It is challenging to measure brain activity in awake toddlers because of their inability to remain still. Several studies, however, have successfully measured brain activity in typically developing toddlers under anesthesia (Kiviniemi et al., 2000), under mild sedation (Fransson et al., CHIR99021 2007), or during natural sleep (Gao et al., 2009 and Liu et al., 2008). Here, we report fMRI data acquired from 72 naturally sleeping toddlers (1–3.5 years old) who were either typically developing, language delayed, or autistic. Compared to both other groups, toddlers with autism exhibited significantly

weaker interhemispheric correlations in inferior frontal gyrus (IFG) and superior temporal gyrus (STG), two areas commonly associated with language production and comprehension. Interhemispheric synchronization strength was positively correlated with verbal ability and negatively correlated with autism severity, and it enabled accurate identification of autistic toddlers with high sensitivity (72%) and specificity (84%). These results

suggest that poor neural synchronization is a notable neurophysiological characteristic that is evident at the earliest stages of autism development and is related to the severity of behavioral symptoms. Finally, the ability to measure this characteristic NLG919 mouse during sleep, when task compliance and subject cooperation are not required, suggests its utility as a possible diagnostic measure to aid growing efforts of identifying autism during infancy Ketanserin (Pierce et al., 2009 and Zwaigenbaum et al., 2009). The data presented in this study were gathered from several studies performed at the Autism Center of Excellence (ACE) in San Diego, CA. In all scans, toddlers were presented with blocks of soft auditory stimuli that were interleaved with silence.

To ensure that the differences in synchronization between the groups were not due to differences in possible auditory-evoked responses, we first “regressed out” the experiment structure from the data of each subject (see Experimental Procedures). This ensured that there was zero correlation between each voxel’s time course and the experiment structure, effectively removing stimulus-evoked responses while leaving spontaneous fMRI fluctuations in the data (see analyses below). Spontaneous fMRI activity during natural sleep exhibited robust and spatially selective correlations between homologous locations across the two hemispheres. To demonstrate this, we sampled activity in six left hemisphere “seed” regions of interest (ROIs) and computed the correlation between each “seed” time course and the time course of every voxel in the cortex.