How we come to process 'what' and 'where' in our visual environment : insights from typical and atypical developmental populations

Human visuoperceptual functions can be divided according to their neuroanatomical substrates: processing 'what' an object is depends largely on ventral occipitotemporal regions, whereas processing 'where' an object is relies more on dorsal occipitoparietal regions. Although much is known regarding the mature incarnation of this ventral/ dorsal dissociation, little is known about how it develops. The current study addressed this with two experiments that involved both typically-developing children and healthy adults, as well as adult individuals with Williams Syndrome (WS), a genetically-based neurodevelopmental disorder for which dissociations in ventral-dorsal stream processing have been reported. The first experiment examined behavioral performance on two matched tasks differing only in their instructions: attend to stimulus identity (face identity-matching; ventral stream) or to stimulus position (face location-matching; dorsal stream). The second experiment examined patterns of brain activation on these same tasks using functional magnetic resonance imaging (FMRI) with each of the three populations. By virtue of including both typical adults and children (chronological age, CA, and mental age, MA, controls for the WS participants) in addition to WS participants, these experiments were well poised to provide both insight into the normal development of ventral and dorsal stream processing, and critical information regarding the integrity of these systems in WS. Thirty-three individuals with WS (age M̲ = 27.2 years), 19 MA controls (age M̲ = 9.0 years), and 24 CA controls (age ̲M̲ = 20.7 years) participated in the behavioral experiment. Although overall, typical children (MA controls) responded less accurately and more slowly than adults (CA controls), like adults, their performance did not differ across the identity- and location-matching tasks. By contrast, WS participants showed a pronounced and selective deficit in location-matching when compared with MA controls, but comparable performance on identity- matching. This finding in WS provides empirical support for numerous reports of face processing skills that far exceed general spatial ability in WS. This dissociation is not observed during the course of normal development. Based on this finding, it was hypothesized that the subsequent brain imaging experiment would reveal differentially impacted function in the ventral and dorsal visual streams in WS when compared with typically- developing participants. In the second experiment, both cognitive and FMRI data were collected from a second set of participants: 15 with WS (age M̲ = 30.1 years), 16 MA controls (age M̲ = 8.9 years), and 17 CA controls (age M̲ = 31.0 years; individually age- and gender-matched to the WS participants). In accordance with their cognitive profile (performance on both tasks that had not quite reached the level of CA controls) MA controls recruited many of the same regions as CA controls, but also differed from them in certain key task-related regions. These regions included the face-preferring temporal fusiform gyrus of the left hemisphere (LH) during identity-matching and superior parietal cortex in the LH during location- matching. Given that most critical task-related regions reside in the right hemisphere (RH) (RH fusiform gyrus for face discrimination, RH parietal cortex for location processing), these findings appear to reflect protracted maturation of task-sensitive regions in the nondominant hemisphere. When compared with profiles of activation in MA and CA controls, brain response in WS participants was abnormal. In accordance with their profound location- matching deficits, WS participants showed a striking lack of location-matching activation in parietal cortex, a finding that isconsistent with a dorsal stream deficit hypothesized in this disorder. However, at variance with hypotheses of spared face discrimination and ventral stream function in WS, the brain response was also abnormal for face identity-matching, even though cognitively, WS participants performed this task at the level of MA controls. Abnormalities in WS manifested as a poorly modulated response of the fusiform gyrus, a lack of response in the amygdala, and surprisingly, an abnormally heightened response in a portion of parietal cortex that was not engaged by the task in MA or CA controls. The latter two findings are of particular interest because they may relate to an unusually strong preference for face -to-face social interaction (often described as "hypersociability") that has been described in the WS populatio