The six seed ROIs selected for this analysis were defined in the left hemisphere according to anatomical criteria (see Experimental Procedures; see also Figure S1 and

Table S1 available online) and included the lateral prefrontal cortex (LPFC), posterior part of inferior frontal gyrus (IFG), “hand knob” area of central sulcus (CS), anterior intraparietal sulcus (aIPS), posterior part of superior temporal gyrus (STG), and lateral occipital sulcus (LO). Selecting right hemisphere ROIs would have yielded a complementary analysis with equivalent findings. Strong correlations with the seed time course were found in voxels adjacent to the location of the seed (white ellipses, Dasatinib solubility dmso Figure 1) and in voxels located in the homologous

area of selleck chemicals the contralateral right hemisphere. Note two important points. First, the voxels that exhibited correlation with each seed showed high spatial selectivity with very little overlap across seeds: this means that the spontaneous activity found for each seed and its corresponding contralateral location was relatively unique and different from that found for each of the other seeds and their contralateral locations. Second, the strength and spread of correlation in the contralateral locations are qualitatively similar across groups in all areas except for STG and IFG, which appear abnormally reduced in the autism group. Voxel-by-voxel comparisons showed that toddlers with autism exhibited significantly weaker interhemispheric correlations than both typically developing and language-delayed toddlers in the STG, a cortical area commonly associated with language processing (Figure 2). The comparisons of the Dichloromethane dehalogenase autism group to each of the other groups were independent of one another, yet both revealed significant synchronization differences only in voxels located within the STG. This analysis was performed by first computing the correlation between the time course of each left-hemisphere voxel and the time course of its corresponding contralateral right-hemisphere

voxel in each subject. This gave us an interhemispheric correlation value for each pair of corresponding left/right voxels, which signified their synchronization strength. We then performed a t test for each voxel, contrasting the correlation values across individuals of different groups. This analysis yields symmetrical results across the two hemispheres, hence the presentation of the voxel-wise group differences only on the left hemisphere. Presenting the results on the right hemisphere yields a reciprocal “mirror image. The results found in STG raised the possibility that poor interhemispheric synchronization may be a characteristic of the language system in toddlers with autism. To evaluate this further, we performed an ROI analysis in six anatomically defined ROIs that included two putative language areas, STG and IFG, and four control areas, LO, aIPS, CS, and LPFC.