Schizophrenia and affective disorder show different visual scanning behavior for faces: a trait versus state-based distinction?

Abstract

Background: Abnormal visual scanpaths to faces and facial expressions in schizophrenia may underlie schizophrenic subjects’ disturbed interpersonal communication. This study is the first to examine the specificity of these impairments to schizophrenia, by including an affective disorder psychiatric control group.

Methods: The visual scanpath performance of 65 schizophrenia, 52 affective disordered, and 61 control subjects were compared in two experiments. In the “face recognition” experiment, subjects viewed four identifiable (non-degraded) neutral faces versus four matched non-identifiable (degraded) control faces. In the “facial affect recognition” experiment, subjects viewed positive (happy), negative (sad), and neutral (control) facial emotion stimuli. Concurrent behavioral tasks were face matching (face recognition) and expression matching (facial affect recognition), each under two multiple-choice conditions (7 or 3 options).

Results: Scanpath disturbances were most apparent in schizophrenia subjects, who maintained a comparatively “restricted” scanpath style to all face stimuli. Schizophrenics subjects also showed the greatest recognition difficulties, particularly for neutral and happy faces. Scanpath parameters for affective disorder subjects differed only from the schizophrenia (but not the control) group, except for attention to facial features where they generally avoided facial features in all expressions and showed the greatest attentional problems of all groups for degraded faces.

Conclusions: Our results suggest that a global restriction of visual scanpaths is specific to schizophrenic psychosis and might be a trait marker for this disorder, whereas scanpath abnormalities in affective disorder might instead reflect severe state-based (or discrete) attentional disturbances.

Introduction

One of the core features of schizophrenia is the breakdown of interpersonal communication Cramer et al 1992, Kee et al 1998. Individuals with schizophrenia show impairments in the accurate identification of facial expressions, which is a particularly important component of effective social and interpersonal communication Bryson et al 1997, Burch 1995, Habel et al 2000, Young et al 1986. One of the possible mechanisms underlying disturbed face perception is a breakdown in the neurocognitive strategies that underlie processing of face stimuli. Accurate recognition of face and facial emotion stimuli relies upon the integration of basic visuomotor and visuo-spatial processes with more detailed interpersonal cues (Bruce and Young 1986). The visual scanpath (pattern of eye movements and foveal fixations) provides an overt and objective measure of these neurocognitive strategies in real time. Evidence from neurocognitive batteries points to the importance of visual scanning strategies to face and emotion perception in schizophrenia (Kee et al 1998).

A growing body of evidence indicates that people with schizophrenia have abnormal visual scanpaths to face stimuli Phillips et al 1998, Streit et al 1997, Sweeney et al 1994, Tsunoda et al 1992. Scanpaths in schizophrenia typically reflect a “restricted” scanpath strategy, characterized by longer fixation durations, shorter eye movements between fixations, and a lack of focused attention to salient facial features (eyes, mouth). A series of studies by our group has focused on particular aspects of abnormal scanpaths in schizophrenia. The restricted scanpath strategy was found to be particularly apparent for face stimuli compared to other complex stimuli, such as geometric figures and degraded (non-identifiable) faces, pointing to the face-specific nature of this disturbance Gordon et al 1992, Manor et al 1999, Williams et al 1999.

In a subsequent study, we examined the proposal that scanpath abnormalities would be even more apparent for facial expressions of emotion, given the increased processing load associated with affective appraisal. Compared to healthy control subjects, schizophrenia subjects showed a greater restriction of visual scanpaths and avoidance of facial features for happy compared to sad expressions (Loughland et al 2002). In an independent study of visual scanpaths and facial emotion, Streit et al (1997) also observed abnormally restricted scanpaths to “affect” stimuli (in an averaged analysis of six basic emotions), but the averaging procedure precluded the consideration of specific scanpath strategies in relation to individual emotions. Our observation that schizophrenia subjects show distinctive scanpath strategies for positive and negative facial emotion suggests that it may be important to consider individual emotions in studies of underlying neurocognitive mechanisms.

Schizophrenia disturbances in the visual scanning of faces have been shown to remain stable over time and across the acute and partially remitted phases of the illness Streit et al 1997, Wölwer et al 1996. An important outstanding issue is the diagnostic specificity of these disturbances to schizophrenia. Without psychiatric control data from a direct comparison with a distinct diagnostic group, it would remain unclear whether schizophrenia scanpath abnormalities to faces are simply the result of a general state of psychosis or illness severity (Archer et al 1992). To date, the isolated visual scanpath studies of other disorders (post-traumatic stress disorder and generalized anxiety disorder) have focused only on non-face threat stimuli Bryant et al 1995, Freeman et al 2000 but show no evidence for a schizophrenia-like scanpath dysfunction.

Face and facial emotion processing difficulties in affective disorder have been examined only in terms of discrimination accuracy using behavioral tasks. Some evidence suggests that individuals with schizophrenia have a greater impairment in facial affect discrimination than those with affective disorder, whereas the two groups may be similarly impaired for face recognition (Feinberg et al 1986). Other studies suggest that the facial emotion impairment in affective disordered individuals may be specific to negative emotion David and Cutting 1990, Gur et al 1992. Gaebel et al (1992) also observed that affect recognition deficits may have a trait-like quality in schizophrenia but may vary with state (e.g, are less apparent post-acute phase) in affective disorder. Importantly for the study of visual scanpaths, (Gaebel et al (1992) proposed that trait- versus state-based emotion deficits in these disorders reflect the function of distinct underlying psychobiological systems and are therefore expressed as different quantitative, qualitative, and temporal patterns of emotion-related dysfunction.

To determine the diagnostic specificity of both face and facial emotion disturbances to schizophrenia, we compared the visual scanpaths of schizophrenia, affective disorder, and healthy control groups on our previous face paradigm (Williams et al 1999) and facial emotion paradigm (Loughland et al 2002). Given that potentially confounding oculomotor disturbances have been associated with certain neurological disorders (e.g., prosopagnosia, certain forms of amnesia) and mental retardation David and Cutting 1990, Feinberg et al 1986, Fletcher et al 1997, Muir et al 1992, Tien et al 1996, we ensured that all subjects were screened for relevant comorbid pathology.

On the basis of our previous scanpath findings and evidence from behavioral studies, we predicted that schizophrenia subjects would be more impaired than both affective disorder and healthy control groups for visual scanning of face and facial expression stimuli. Affective disordered individuals, on the other hand, would show only minimal scanpath disturbances relative to healthy control subjects, and these would be limited to the more attention-based parameters (eg., allocation of fixations to features), consistent with a more state-based impairment. The pattern of concurrent recognition accuracy should parallel this pattern of visual scanning performance.