Elsevier

Brain and Cognition

Volume 80, Issue 1, October 2012, Pages 33-44
Brain and Cognition

Neural adaptation across viewpoint and exemplar in fusiform cortex

https://doi.org/10.1016/j.bandc.2012.04.009Get rights and content

Abstract

The visual system has the remarkable ability to generalize across different viewpoints and exemplars to recognize abstract categories of objects, and to discriminate between different viewpoints and exemplars to recognize specific instances of particular objects. Behavioral experiments indicate the critical role of the right hemisphere in specific-viewpoint and -exemplar visual form processing and the left hemisphere in abstract-viewpoint and -exemplar visual form processing. Neuroimaging studies indicate the role of fusiform cortex in these processes, however results conflict in their support of the behavioral findings. We investigated this inconsistency in the present study by examining adaptation across viewpoint and exemplar changes in the functionally defined fusiform face area (FFA) and in fusiform regions exhibiting adaptation. Subjects were adapted to particular views of common objects and then tested with objects appearing in four critical conditions: same-exemplar, same-viewpoint adapted, same-exemplar, different-viewpoint adapted, different-exemplar adapted, and not adapted. In line with previous results, the FFA demonstrated a release from neural adaptation for repeated different viewpoints and exemplars of an object. In contrast to previous work, a (non-FFA) right medial fusiform area also demonstrated a release from neural adaptation for repeated different viewpoints and exemplars of an object. Finally, a left lateral fusiform area demonstrated neural adaptation for repeated different viewpoints, but not exemplars, of an object. Test-phase task demands did not affect adaptation in these regions. Together, results suggest that dissociable neural subsystems in fusiform cortex support the specific identification of a particular object and the abstract recognition of that object observed from a different viewpoint. In addition, results suggest that areas within fusiform cortex do not support abstract recognition of different exemplars of objects within a category.

Highlights

► We examined adaptation for common objects across viewpoint and exemplar using fMRI. ► Adaptation for objects was viewpoint and exemplar specific in the fusiform face area. ► Adaptation was also viewpoint and exemplar specific in a right medial fusiform area. ► In a left lateral fusiform area, adaptation was viewpoint abstract but exemplar specific. ► Findings are partially in line with dual subsystems theory of object recognition.

Introduction

The visual system has the remarkable ability to categorize a wide array of visual input forms despite the lack of uniform appearance, as well as make fine grain discriminations between highly similar inputs. Although two exemplars of a cup may produce different visual inputs, it is still possible to classify both as cups, while also distinguishing one particular cup from another. Likewise, it is possible to identify the same exemplar of a cup presented at a different viewpoint despite different input forms, however different viewpoints may also be easily distinguished from each other. To this end, object recognition can be achieved by two processes: abstract and specific. Abstract recognition refers to the classification of two objects being the same despite dissimilar appearances, and involves the mapping of different input forms, even somewhat dissimilar ones, to the same output form. Abstract recognition not only allows one to classify different exemplars of an object as belonging to a common category, but it also mediates processes that make it possible to recognize the same exemplar across different viewpoints. Specific object recognition refers to the identification of an object based on its specific form, and involves the mapping of different input forms, even somewhat similar ones, to different output forms. Specific recognition allows one to discriminate similar exemplars within a category, such as individual faces, and is also sensitive to changes in viewpoint.

Abstract and specific recognition processes may be measured behaviorally through priming effects (e.g., Burgund and Marsolek, 2000, Marsolek, 1999) and neurally through adaptation effects (e.g., Koutstaal et al., 2001, Vuilleumier et al., 2002). With both techniques, comparisons between the measure (performance or neural activity) obtained for test objects preceded by the exact same object and test objects preceded by an object differing from the test object on the property of interest, such as exemplar or viewpoint, reveal the degree to which generalization across the manipulated property occurred. Abstract recognition is evidenced by equivalent performance or neural activity for test objects preceded by identical objects and test objects preceded by objects that differ with respect to a property of interest (e.g., viewpoint). In contrast, specific recognition is evidenced by a difference in performance or neural activity between test objects preceded by identical and different objects.

Studies examining behavioral priming with divided-visual-field presentations have observed greater exemplar-specific priming (Burgund and Marsolek, 1997, Marsolek, 1999, Marsolek et al., 1992) and viewpoint-specific priming (Burgund & Marsolek, 2000) during right hemisphere presentations than left, suggesting that dual subsystems, operating with different relative efficiencies in the right and left hemispheres, support specific and abstract recognition processes, respectively. Functional magnetic resonance imaging (fMRI) studies examining neural adaptation for same compared to different-exemplar or -viewpoint objects suggest that areas within the fusiform cortex in the temporal lobe are critical for specific and abstract recognition (Andresen et al., 2009, Chouinard et al., 2008, Davies-Thompson et al., 2009, Ewbank and Andrews, 2008, Fairhall et al., 2011, Fang et al., 2007, James et al., 2002, Kim et al., 2009, Koutstaal et al., 2001, Mur et al., 2010, Pourtois et al., 2005a, Pourtois et al., 2005b, Pourtois et al., 2009, Simons et al., 2003, Vuilleumier et al., 2002, Xu et al., 2009), however results from these studies are only partially consistent with the dual subsystems hypothesis. As such, the different roles that areas within fusiform cortex play in specific and abstract visual form recognition remain to be clarified.

The fusiform face area (FFA) is an area within the right posterior, lateral fusiform that exhibits a greater neural response to faces than to other non-face objects (Kanwisher et al., 1997, Kanwisher et al., 1998, Loffler et al., 2005, McCarthy et al., 1997, Yovel and Kanwisher, 2004). Critically, neural adaptation in the FFA is exemplar and viewpoint specific: activity decreases are greater for repetitions of the same face compared to a different face or a different picture of the same face (exemplar-specific adaptation; Davies-Thompson et al., 2009, Xu et al., 2009) and greater for repetitions of the same view of a face than a different view of the same face (viewpoint-specific adaptation; Ewbank and Andrews, 2008, Fang et al., 2007, Pourtois et al., 2005a, Pourtois et al., 2005b, Pourtois et al., 2009, Xu et al., 2009). Moreover, studies examining adaptation for common objects have observed exemplar-specific effects (Chouinard et al., 2008, Ewbank et al., 2005, Fairhall et al., 2011, Kim et al., 2009, Koutstaal et al., 2001, Simons et al., 2003, Vuilleumier et al., 2002) and viewpoint-specific effects (Andresen et al., 2009, Pourtois et al., 2009, Vuilleumier et al., 2002) in lateral fusiform areas overlapping with the typical FFA. These results, as well as others (e.g., Gauthier, Tarr, et al., 2000; Haxby, Gobbini, Furey, & Ishai, 2001), support the idea that the FFA supports specific object recognition, in line with the dual subsystems hypothesis.

More controversial is the extent to which areas within the fusiform cortex generalize across object exemplars and viewpoints. Some studies have observed neural adaptation for different object exemplars (Dehaene et al., 2001, Dehaene et al., 2004, Koutstaal et al., 2001, Simons et al., 2003) and viewpoints (Andresen et al., 2009, Pourtois et al., 2005a, Vuilleumier et al., 2002) in the left fusiform, in line with the dual subsystems hypothesis. However, other studies have produced less consistent results. For example, Vuilleumier et al. (2002) observed a release from adaptation for different exemplars of objects in left lateral fusiform, demonstrating exemplar-specific rather than exemplar-abstract recognition. In addition, neural adaptation across different viewpoints has been observed in right lateral fusiform (James et al., 2002), right medial fusiform (Pourtois et al., 2005b, Pourtois et al., 2009), and in the FFA (Ewbank and Andrews, 2008, Mur et al., 2010). Thus, it is unclear to what extent the dual subsystem theory accurately characterizes fusiform function, particularly with regard to abstract processing.

In the present study, we attempt to further clarify the roles of different regions within the fusiform cortex with regard to specific and abstract recognition by examining adaptation across viewpoint and exemplar changes. Subjects were exposed to particular views of common objects and then tested with objects appearing in four critical conditions: same-exemplar, same-viewpoint adapted (SE), same-exemplar, different-viewpoint adapted (DV), different-exemplar adapted (DE), and not adapted (NA; see Fig. 1). Subjects also completed an FFA localizer task in which they viewed pictures of faces and houses (see Fig. 2). Regions of interest were defined in two ways. An FFA was defined by comparing activity during face and house trials during the localizer task—a region within fusiform cortex exhibiting greater activity for faces than houses was considered the FFA. In addition, more general object processing areas were defined by comparing activity during NA and SE trials to find regions exhibiting adaptation. We reasoned that regions exhibiting adaptation for repeated objects are critically involved in object processing (see also, Grill-Spector and Malach, 2001, Grill-Spector et al., 1999); thus this method was appropriate for our goal.1 Adaptation across different viewpoints and exemplars of objects was then examined in each of these regions of interest. Previous studies have focused on neural adaptation across viewpoints (Andresen et al., 2009, Ewbank and Andrews, 2008, Fang et al., 2007, James et al., 2002, Mur et al., 2010, Pourtois et al., 2005a, Pourtois et al., 2005b, Pourtois et al., 2009) or exemplars (Chouinard et al., 2008, Fairhall et al., 2011, Kim et al., 2009, Koutstaal et al., 2001, Simons et al., 2003), and few studies have examined adaptation across viewpoint and exemplar within the same experiment (but see, Vuilleumier et al., 2002, Xu et al., 2009). The present method will allow viewpoint and exemplar adaptation to be compared directly.

In addition, the present study examined the effect of task demands on fusiform adaptation by using two tasks at test—an object-decision task (i.e., decide if image depicts real object or not; e.g., Vuilleumier et al., 2002) and a size-judgment task (i.e., decide if real-life referent of depicted object would fit within a certain size box; e.g., Fairhall et al., 2011, Horner and Henson, 2011, Koutstaal et al., 2001, Simons et al., 2003). Task demands have been shown to affect fusiform function (Chouinard et al., 2008, Cohen Kadosh et al., 2009, Gauthier et al., 1997, Gauthier et al., 2002, Joseph et al., 2006, Large et al., 2007), and these particular tasks have lead to different conclusions regarding adaptation across object exemplars in the left lateral fusiform. Specifically, some experiments employing the size-judgment task have observed adaptation across exemplars in the left lateral fusiform (Koutstaal et al., 2001, Simons et al., 2003; but see, Fairhall et al., 2011), however an experiment employing the object-decision task observed a release from adaptation in this same region (Vuilleumier et al., 2002). Critically, these tasks differ in terms of the depth of processing required by each. In particular, the size-judgment task requires semantic processing in addition to visual processing, whereas the object-decision task can be performed on the basis of visual familiarity alone. It is possible that semantic processing is required for different-exemplar adaptation effects to manifest. Thus, the inclusion of both tasks in the present experiment allows us to examine this possibility.

Finally, unlike the majority of previous studies examining specific and abstract object recognition in fusiform cortex, which have used the same task for adaptation and test phases of the experiment (e.g., Chouinard et al., 2008, Fairhall et al., 2011, Fang et al., 2007, James et al., 2002, Koutstaal et al., 2001, Pourtois et al., 2005a, Pourtois et al., 2005b, Pourtois et al., 2009, Simons et al., 2003, Vuilleumier et al., 2002; but see, Andresen et al., 2009, Ewbank and Andrews, 2008), the present study used different tasks in adaptation and test phases. Critically, the use of different tasks during adaptation and test phases avoids the potential contamination of adaptation effects by response learning effects—increases in adaptation due to the association between stimulus and response learned during an earlier exposure (Dobbins, Schnyer, Verfaellie, & Schacter, 2004). Although the extent to which adaptation in fusiform cortex is affected by response learning is debated (e.g., Dobbins et al., 2004, Horner and Henson, 2008, Horner and Henson, 2011, Wig et al., 2009), the present study aims to obtain a clearer view on adaptation in the fusiform by eliminating the possibility of response learning effects.

Section snippets

Design

A 4 × 2 mixed factorial design was employed in which adaptation type (same-exemplar, same-viewpoint adapted [SE], same-exemplar, different-viewpoint adapted [DV], different-exemplar adapted [DE], and not adapted [NA; see Fig. 1]) was a within-subject variable and task at test (object decision, size judgment) was a between-subjects variable.

Subjects

Twenty volunteers (10 male; mean age 21) from Rice University and Baylor College of Medicine participated in the study. All subjects completed a screening

Behavioral

Error rates and correct response times within 2.5 standard deviations of the grand mean were each analyzed in a 2-way repeated-measures ANOVA with adaptation type (SE, DV, DE, and NA) as a within-subjects variable and task (object decision, size judgment) as a between-subjects variable. Error rates were low (∼10%) and did not exhibit any significant effects (all ps > .345), however the general trend was in line with typical priming effects (see Fig. 3A). The analysis of response times revealed a

Discussion

In this study, we investigated the role of the fusiform cortex in the mediation of two object recognition processes: the specific recognition of a particular object exemplar from a particular viewpoint, and the abstract recognition of different object viewpoints and exemplars. Critically, this research explored the extent to which regions within the fusiform cortex, including the functionally defined FFA, differentially support recognition processes specific to object viewpoint and category and

References (62)

  • K.J. Hayworth et al.

    Neural evidence for intermediate representations in object recognition

    Vision Research

    (2006)
  • A.J. Horner et al.

    Priming, response learning and repetition suppression

    Neuropsychologia

    (2008)
  • T.W. James et al.

    Differential effects of viewpoint on object-driven activation in dorsal and ventral streams

    Neuron

    (2002)
  • J.E. Joseph et al.

    FMRI correlates of cortical specialization and generalization for letter processing

    Neuroimage

    (2006)
  • N. Kanwisher et al.

    The effect of face inversion on the human fusiform face area

    Cognition

    (1998)
  • J.G. Kim et al.

    Adaptation to objects in the lateral occipital complex (LOC): Shape or semantics?

    Vision Research

    (2009)
  • W. Koutstaal et al.

    Perceptual specificity in visual object priming: Functional magnetic resonance imaging evidence for laterality difference in fusiform cortex

    Neuropsychologia

    (2001)
  • M. Large et al.

    Task-related laterality effects in the lateral occipital complex

    Brain Research

    (2007)
  • C.J. Marsolek et al.

    Identifying objects impairs knowledge of other objects: A relearning explanation for the neural repetition effect

    Neuroimage

    (2010)
  • G. Pourtois et al.

    View-independent coding of face identity in frontal and temporal cortices is modulated by familiarity: An event-related fMRI study

    Neuroimage

    (2005)
  • J.S. Simons et al.

    Neural mechanisms of visual object priming: Evidence for perceptual and semantic distinctions in fusiform cortex

    Neuroimage

    (2003)
  • A.Z. Snyder

    Difference image vs. ratio image error function forms in PET realignment

  • A. Soldan et al.

    Neural mechanisms of repetition priming of familiar and globally unfamiliar visual objects

    Brain Research

    (2010)
  • X. Xu et al.

    Adaptation in the fusiform face area (FFA): Image or person?

    Vision Research

    (2009)
  • G. Yovel et al.

    Face perception: Domain specific, not process specific

    Neuron

    (2004)
  • E. Zarahn et al.

    Empirical analyses of BOLD fMRI statistics I. Spatially unsmoothed data collected under null-hypothesis conditions

    Neuroimage

    (1997)
  • E.D. Burgund

    Viewpoint dependence in dissociable specific and abstract visual-form subsystems

    Dissertation Abstracts International: Section B: The Sciences and Engineering

    (2000)
  • E.D. Burgund et al.

    Viewpoint-invariant and viewpoint-dependent object recognition in dissociable neural subsystems

    Psychonomic Bulletin and Review

    (2000)
  • K. Cohen Kadosh et al.

    Task-dependent activation of face-sensitive cortex: An fMRI adaptation study

    Journal of Cognitive Neuroscience

    (2009)
  • J.D. Cohen et al.

    PsyScope: An interactive graphic system for designing and controlling experiments in the psychology laboratory using Macintosh computers

    Behavioral Research Methods, Instruments, and Computers

    (1993)
  • S. Dehaene et al.

    Cerebral mechanisms of word masking and unconscious repetition priming

    Nature Neuroscience

    (2001)
  • Cited by (18)

    • Visuo-haptic object perception

      2019, Multisensory Perception: From Laboratory to Clinic
    • Beyond the FFA: Brain-behavior correspondences in face recognition abilities

      2017, NeuroImage
      Citation Excerpt :

      In addition, there is support for the notion of smaller patches of distributed representations in the ventral visual pathway (Haxby et al. 2000; Kriegeskorte et al. 2007; Axelrod and Yovel 2015). There is also evidence from fMRI-adaptation paradigms that there are independent populations of neurons that encode information about the whole face, face parts, and global head shape (Betts and Wilson 2010; Harris and Aguirre 2010) as well as separate populations of viewpoint-dependent neurons distributed throughout the FG (Harvey and Burgund 2012). Integration of these independent populations of neurons could functionally influence the ability to form more invariant representations of face identity, but also manifest as larger patches of category-selective functional tissue (as defined by typical localizer contrasts).

    • Role of features and categories in the organization of object knowledge: Evidence from adaptation fMRI

      2016, Cortex
      Citation Excerpt :

      In order to ensure that subjects engaged in deep semantic processing for each word pair, we asked subjects to make a decision about which object they encounter more frequently in their daily life when they see a question mark after randomly selected word pair presentations. In order to eliminate the impact of the question on word presentation, we did not present the question for all word pairs, but instead we presented the question for eight word pairs randomly distributed across each run (2 in each condition per run, excluding the identical and the corresponding unrelated conditions) resulting in 12 questions per condition for taxonomically related/unrelated and thematically related/unrelated conditions across all runs (for a similar task see Harvey & Burgund, 2012; see Fig. 1). In addition, we used longer jittered delays (4, 6, or 8 sec) between the question and next word pair.

    • Separability of abstract-category and specific-exemplar visual object subsystems: Evidence from fMRI pattern analysis

      2015, Brain and Cognition
      Citation Excerpt :

      Each of these procedural aspects (a single study-test cycle with a longer study-to-test interval, briefer test presentations, different encoding and test tasks, and a more tightly controlled measure of AC priming) comprise changes from the previous fMRI experiments that, while associated with the indicated inferential advantages, likely led to weaker overall BOLD signal differences between conditions. In fact, when Harvey and Burgund (2012) used both a single study-test cycle and different encoding and test tasks in a similar object priming study, they found no hemisphere difference in SE priming in univariate analyses. For these reasons, we used MVPA to enable more sensitive statistical testing.

    • Fusing concurrent EEG-fMRI with dynamic causal modeling: Application to effective connectivity during face perception

      2014, NeuroImage
      Citation Excerpt :

      A different pattern emerged for inverted faces. In all of our models, face input (upright and inverted) was always directed via the OFA and non-face input (chairs and scrambled images) via the mFG, an area known to be more involved in object processing (Ishai et al., 1999; Pourtois et al., 2009; Harvey and Burgund, 2012). Our best fitting models showed that the functional connection from the OFA to the mFG was significantly enhanced in response to inverted faces, explaining the greater activation commonly observed in the mFG for inverted compared with upright faces (Aguirre et al., 1999; Haxby et al., 1999; Epstein et al., 2006).

    View all citing articles on Scopus
    View full text