Elsevier

Neuropsychologia

Volume 49, Issue 11, September 2011, Pages 3145-3150
Neuropsychologia

Brief Communication
Impairment of holistic face perception following right occipito-temporal damage in prosopagnosia: Converging evidence from gaze-contingency

https://doi.org/10.1016/j.neuropsychologia.2011.07.010Get rights and content

Abstract

Gaze-contingency is a method traditionally used to investigate the perceptual span in reading by selectively revealing/masking a portion of the visual field in real time. Introducing this approach in face perception research showed that the performance pattern of a brain-damaged patient with acquired prosopagnosia (PS) in a face matching task was reversed, as compared to normal observers: the patient showed almost no further decrease of performance when only one facial part (eye, mouth, nose, etc.) was available at a time (foveal window condition, forcing part-based analysis), but a very large impairment when the fixated part was selectively masked (mask condition, promoting holistic perception) (Van Belle et al., 2010a, Van Belle et al., 2010b). Here we tested the same manipulation in a recently reported case of pure prosopagnosia (GG) with unilateral right hemisphere damage (Busigny, Joubert, Felician, Ceccaldi, & Rossion, 2010). Contrary to normal observers, GG was also significantly more impaired with a mask than with a window, demonstrating impairment with holistic face perception. Together with our previous study, these observations support a generalized account of acquired prosopagnosia as a critical impairment of holistic (individual) face perception, implying that this function is a key element of normal human face recognition. Furthermore, the similar behavioral pattern of the two patients despite different lesion localizations supports a distributed network view of the neural face processing structures, suggesting that the key function of human face processing, namely holistic perception of individual faces, requires the activity of several brain areas of the right hemisphere and their mutual connectivity.

Introduction

The scientific literature about the mechanisms responsible for the remarkable ability with which humans can recognize faces has traditionally been characterized by a debate about the relative contribution of an analytical vs. holistic way of processing the visual stimulus. Analytical processing involves selecting first the most diagnostic information from the face in a local part-by-part fashion, bringing the information from different parts together only at a later stage. Evidence for this processing mode in face recognition tasks comes from studies showing that it is indeed possible to recognize faces based on parts only (e.g., Davies et al., 1977, Sadr et al., 2003) and that facial parts such as the eyes seem to be more important for face recognition than others, as shown by eye tracking studies (e.g., Yarbus, 1967) and response classification experiments (e.g., Gosselin and Schyns, 2001, Haig, 1985). However, other studies show that facial parts are processed interactively, whereby perception of one part is influenced by how we perceive the other parts, as demonstrated by the whole-part face superiority effect (Tanaka & Farah, 1993) and the composite face effect (Young, Hellawell, & Hay, 1987). Along these lines, proponents of the holistic processing view claim that faces are first perceived as a whole rather than a collection of independent parts (Galton, 1883), perhaps as a coarse global representation that can be gradually refined over time (Sergent, 1984, Watt, 1987).

Gaze-contingency is a method that was originally used to investigate the perceptual span in reading (Rayner, 1975) and later in visual scene recognition (van Diepen, Wampers, d’Ydewalle, & Underwood, 1998). Recently, introducing this method in face perception research allowed investigating the amount and type of information that could be simultaneously perceived and potentially used during a face perception task (Van Belle et al., 2010a, Van Belle et al., 2010b). More specifically, Van Belle, De Graef, Verfaillie, Busigny, et al. (2010a) used gaze-contingency to test a well-known brain-damaged patient presenting with a face-specific recognition impairment (prosopagnosia, Bodamer, 1947). In a delayed face matching task, the prosopagnosic patient PS (Rossion et al., 2003) was relatively less impaired than typical observers if she was forced to analyze the faces part-by-part through a gaze-contingent window. In contrast, if a gaze-contingent mask obstructed only her fixated part, the patient's performance decreased dramatically as compared to normal observers.

These results led to the conclusion that the cause of PS’ face recognition impairment is the inability to perceive individual faces holistically, supporting the view that holistic perception is necessary for visual expertise in processing faces.

So far, these findings were obtained in only one case of prosopagnosia, a patient who has distributed lesions, including a large lesion in the right inferior occipital cortex, in the left middle fusiform gyrus, and to a smaller extent in the right middle temporal gyrus (Sorger, Goebel, Schiltz, & Rossion, 2007). Nevertheless, there is evidence that holistic perception of individual faces is primarily a function of the right hemisphere (e.g., Hillger and Koenig, 1991, Schiltz and Rossion, 2006), and it has been suggested that an impairment of this function is a common characteristic of all cases of prosopagnosia following brain damage (Ramon, Busigny & Rossion, 2010; see also Barton, Press, Keenan, & Connor, 2002). To provide further evidence for these views, and more generally help getting a clearer view on the neural mechanisms responsible for holistic processing, the present study reports the test of the same gaze-contingency experiment on another case of acquired prosopagnosia (GG, Busigny et al., 2010). Contrary to PS, GG has a single lesion, unilaterally in the right hemisphere, which encompasses a large section of the medial section of the ventral occipito-temporal cortex including the lingual, medial fusiform, and parahippocampal gyri. Despite the extent of the brain damage, his object recognition is entirely preserved, so that, like PS, GG suffers from a face-specific visual agnosia (“pure prosopagnosia”, see Busigny et al., 2010).

Section snippets

Materials and methods

GG is a right-handed male born in 1942 who suffered from brain damage after a cerebral vascular accident in 2002. His remaining complaints are a left hemianopia and prosopagnosia. GG's object recognition and perception is intact, even for tasks requiring holistic processing of objects (for more details about GG's case, see Busigny et al., 2010).

The course of a trial in the delayed face matching task was identical to the study with PS (see Fig. 2 in Van Belle et al., 2010a). Each trial started

Behavioral data

For normal observers, performance differed significantly between conditions (main effects: accuracy rates: F(2, 12) = 8.45; p = .005, RT: F(2, 12) = 11.53; p = .0016, number of fixations: F(2, 12) = 16.97; p = .0003). Specifically, performance was higher in full view than with the mask (Accuracy: t(12) = 3.88; p = .006, RT: t(12) = 4.78; p = .0012, fixations: t(12) = 5.78; p = .0002), and the window (t(12) = 3.13; p = .022, t(12) = 2.77; p = .042, t(12) = 2.28; p = .098, respectively). Accuracy and response times in the mask and

Discussion

Like the previously reported case of acquired prosopagnosia PS, GG has a strong impairment with the central mask condition when matching/discriminating individual faces. This condition prevents using the fixated part, and thus promotes reliance on holistic perception to realize the task. In contrast, GG showed a similar performance level as controls in the window condition, the condition that forces relying on a single part at a time. This observation demonstrates again the key role of holistic

Acknowledgements

We thank all participants for their cooperation. GVB, TB and BR are supported by the Belgian National Fund for Scientific Research (FNRS). This work was supported by grants from the PRODEX program, FNRS (MIS), Action de Recherche Concertée (Belgium), and the European Space Agency of the European Union. This paper presents research results of the Belgian Network DYSCO (Dynamical Systems, Control and Optimization), funded by the Interuniversity Attraction Poles Programmes, initiated by the

References (38)

  • A.L. Benton et al.

    Impairment in facial recognition in patients with cerebral disease

    Transactions of the American Neurological Association

    (1968)
  • Bodamer, J. (1947). Die prosopagnosie. Archiv für Psychiatrie und Nervenkrankheiten, 179, 6–54 (Partial English...
  • C.M. Bukach et al.

    Perceptual expertise effects are not all or none: spatially limited perceptual expertise for faces in a case of prosopagnosia

    Journal of Cognitive Neuroscience

    (2006)
  • R. Caldara et al.

    Does prosopagnosia take the eyes out of face representations? Evidence for a defect in representing diagnostic facial information following brain damage

    Journal of Cognitive Neuroscience

    (2005)
  • J.R. Crawford et al.

    Comparing an individual's test score against norms derived from small samples

    The Clinical Neuropsychologist

    (1998)
  • M. Catani et al.

    Occipito-temporal connections in the human brain

    Brain

    (2003)
  • G. Davies et al.

    Cue saliency in faces as assessed by the ‘Photofit’ technique

    Perception

    (1977)
  • F. Galton

    Inquiries into human faculty and its development

    (1883)
  • N.D. Haig

    How faces differ—A new comparative technique

    Perception

    (1985)
  • Cited by (41)

    • Category-selective deficits are the exception and not the rule: Evidence from a case-series of 64 patients with ventral occipito-temporal cortex damage

      2021, Cortex
      Citation Excerpt :

      In contrast, patients with (pure) prosopagnosia have impaired face recognition, in the absence of low-level visual processing impairments, and normal word and object recognition (Rossion, 2018; Susilo et al., 2015). Prosopagnosia typically follows right hemisphere or bilateral vOTC damage (Barton et al., 2002; Van Belle et al., 2011). Early reviews of such single case studies showed relative independence between word and face recognition (Farah, 1991, 1992, 2004, p. 1992), with object recognition falling in between, which is still reflected in current textbook knowledge (Gazzaniga et al., 2019).

    • Can processing of face trustworthiness bypass early visual cortex? A transcranial magnetic stimulation masking study

      2020, Neuropsychologia
      Citation Excerpt :

      Faces seem to be processed differently between the two cerebral hemispheres. Previous studies have established the right hemispheric dominance for face perception in general (de Heering and Rossion, 2015; Gainotti and Marra, 2011; Pitcher et al., 2007; Rangarajan et al., 2014; Rossion et al., 2012; Van Belle et al., 2011). More specifically, some evidence suggests that especially the right hemisphere is involved in the processing of emotional expressions (Demaree et al., 2005) and face trustworthiness (Dzhelyova et al., 2012; Okubo et al., 2013).

    View all citing articles on Scopus
    View full text