, BA 47) which lends support instead to Badre et al.’s model.

It is notable that only left PRc was implicated in this network. It is possible that participants engaged verbal strategies to attempt to discriminate stimuli in order to give meaning to otherwise nonsense shapes (for example, imagining the blobs as amoebas). As language faculties have been shown to be left lateralized in right handed people(Frost et al., 1999), this may have resulted in a left hemisphere relationship with performance.

Additionally, it is notable that the correlation with left PRc was a negative correlation. This may suggest that for ambiguous discriminations, a left-hemisphere dependent verbal strategy was not as effective and such reliance on left PRc resulted in poorer performance.

4.4 The PRc and non-visual conjunctive representations

It is possible that the role of the PRc in distinguishing between ambiguous stimuli is not limited to perceptual discriminations. Research in rats has shown that the perirhinal cortex is involved in complex auditory (Campolattaro & Freeman, 2006; Lindquist, Jarrard, & Brown, 2004), olfactory (Feinberg, Allen, Ly, & Fortin, 2012) and somatosensory (Ramos, 2014) stimulus discrimination. Further research demonstrating these effects in both animal and human subjects could implicate the PRC in cross-modal feature conjunctive processes. Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 34

4.5 Strengths and Limitations

This study is the first to examine the perceptual role of the PRc in the context of whole brain networks of activation and the first to assess these networks in relationship to task performance. Additionally, functional and structural images were acquired at a higher resolution (2x2x2) than previous studies (4x4x4). However, there are limitations to this design. It is possible that the use of strictly high ambiguity stimuli and difficult size comparisons may have resulted in perceptual fatigue across time. This may have resulted in overall lower average performance scores and affected participant effort. Additionally, this may have lead to functional differences between early “non fatigued” trials and later

“fatigued” trials. Further, participants were not asked to disclose what discrimination strategies they used during the task. This information may have allowed us to better explain the patterns of activity seen in both squares and blobs tasks, and to determine if participants were using a verbal strategy to solve ambiguous discrimination trials. Finally, because our research question applied specifically to the PRc, we examined this ROI in particular.

However, in future studies, other MTL ROIs should be examined in order to determine if this effect is truly specific to the PRc or if it can be generalized to other MTL structures as well.

4.5 Conclusions

In this study, we identified bilateral task-related PRc activity during a complex feature conjunctive discrimination task. This activity was observed in the context of larger patterns of co-activating regions and identified hemisphere-specific networks of activation supporting accuracy on blob and square tasks. Additionally, we integrated PRc activity into the context of Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 35 larger patterns of task related activation and performance related hierarchical PFC function.

We have shown that left and right PRc are differentially involved in these visual discrimination tasks and that the left PRc may have more of a role in distinguishing between highly ambiguous stimuli than the right PRc, which maintains functional correlations with more posterior visual areas. With this work, we continue to expand the discussion of PRc in perceptual function, specifically in performance of discrimination between highly ambiguous stimuli.

It is important that further research continues to use whole brain analysis to replicate these findings and to further investigate networks of regions that are shared between both left and right PRc as well as the networks that are unique to each.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 36

5. References

Badre, D. (2008). Cognitive control, hierarchy, and the rostro–caudal organization of the frontal lobes.

Trends in Cognitive Sciences, 12(5), 193–200. https://doi.org/10.1016/j.tics.2008.02.004

Barense, M. D., Bussey, T. J., Lee, A. C. H., Rogers, T. T., Davies, R. R., Saksida, L. M., … Graham, K. S.

(2005). Functional specialization in the human medial temporal lobe. The Journal of

Neuroscience: The Official Journal of the Society for Neuroscience, 25(44), 10239–10246.

https://doi.org/10.1523/JNEUROSCI.2704-05.2005

Barense, M. D., Gaffan, D., & Graham, K. S. (2007). The human medial temporal lobe processes online

representations of complex objects. Neuropsychologia, 45(13), 2963–2974.

https://doi.org/10.1016/j.neuropsychologia.2007.05.023

Barense, M. D., Groen, I. I. A., Lee, A. C. H., Yeung, L.-K., Brady, S. M., Gregori, M., … Henson, R. N. A.

(2012). Intact memory for irrelevant information impairs perception in amnesia. Neuron, 75(1),

157–167. https://doi.org/10.1016/j.neuron.2012.05.014

Barense, M. D., Henson, R. N. A., Lee, A. C. H., & Graham, K. S. (2010). Medial temporal lobe activity

during complex discrimination of faces, objects, and scenes: Effects of viewpoint. Hippocampus,

20(3), 389–401. https://doi.org/10.1002/hipo.20641

Bartko, S. J., Winters, B. D., Cowell, R. A., Saksida, L. M., & Bussey, T. J. (2007a). Perceptual Functions of

Perirhinal Cortex in Rats: Zero-Delay Object Recognition and Simultaneous Oddity

Discriminations. The Journal of Neuroscience, 27(10), 2548–2559.

https://doi.org/10.1523/JNEUROSCI.5171-06.2007 Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 37

Bartko, S. J., Winters, B. D., Cowell, R. A., Saksida, L. M., & Bussey, T. J. (2007b). Perirhinal cortex

resolves feature ambiguity in configural object recognition and perceptual oddity tasks. Learning

& Memory, 14(12), 821–832. https://doi.org/10.1101/lm.749207

Beck, A. T., & Steer, R. A. (1993). BDI, Beck Depression Inventory manual. San Antonio: Psychological

Corp. : Harcourt Brace & Co.

Bialystok, E., Craik, F. I. M., & Freedman, M. (2007). Bilingualism as a protection against the onset of

symptoms of dementia. Neuropsychologia, 45(2), 459–464.

https://doi.org/10.1016/j.neuropsychologia.2006.10.009

Bialystok, E., Craik, F. I. M., Grady, C., Chau, W., Ishii, R., Gunji, A., & Pantev, C. (2005). Effect of

bilingualism on cognitive control in the Simon task: evidence from MEG. NeuroImage, 24(1), 40–

49. https://doi.org/10.1016/j.neuroimage.2004.09.044

Buckley MJ, Booth MC, Rolls ET, & Gaffan D. (2001). Selective perceptual impairments after perirhinal

cortex ablation. The Journal of Neuroscience : The Official Journal of the Society for

Neuroscience, 21(24), 9824–36.

Bussey, Saksida, L. M., & Murray, E. A. (2002). Perirhinal cortex resolves feature ambiguity in complex

visual discriminations. European Journal of Neuroscience, 15(2), 365–374.

https://doi.org/10.1046/j.0953-816x.2001.01851.x

Bussey, T. J., & Saksida, L. M. (2002). The organization of visual object representations: a connectionist

model of effects of lesions in perirhinal cortex. European Journal of Neuroscience, 15(2), 355–

364. https://doi.org/10.1046/j.0953-816x.2001.01850.x

Bussey, T. j., & Saksida, L. m. (2007). Memory, perception, and the ventral visual-perirhinal-hippocampal

stream: Thinking outside of the boxes. Hippocampus, 17(9), 898–908.

https://doi.org/10.1002/hipo.20320 Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 38

Bussey, T. J., Saksida, L. M., & Murray, E. A. (2003). Impairments in visual discrimination after perirhinal

cortex lesions: testing “declarative” vs. “perceptual-mnemonic” views of perirhinal cortex

function. The European Journal of Neuroscience, 17(3), 649–660.

Bussey TJ, Saksida LM, & Murray EA. (2005). The perceptual-mnemonic/feature conjunction model of

perirhinal cortex function. The Quarterly Journal of Experimental Psychology. B, Comparative

and Physiological Psychology, 58(3-4).

Cabeza, R., & Nyberg, L. (2000). Imaging Cognition II: An Empirical Review of 275 PET and fMRI Studies.

Journal of Cognitive Neuroscience, 12(1), 1–47. https://doi.org/10.1162/08989290051137585

Campolattaro, M. M., & Freeman, J. H. (2006). Perirhinal cortex lesions impair simultaneous but not

serial feature-positive discrimination learning., Perirhinal Cortex Lesions Impair Simultaneous

but Not Serial Feature-Positive Discrimination Learning. Behavioral Neuroscience, Behavioral

Neuroscience, 120, 120(4, 4), 970, 970–975. https://doi.org/10.1037/0735-7044.120.4.970,

10.1037/0735-7044.120.4.970

Collins DL, Neelin P, Peters TM, & Evans AC. (1994). Automatic 3D intersubject registration of MR

volumetric data in standardized Talairach space. Journal of Computer Assisted Tomography,

18(2).

Corkin, S., Amaral, D. G., González, R. G., Johnson, K. A., & Hyman, B. T. (1997). H. M.’s medial temporal

lobe lesion: findings from magnetic resonance imaging. The Journal of Neuroscience: The Official

Journal of the Society for Neuroscience, 17(10), 3964–3979.

Della-Maggiore, V., Chau, W., Peres-Neto, P. R., & McIntosh, A. R. (2002). An empirical comparison of

SPM preprocessing parameters to the analysis of fMRI data. NeuroImage, 17(1), 19–28.

Devlin, J. T., & Price, C. J. (2007). Perirhinal Contributions to Human Visual Perception. Current Biology,

17(17), 1484–1488. https://doi.org/10.1016/j.cub.2007.07.066 Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 39

Eichenbaum, H., Yonelinas, A. P., & Ranganath, C. (2007). The Medial Temporal Lobe and Recognition

Memory. Annual Review of Neuroscience, 30(1), 123–152.

https://doi.org/10.1146/annurev.neuro.30.051606.094328

Eickhoff, S. B., Laird, A. R., Grefkes, C., Wang, L. E., Zilles, K., & Fox, P. T. (2009). Coordinate-based

activation likelihood estimation meta-analysis of neuroimaging data: a random-effects approach

based on empirical estimates of spatial uncertainty. Mapping, 30(9), 2907–2926.

https://doi.org/10.1002/hbm.20718

Feinberg, L. M., Allen, T. A., Ly, D., & Fortin, N. J. (2012). Recognition memory for social and non-social

odors: Differential effects of neurotoxic lesions to the hippocampus and perirhinal cortex.

Neurobiology of Learning and Memory, 97(1), 7–16. https://doi.org/10.1016/j.nlm.2011.08.008

Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J.-P., Frith, C. D., & Frackowiak, R. S. J. (1994).

Statistical parametric maps in functional imaging: A general linear approach. Human Brain

Mapping, 2(4), 189–210. https://doi.org/10.1002/hbm.460020402

Friston, K. J., Jezzard, P., & Turner, R. (1994). Analysis of functional MRI time-series. Human Brain

Mapping, 1(2), 153–171. https://doi.org/10.1002/hbm.460010207

Friston, K. J., Zarahn, E., Josephs, O., Henson, R. N. A., & Dale, A. M. (1999). Stochastic Designs in Event-

Related fMRI. NeuroImage, 10(5), 607–619. https://doi.org/10.1006/nimg.1999.0498

Frost, J. A., Binder, J. R., Springer, J. A., Hammeke, T. A., Bellgowan, P. S. F., Rao, S. M., & Cox, R. W.

(1999). Language processing is strongly left lateralized in both sexesEvidence from functional

MRI. Brain, 122(2), 199–208. https://doi.org/10.1093/brain/122.2.199

Gauthier, I., & Tarr, M. J. (1997). Becoming a “Greeble” Expert: Exploring Mechanisms for Face

Recognition. Vision Research, 37(12), 1673–1682. https://doi.org/10.1016/S0042-

6989(96)00286-6 Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 40

Goel, V., Gold, B., Kapur, S., & Houle, S. (1998). Neuroanatomical correlates of human reasoning. Journal

of Cognitive Neuroscience, 10(3), 293–302.

Hayden, K. M., Zandi, P. P., West, N. A., Tschanz, J. T., Norton, M. C., Corcoran, C., … Welsh-Bohmer, K.

A. (2009). Effects of Family History and APOE ε4 Status on Cognitive Decline in the Absence of

AD: The Cache County Study. Archives of Neurology, 66(11), 1378–1383.

https://doi.org/10.1001/archneurol.2009.237

Hubel, D. H., & Wiesel, T. N. (1962). Receptive fields, binocular interaction and functional architecture in

the cat’s . The Journal of Physiology, 160(1), 106–154.

https://doi.org/10.1113/jphysiol.1962.sp006837

Hubel, D. H., & Wiesel, T. N. (1968). Receptive fields and functional architecture of monkey striate

cortex. The Journal of Physiology, 195(1), 215–243.

https://doi.org/10.1113/jphysiol.1968.sp008455

Knecht, S., Dräger, B., Deppe, M., Bobe, L., Lohmann, H., Flöel, A., … Henningsen, H. (2000). Handedness

and hemispheric language dominance in healthy humans. Brain, 123(12), 2512–2518.

https://doi.org/10.1093/brain/123.12.2512

Koechlin, E., Ody, C., & Kouneiher, F. (2003). The Architecture of Cognitive Control in the Human

Prefrontal Cortex. Science, 302(5648), 1181–1185. https://doi.org/10.1126/science.1088545

Kopelman, M. D., Bright, P., Buckman, J., Fradera, A., Yoshimasu, H., Jacobson, C., & Colchester, A. C. F.

(2007). Recall and recognition memory in amnesia: Patients with hippocampal, medial temporal,

temporal lobe or frontal pathology. Neuropsychologia, 45(6), 1232–1246.

https://doi.org/10.1016/j.neuropsychologia.2006.10.005

Lancaster, J. L., Tordesillas-Gutiérrez, D., Martinez, M., Salinas, F., Evans, A., Zilles, K., … Fox, P. T. (2007).

Bias between MNI and Talairach coordinates analyzed using the ICBM-152 brain template.

Human Brain Mapping, 28(11), 1194–1205. https://doi.org/10.1002/hbm.20345 Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 41

Law, J. R., Flanery, M. A., Wirth, S., Yanike, M., Smith, A. C., Frank, L. M., … Stark, C. E. L. (2005).

Functional magnetic resonance imaging activity during the gradual acquisition and expression of

paired-associate memory. The Journal of Neuroscience: The Official Journal of the Society for

Neuroscience, 25(24), 5720–5729. https://doi.org/10.1523/JNEUROSCI.4935-04.2005

Lee, A. C. H., Buckley, M. J., Pegman, S. J., Spiers, H., Scahill, V. L., Gaffan, D., … Graham, K. S. (2005).

Specialization in the medial temporal lobe for processing of objects and scenes. Hippocampus,

15(6), 782–797. https://doi.org/10.1002/hipo.20101

Lee, A. C. H., Bussey, T. J., Murray, E. A., Saksida, L. M., Epstein, R. A., Kapur, N., … Graham, K. S. (2005).

Perceptual deficits in amnesia: challenging the medial temporal lobe “mnemonic” view.

Neuropsychologia, 43(1), 1–11. https://doi.org/10.1016/j.neuropsychologia.2004.07.017

Lee, A. C. H., Scahill, V. L., & Graham, K. S. (2008). Activating the Medial Temporal Lobe during Oddity

Judgment for Faces and Scenes. , 18(3), 683–696.

https://doi.org/10.1093/cercor/bhm104

Lindquist, D. H., Jarrard, L. E., & Brown, T. H. (2004). Perirhinal cortex supports delay fear conditioning to

rat ultrasonic social signals. The Journal of Neuroscience: The Official Journal of the Society for

Neuroscience, 24(14), 3610–3617. https://doi.org/10.1523/JNEUROSCI.4839-03.2004

McIntosh, A. R., Chau, W. K., & Protzner, A. B. (2004). Spatiotemporal analysis of event-related fMRI

data using partial least squares. NeuroImage, 23(2), 764–775.

https://doi.org/10.1016/j.neuroimage.2004.05.018

Meunier, M., Bachevalier, J., Mishkin, M., & Murray, E. A. (1993). Effects on visual recognition of

combined and separate ablations of the entorhinal and perirhinal cortex in rhesus monkeys. The

Journal of Neuroscience, 13(12), 5418–5432. Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 42

Murray, E. A., & Mishkin, M. (1986). Visual recognition in monkeys following rhinal cortical ablations

combined with either amygdalectomy or hippocampectomy. The Journal of Neuroscience, 6(7),

1991–2003.

Newsome, R. N., Duarte, A., & Barense, M. D. (2012). Reducing perceptual interference improves visual

discrimination in mild cognitive impairment: Implications for a model of perirhinal cortex

function. Hippocampus, 22(10), 1990–1999. https://doi.org/10.1002/hipo.22071

Norman, M. A., Evans, J. D., Miller, W. S., & Heaton, R. K. (2000). Demographically corrected norms for

the California Verbal Learning Test. Journal of Clinical and Experimental Neuropsychology, 22(1),

80–94. https://doi.org/10.1076/1380-3395(200002)22:1;1-8;FT080

Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory.

Neuropsychologia, 9(1), 97–113. https://doi.org/10.1016/0028-3932(71)90067-4

O’Neil, E. B., Cate, A. D., & Köhler, S. (2009). Perirhinal Cortex Contributes to Accuracy in Recognition

Memory and Perceptual Discriminations. Journal of Neuroscience, 29(26), 8329–8334.

https://doi.org/10.1523/JNEUROSCI.0374-09.2009

Orban, G. A., Dupont, P., Vogels, R., Bormans, G., & Mortelmans, L. (1997). Human Brain Activity Related

to Orientation Discrimination Tasks. EJN European Journal of Neuroscience, 9(2), 246–259.

Paolo, A. M., Troster, A. I., & Ryan, J. J. (1997). California Verbal Learning Test: normative data for the

elderly. Journal of Clinical and Experimental Neuropsychology, 19(2), 220–234.

https://doi.org/10.1080/01688639708403853

Park, H.-J., & Friston, K. (2013). Structural and Functional Brain Networks: From Connections to

Cognition. Science, 342(6158), 1238411. https://doi.org/10.1126/science.1238411

Rajah, M. N., Languay, R., & Valiquette, L. (2010). Age-related changes in activity are

associated with behavioural deficits in both temporal and spatial context memory retrieval in Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 43

older adults. Cortex; a Journal Devoted to the Study of the Nervous System and Behavior, 46(4),

535–549. https://doi.org/10.1016/j.cortex.2009.07.006

Ramos, J. M. J. (2014). Essential Role of the Perirhinal Cortex in Complex Tactual Discrimination Tasks in

Rats. Cerebral Cortex, 24(8), 2068–2080. https://doi.org/10.1093/cercor/bht054

Ranganath, C. (2010). A unified framework for the functional organization of the medial temporal lobes

and the phenomenology of episodic memory. Hippocampus, 20(11), 1263–1290.

https://doi.org/10.1002/hipo.20852

Rossion, B., Hanseeuw, B., & Dricot, L. (2012). Defining face perception areas in the human brain: a

large-scale factorial fMRI face localizer analysis. Brain and Cognition, 79(2), 138–157.

https://doi.org/10.1016/j.bandc.2012.01.001

Ryan, L., Cardoza, J. a., Barense, M. d., Kawa, K. h., Wallentin-Flores, J., Arnold, W. t., & Alexander, G. e.

(2012). Age-related impairment in a complex object discrimination task that engages perirhinal

cortex. Hippocampus, 22(10), 1978–1989. https://doi.org/10.1002/hipo.22069

Saksida, L. M., Bussey, T. J., Buckmaster, C. A., & Murray, E. A. (2006). No effect of hippocampal lesions

on perirhinal cortex-dependent feature-ambiguous visual discriminations. Hippocampus, 16(4),

421–430. https://doi.org/10.1002/hipo.20170

Sands, J. D. (1981). The relationship of stressful life events to intellectual functioning in women over 65.

The International Journal of Aging & Human Development, 14(1), 11–22.

https://doi.org/10.2190/7RDC-2RUK-0GLR-J2KB

Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of

Neurology, Neurosurgery, and Psychiatry, 20(1), 11–21.

Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., … Dunbar, G. C. (1998).

The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 44

a structured diagnostic psychiatric interview for DSM-IV and ICD-10. The Journal of Clinical

Psychiatry, 59 Suppl 20, 22–33;quiz 34–57.

Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A Compendium of Neuropsychological Tests:

Administration, Norms, and Commentary. Oxford University Press.

Suzuki, W. A., Zola-Morgan, S., Squire, L. R., & Amaral, D. G. (1993). Lesions of the perirhinal and

parahippocampal cortices in the monkey produce long-lasting memory impairment in the visual

and tactual modalities. The Journal of Neuroscience: The Official Journal of the Society for

Neuroscience, 13(6), 2430–2451.

Talairach, J., & Tournoux, P. (1988). Co-planar Stereotaxic Atlas of the Human Brain: 3-dimensional

Proportional System : an Approach to Cerebral Imaging. G. Thieme.

Williams, P., & Simons, D. J. (2000). Detecting Changes in Novel, Complex Three-dimensional Objects.

Visual Cognition, 7(1-3), 297–322. https://doi.org/10.1080/135062800394829

Winters, B. D., & Bussey, T. J. (2005a). Glutamate Receptors in Perirhinal Cortex Mediate Encoding,

Retrieval, and Consolidation of Object Recognition Memory. The Journal of Neuroscience,

25(17), 4243–4251. https://doi.org/10.1523/JNEUROSCI.0480-05.2005

Winters, B. D., & Bussey, T. J. (2005b). Removal of cholinergic input to perirhinal cortex disrupts object

recognition but not spatial working memory in the rat. The European Journal of Neuroscience,

21(8), 2263–2270. https://doi.org/10.1111/j.1460-9568.2005.04055.x

Zhang, J. X., Leung, H.-C., & Johnson, M. K. (2003). Frontal activations associated with accessing and

evaluating information in working memory: an fMRI study. NeuroImage, 20(3), 1531–1539.

Zola-Morgan, S., Squire, L. R., Amaral, D. G., & Suzuki, W. A. (1989). Lesions of perirhinal and

parahippocampal cortex that spare the amygdala and produce severe

memory impairment. Journal of Neuroscience, 9(12), 4355–4370.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 45

6. Tables

Table 1 Participant means for age, education, and neuropsychological tests

Age Education (years) MMSE BDI NART

Mean 20.79 14.50 29.48 5.65 33.96 SD 2.65 1.53 .73 7.37 8.22

Note: This table presents the group means and standard errors (SE) for education and other neuropsychological measures taken. MMSE, mini-mental status examination; BDI, Beck Depression Inventory; NART, American National Adult Reading Test

Table 2: Mean accuracy (Acc) and reaction time (RT) in scanned tasks

Blobs Squares

Mean Acc(%) 0.6698 ( 0.6782 ( Mean RT(ms) 3168.44 ( 2341.79 (

Note: Accuracy values are shown as proportion correct per task type with standard error (SE). Reaction time values are shown in milliseconds (ms) per task type with SE.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 46

Table 3: Local maxima for TPLS LV 1: blobs versus squares main effect *note Talairach coordinates are rounded to the nearest whole number

Temporal BSR Spatial Talairach coordinates Hemi Gyral location BA Lag extent x y z

Lateral (Posterior 2-4 11.13 5119 38 -1 14 Right Insula) 13 2 9.75 2330 -39 -43 61 Left Postcentral 5 2,4 8.72 5098 -12 -73 -3 Left Lingual 18 3,4 7.94 5624 -59 -50 15 Left Superior Temporal 22 1,5 7.86 1656 -33 -22 56 Left Precentral 4

1-5 -8.77 1981 -38 -6 32 Left Precentral 6 2,3 -8.58 472 29 -36 -15 Right Parahippocampal 36/20* Inferior Parietal 1,5 -8.46 3217 -33 -60 41 Left Lobule 39 1,5 -8.30 3651 32 -62 39 Right 19

3 -4.28 12 -28 0 -32 Left Parahippocampal 36/20* Notes: Temporal lag represents the time after event onset, when a cluster of voxels exhibited a contrast effect of interest. The bootstrap ratio threshold was set to ±>3.28, and identified dominant and stable activation clusters. The spatial extent refers to the total number of voxels included in the voxel cluster (threshold = 10). The stereotaxic coordinates are measured in millimeters, and gyral location and Brodmann areas (BAs) were determined by referring to Talairach and Tournoux (1988). Regions marked with * were ROIs for which mean activity was extracted and plotted in a bar graph. Hemi, in which the activation occurred.

Table 4: Local maxima for Seed BPLS LV 1: Square accuracy network correlations with PRc *note Talairach coordinates are rounded to the nearest whole number

Temporal BSR Spatial Talairach coordinates Hemi Gyral location BA Lag extent x y z

2 12.20 3532 13 -90 5 Right 17 4,5, 6 8.17 96 15 -55 5 Right Lingual 18 1 7.42 83 -29 -81 -8 Left Middle Occipital 18 2,3,4,6 11.57 972 -28 -62 48 Left 7

2,3,5,6 9.40 734 -26 -8 64 Left Superior Frontal 6

Notes: Temporal lag represents the time after event onset, when a cluster of voxels exhibited a contrast effect of interest. The bootstrap ratio threshold was set to ±>3.28, and identified dominant and stable activation clusters. The spatial extent refers to the total number of voxels included in the voxel cluster (threshold = 10). The stereotaxic coordinates are measured in millimeters, and gyral location and Brodmann areas (BAs) were determined by referring to Talairach and Tournoux (1988). Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 47

Table 5: Local maxima for Seed BPLS LV 2: Blob accuracy network correlations with PRc *note Talairach coordinates are rounded to the nearest whole number

Temporal BSR Spatial Talairach coordinates Hemi Gyral location BA Lag extent x y z

1 5.42 57 22 10 -18 Right Inferior Frontal 47 6 3.87 21 23 58 30 Right Superior Frontal 9 3 3.85 30 -42 17 30 Left Middle Frontal 9 1,6,5 6.34 82 -18 35 49 Left Superior Frontal 8

4,3 5.37 86 -48 -14 49 Left Precentral 4

Notes: Temporal lag represents the time after event onset, when a cluster of voxels exhibited a contrast effect of interest. The bootstrap ratio threshold was set to ±>3.28, and identified dominant and stable activation clusters. The spatial extent refers to the total number of voxels included in the voxel cluster (threshold = 10). The stereotaxic coordinates are measured in millimeters, and gyral location and Brodmann areas (BAs) were determined by referring to Talairach and Tournoux (1988).

7. Figures Figure 1: Sample task stimuli A. Match blobs B. No match blobs

C. Match squares D. No match squares

Figure 1. Example task stimuli. There were four conditions: (A) matched blob, (B) no match blob, (C) match square, (D) no match square. The blobs were composed of three features: inner shape, outer shape, and fill pattern. For no match trials, only one of these three features differed, whereas for match trials, all features were identical. In the control task, participants decided if two rotated squares were the same size Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 48

Figure 2: Paired samples t-test comparing mean blobs and squares accuracy

Figure 2. Paired samples t-test comparing mean accuracy for blob trials and mean accuracy for square trials. No significant difference between tasks was found t(23) = .327, p>.05.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 49

Figure 3: Task specific blob (blue) vs squares (red) network A

B

200

150

100

50

0

-50 Signal % Change -100

-150

-200

BlobCorrect BlobIncorrect SqCorrect SqIncorrect

Figure 3. (A) Singular image for TPLS LV1 (blobs versus squares), at a bootstrap ratio of ±3.28, (P < 0.001), which reflects reliable activations at time lags 2–5. Red regions were activated to a greater extent for squares > blobs, while blue regions showed the opposite effect. (B) Bar graph representing mean activation with standard error bars showing activation differences specific to task, but not distinguishing between correct and incorrect trials.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 50

Figure 4: Bilateral PRc activation during blob trials

A

B

PRc ROIS 0.20

0.15

0.10

0.05

0.00 Baseline Corrected Mean % Signal Change Signal % Mean Corrected Baseline

-0.05

-0.10 BlobCorrect BlobIncorrect SquareCorrect SquareIncorrect L3620 R3620

Figure 4. (A) Singular image for TPLS LV1 (blobs versus squares) showing bilateral PRc saliences, at a bootstrap ratio of ±3.28, (P < 0.001), which reflects reliable activations at time lags 2–5. (B) Bar graph representing mean activation with standard error bars in bilateral PRc ROIs for this LV. Regions are identified by their hemisphere and . L, left; R, right

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 51

Figure 5: Square accuracy network (Seed PLS LV1) A

B

1

0.8

0.6

0.4

0.2

0 % Signal Change -0.2

-0.4

-0.6

BlobAcc BlobL36/20 BlobR36/20 SqACC SqL36/20 SqR36/20

Figure 5. (A) Singular image for Seed BPLS LV1 at a bootstrap ratio of ±3.28, (P < 0.001), which reflects reliable activations at time lags 1–5. Red regions were correlated with activity in bilateral PRc and performance on the square task. (B) Bar graph representing correlation between activity in PRc seeds and task accuracy. Regions are identified in the legend by task (Blob, blob; Sq, square), hemisphere (L, left; R, right) and Brodmann area (36/20). Accuracy is identified in the legend by task (Blob, blob; Sq, square) and “Acc”

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 52

Figure 6: Blob accuracy network (Seed PLS LV2)

A

B

1 0.8 0.6 0.4 0.2 0 -0.2

-0.4 %Signal %Signal Change -0.6 -0.8 -1 -1.2

BlobAcc BlobL36/20 BlobR36/20 SqACC SqL36/20 SqR36/20

Figure 6. (A) Saliences identified in singular image for seed BPLS LV1 at a bootstrap ratio of ±3.28, (P < 0.001). Salient regions are displayed using PLS to better visualize lateral peaks. Image reflects reliable activations at time lags 1,3,6 to capture peak activation lags for all key regions. Red regions were negatively correlated with activity in left PRc and positively correlated with performance on the blob task. (B) Bar graph representing correlation between activity in PRc seeds and task accuracy. Regions are identified in the legend by task (Blob, blob; Sq, square), hemisphere (L, left; R, right) and Brodmann area (36/20). Accuracy is identified in the legend by task (Blob, blob; Sq, square) and “Acc”.

Perception and the Medial Temporal Lobe: The Role of the Perirhinal Cortex 53

Figure 7: Simplified network schematics

A. LV 1 Seed B-PLS Network B. LV 2 Seed B-PLS Network

C. Schematic of results in the context of Badre et al.’s theory of hierarchical PFC function

Figure 7 Simplified schematic depicting two distinct networks activation correlated with PRc and supporting accuracy on square and blob tasks. (A) Key regions in network of activation positively correlated with PRc and supporting accuracy on square task. (B) Key regions in network of activation negatively correlated with left PRc and supporting accuracy on blob task. (C) Simplified schematic of Badre et al’s abstract relational hierarchy of PFC function depicting the regions implicated in each level of abstraction along an anterior to posterior gradient. The bolded anterior 3rd order region indicates where our results fall along the gradient.