Electrophysiological Studies of Face Processing in Williams Syndrome.

Abstract

* Williams Syndrome (WMS) is a genetically based disorder characterized by pronounced variability in performance across different domains of cognitive functioning. This study examined brain activity linked to face-processing abilities, which are typically spared in individuals with WMS. Subjects watched photographic pairs of upright or inverted faces and indicated if the second face matched or did not match the first face. Results from a previous study with normal adults showed dramatic differences in the timing and distribution of ERP effects linked to recognition of upright and inverted faces, In normal adults, upright faces elicited ERP differences to matched vs. mismatched faces at approximately 320 msec (N320) after the onset of the second stimulus. This "N320" effect was largest over anterior regions of the right hemisphere. In contrast, the mismatch/match effect for inverted faces consisted of a large positive component between 400 and 1000 msec (P500) that was largest over parietal regions and was sym metrical. In contrast to normal adults, WMS subjects showed an N320-mismatch effect for both upright and inverted faces. Additionally, the WMS subjects did not display the N320 right-hemisphere asymmetry observed in the normal adults. WMS subjects also displayed an abnormally small negativity at 100 msec (N100) and an abnormally large negativity at 200 msec (N200) to both upright and inverted faces. This ERP pattern was observed in all subjects with WMS but was not observed in the normal controls. These results may be linked to increased attention to faces in subjects with WMS and might be specific to the disorder. These results were consistent with our ERP studies of language processing in WMS, which suggested abnormal cerebral specialization for spared cognitive functions in individuals with WMS.

INTRODUCTION

The relative influence of genetic, maturational, and experiential factors on the development of cerebral specialization is a central issue in cognitive neuroscience. The unusual neurocognitive and genetic profiles in Williams Syndrome (WMS) provide an opportunity to examine how these factors interact to shape cerebral specializations for language and nonlanguage cognitive functions, Williams syndrome is a genetically based disorder characterized by remarkable sparing in some domains, such as language and face recognition, in contrast with marked deficits in other domains, such as spatial abilities (see Bellugi, Lichtenberger, Jones, Lai, & St. George, this volume). This disorder is also associated with concomitant abnormalities in brain structure, such as curtailment of the posterior-parietal and occipital regions (see Galaburda & Bellugi, this volume; Reiss et al., this volume). One approach to studying structure-function relations would be to link abnormalities in brain structure with specific cognitive def icits. For example, deficits in spatial abilities may be linked with abnormal structure (Galaburda & Bellugi, this volume) and abnormal function (Atkinson, Braddick, Nokes, Anker, & Braddick, 1997) in the dorsal-visual stream, However, this approach does not provide information about the organization of spared cognitive functions such as language and face processing.

In early neuroanatomical studies, MRI analyses of individuals with WMS suggested normal volumetric measures of frontal and cerebellar structures, which might underlie spared language abilities (Jernigan & Bellugi, 1994). The volume of gray matter in the inferior-posterior medial cortex was positively correlated with face-recognition abilities in individuals with WMS (Jones, Rossen, Hickok, Jernigan, & Bellugi, 1995). Of particular interest was whether a configuration of relatively normal brain structure in regions typically associated with the spared cognitive abilities would be indicative of normal brain function in WMS, as suggested by Bellugi, Mills, Jernigan, Hickok, and Galaburda (1999b). Alternatively, the brain systems that underlie the spared cognitive functions might be abnormally organized due to interactions with known structural abnormalities in other parts of the WMS brain (Galaburda & Bellugi, this volume; Reiss et al., this volume). From a developmental perspective, it was also important to ex amine whether the functional organization of brain systems linked to face processing in WMS might be similar to that found in normal brains at an earlier point in development. This result would indicate normal but delayed brain development. In contrast, it is also possible that WMS brains process this information in a different, perhaps unique, way.

In this study, we employed the event-related potential (ERP) technique to examine the organization of brain activity for face recognition, a spared cognitive function in WMS. We tested the hypothesis that the brain systems underlying face processing may be abnormally organized in WMS.

The Neural Basis of Abnormal Language Processing in WMS

In WMS, evidence from ERP studies of auditory language processing suggested abnormal patterns of cerebral specialization for language processing (Bellugi et al., 1999b; Bellugi, Lichtenberger, Mills, Galaburda, & Korenberg, 1999a; Mills, 1998; Neville, Mills, & Bellugi, 1994). Like face processing, auditory language comprehension and production have been shown to be remarkably spared in WMS adolescents and adults, in spite of the late onset of auditory language acquisition. Recent electrophysiological evidence (cited above) suggested that there were marked differences in the organization of language-relevant brain systems that might be unique to individuals with WMS. In normal adults and school-aged children, ERPs to closed class (i.e., grammatical function) words display a left-anterior asymmetry from at least 9 years of age. The presence of this asymmetry has been linked to performance on tests of comprehension of syntax (Neville, Coffey, Holcomb, & Tallal, 1993). Although WMS adolescents and adults have re latively spared grammatical abilities, most WMS subjects did not show the left-anterior asymmetry to closed class words. Additionally, WMS subjects showed an abnormally organized ERP response to processing semantic information in auditory sentences. Several ERP studies of normal adults and children have shown that a semantically anomalous word at the end of a sentence produces a robust negativity, called an N400, that has been linked to integration of word meaning (see Kutas & Hillyard, 1980). In adults, the visual N400 tends to be largest over posterior regions of the right hemisphere. However, in WMS subjects the N400 to semantic violations tended to be larger than normal and had a different distribution. In WMS the N400 was larger over anterior than posterior regions and was larger from the left than the right hemisphere. In summary, the ERP studies of sentence processing described here suggest that the organization of neural systems that mediate different aspects of language, a spared cognitive function i n WMS, is abnormally organized.

Face Processing and Other Spatial Abilities in WMS

Adolescents and adults with WMS have been shown to be quite adept at discriminating and learning to recognize unfamiliar faces. Behavioral studies suggest that most individuals with WMS are at, or close to, normal levels of performance on standardized tests of face processing, such as the Benton Test of Facial Recognition (Benton, Hamsher, Varney, & Spreen, 1983a), the Mooney Closure Test (Mooney, 1957), and the Warrington Recognition Memory Test (Warrington, 1984) (see Bellugi et al., 1999a; Bellugi et al., 1999b). This is in marked contrast to their impaired performance on tests of other spatial abilities. For example, most individuals with WMS are unable to match the angular orientation of two lines with lines in an array on the Benton Judgment of Line Orientation (Benton, Hamsher, Varney, & Spreen, 1983b) and perform very poorly on other form copying tasks such as the Test of Visual-Motor Integration (VMI; Beery, 1997), the block construction tasks in the Wechsler Intelligence Scale for Children-Revised (WISC; Wechsler, 1974) and in the Wechsler Adult Intelligence Scale-Revised (WAIS; Wechsler, 1981) (see Bellugi et al., this volume). Additionally, when asked to copy a line drawing of a house, WMS subjects tend to reproduce the local features, e.g., door, windows, chimney, but do not preserve the overall global configuration of the drawing (Bellugi et al., this volume). The tendency for WMS subjects to reproduce only the local elements in an array is also displayed in a hierarchical forms task (Bihrle Bellugi, Delis, & Marks, 1989). For example, when asked to copy a large "Y" comprised of smaller "H"s, WMS subjects only produce the small "H"s, i.e., the local elements. This is of particular interest in relation to face-processing abilities that are generally thought to call on global or configural processing strategies in normal adults (Farah, Wilson, Drain, & Tanaka, 1998; Farah, Tanaka, & Drain, 1995; Tanaka & Farah, 1991; Diamond & Carey, 1986; Carey, Diamond, & Woods, 1980).

The hypothesis that upright faces are processed in a global or configural manner is supported by studies showing a disproportionate inversion effect for faces over other objects (Valentine, 1988; Diamond & Carey, 1986; Yin, 1969). The decrement in performance for recognition of inverted vs. upright faces is considerably larger than for other objects such as houses or cars. A series of recent studies by Farah et al. (1998) suggests that upright faces are recognized "holistically," whereas inverted faces and other types of objects are recognized by decomposition of their parts. One might predict that if WMS subjects show a bias for local processing, they would use similar strategies for processing upright and inverted faces. That is, they might not show an inversion effect. A preliminary study with WMS adults and adolescents supported this hypothesis (Rossen, Jones, Wang, & Kilma, 1995). A recent study on children with WMS (ages 6 to 14 years) directly investigated the link between local! global strategies and performance on recognition of upright and inverted faces. In that study, most WMS children showed a preference for local processing on the hierarchical forms task. In contrast to the earlier study with adults, most WMS children showed a larger inversion effect than did normal age-matched controls (Jones, Hickok, & Lai, 1998). However, in that study, WMS children were presented with an example stimulus in the upright orientation and asked to find the matching face from an array of inverted stimuli. The mental rotation component, rather than the inversion of the stimuli, could have accounted for the increased decrement in performance. Moreover, in that study, behavioral scores on the matching task for upright and inverted faces were not correlated with scores on local/global processing. The results suggest that global processing of faces and hierarchical forms do not index the same processes.

Face-Specific Brain Mechanisms

The idea that there are brain systems specific to face recognition is supported by several lines of research. One line of evidence comes from brain-injured patients with prosopagnosia. These patients typically display an inability to recognize familiar faces without a concomitant decrement in other forms of object recognition. The lesions that produce the disorder are usually bilateral and extend along temporal and occipital cortical regions (Damasio, Tranel, & …