Sixty percent of children with Tuberous Sclerosis Complex (TSC) are diagnosed with autism spectrum disorder (ASD). In this study, we investigated the neural correlates of face processing in children with TSC, using event related potentials (ERPs).We focused on face processing for two reasons. First, given the extensive literature on atypical face perception in ASD, face processing could serve as an important biomarker of ASD in this population. Secondly, face processing represents a construct that requires a combination of low-level visual processing and higher-order processing (categorization of identity), each of which could be impaired in TSC given the aberrant structural connectivity in visual projections shown in the TSC mouse model.
Our goal was to characterize face processing in children with TSC. Our guiding hypothesis was that children with TSC would show impairments in neural markers of early visual processing and in face perception, and that these differences would be more prominent in children with TSC/ASD.
We studied 19 children with TSC under age 4 and 20 age matched controls using an ERP paradigm of familiar-unfamiliar faces. Six children with TSC (32%) had ASD. Components of interest included the temporal-occipital P1 (early visual processing), N290 and P400 (face processing). A repeated measures analysis of variance of hemisphere (right, left), condition (mother, stranger) and group was performed with both amplitude and latency as dependent variables. Several different groups were analyzed: (1) Controls; (2) TSC; (3) TSC/ASD; (4) TSC/no ASD; (5) TSC with temporal lobe tubers; (6) TSC with occipital lobe tubers.
Despite extensive cortical pathology, children with TSC showed robust ERP evidence of face processing. There was a main effect of diagnostic group for the N290 latency, with the TSC group showing a longer N290 latency than controls (276 msec vs. 259 msec; p=0.05). There was also a region by group interaction (F=3.63, p=0.04), with the TSC group failing to show the expected hemispheric differences in face processing. Post-hoc analysis showed that this interaction was driven by two factors. First, there was a significant difference for N290 latency between TSC and controls in the right hemisphere [t(33.2)=2.68, p=0.01)]. Secondly, the control group showed an expected significant hemispheric difference for the N290, with N290 latency longer on the left compared to the right [t(19)=2.98; p=0.008)]. The TSC group did not demonstrate this hemispheric difference. On subgroup analysis, the longest N290 latency was seen in (1) children with ASD/TSC and (2) children with temporal lobe tubers, regardless of ASD diagnosis.
This is the first study to investigate face processing in children with TSC. These promising results demonstrate that children with TSC are slower to process faces, and that they lack the hemispheric specialization for face processing seen in typically developing children. We also characterize a trend towards slower face processing in children with TSC and ASD. Through this work, we have begun to define an electrophysiological biomarker of a perceptual domain critical for normative social development, with the ultimate goal of defining functional pathways to ASD in TSC.