The sensory and cognitive features of autism spectrum disorders (ASDs) must have origins in altered neural coding and information processing. But what these neural coding changes are, and whether they are shared in different forms of ASD, are unknown.

Transgenic mouse models of ASD provide an opportunity to study neural coding defects and the circuit mechanisms that underlie them. A canonical model is that ASD involves circuit hyperexcitability and excess spiking, which may contribute to sensory hypersensitivity in some forms of ASD.

Dan Feldman and his team recently tested this hypothesis in the somatosensory cortex of several genetic mouse models of autism and, surprisingly, found that spiking was normal or slightly reduced, not elevated¹. Thus, hyperexcitable circuits may not be the common basis for cortical dysfunction in ASD and additional hypotheses are needed. These may include blunted sensory tuning, disrupted adaptation, blurred maps, changes in brain rhythms or reduced signal-to-noise in population activity, all of which could degrade population coding without adding excess spikes.

In the current project, the Feldman lab plans to survey for neural coding abnormalities in the somatosensory cortex in multiple genetic mouse models of ASD in order to identify novel coding changes. They will focus initially on Fmr1^-/y, Cntnap2^-/-, and Tsc2^+/- mice, extending to other ASD mouse models to define the scope of common coding phenotypes. Dense multi-site spike recordings and 2-photon calcium imaging will be used to assess neural coding in awake mice under sparse and dense sensory loads. Coding deficits that emerge under specific sensory conditions can suggest underlying circuit mechanisms, which will also be investigated.

Findings from these studies may identify coding changes that are common across multiple genetic forms of ASD or may reveal a wide set of unique disruptions. Cortical dysfunction under dense sensory conditions could relate to sensory overload or gating phenotypes in autism. An accurate understanding of coding changes is essential to test future therapeutic strategies aimed at normalizing cortical function.