We examine hypotheses for the neural basis of the profile of visual cognition in young children with Williams syndrome (WS). These are:(a)that it is a consequence of anomalies in sensory visual processing,(b)that it is a de.cit of the dorsal relative to the ventral cortical stream,(c)that it reflects de.cit of frontal function, in particular of frontoparietal interaction, and (d)that it is related to impaired function in the right hemisphere relative to the left. The tests reported here are particularly relevant to hypotheses 2 and 3. They form part of a more extensive program of investigating visual, visuospatial, and cognitive function in large group of children with WS children, aged 8 months to 15 years. To compare performance across tests, avoiding floor and ceiling effects, we have measured performance in children with WS in terms of the "age equivalence " for typically developing children. In this article the relation between dorsal and ventral function is tested by motion and form coherence thresholds, respectively. We confirm the presence of a subgroup of children with WS who perform particularly poorly on the motion (dorsal) task. However, such performance is also characteristic of normally developing children up to 5 years; thus the WS performance may reflect an overall persisting immaturity of visuospatial processing that is particularly evident in the dorsal stream. Looking at the performance on the global coherence tasks of the entire WS group, we find that there is also a subgroup who have both high form and motion coherence thresholds, relative to the performance of children of the same chronological age and verbal age on the British Picture Vocabulary Scale, suggesting a more general global processing deficit. Frontal function was tested by a counterpointing task, ability to retrieve a ball from a "detour box," and the Stroop--like "day.night " task, all of which require inhibition of a familiar response. When considered in relation to overall development as indexed by vocabulary, the day.night task shows little specific impairment, the detour box shows a significant delay relative to controls,and the counterpointing task shows a marked and persistent deficit in many children. We conclude that frontal control processes show most impairment in WS when they are associated with spatially directed responses, reflecting a deficit of frontoparietal processing. However, children with WS may successfully reduce the effect of this impairment by verbally mediated strategies. On all these tasks we find a range of difficulties across individual children and a small subset of children with WS who show very good performance, equivalent to chronological age norms of typically developing children. Overall, we conclude that children with WS have specific processing difficulties with tasks involving frontoparietal circuits within the spatial domain. However, some children with WS can achieve similar performance to typically developing children on some tasks involving the dorsal stream although the strategies and processing may be different in the 2 groups.