Predictive coding impairment in mouse models of autistic spectrum disorders

Theoretical work and psychophysical studies have led to the hypothesis that autism phenotypes may be manifestations of an underlying impairment in predictive abilities. Within this theoretical framework, autism is characterized by a greater weighting of sensory information in updating internal representations of the environment. With compromised prediction skills, an individual with autism faces an environment in which events occur unexpectedly and without cause.

Therefore, an insistence on sameness or stereotyped and repetitive behaviours can be seen as attempts to provide a reassuring sense of a predictable environment in a world otherwise filled with unexpected changes.

Testing the hypothesis of predictive coding requires an experimental design in which predictions are constrained experimentally. One possibility is to use learned associations between behaviour and sensory feedback. For example, sensory feedback couples in a predictable way to motor output.

Hence, the experimental assumption is that signals generated during movement that are fed back to sensory areas should constitute an experience-dependent prediction of sensory feedback. Recent studies in the mouse primary visual cortex (V1) have tested whether top-down projections to V1 carry a prediction of visual input based on motor output. Using a virtual reality environment combined with two- photon calcium imaging, these experiments have identified a neuronal circuit for visual flow predictions between the cortical area Anterior Cingulate Cortex (ACC) and V1 (Leinweber et al., 2017).

Aims

The aim of this project is to test whether visual flow predictions are disrupted in the primary visual cortex of mouse models of ASDs. We plan to use mouse models of fragile X syndrome (Fmr1 -/y ), and SYNGAP happloinsufficiency (Syngap1 +/- ), as two well established models to determine how these mutations affect the propagation of neural activity through cortical networks, focussing on sensory areas where inputs are readily controllable. Specifically, the project is organized around 3 aims:

Aim 1: To characterize movement-related activity of ACC axons in V1 of mouse models of fragile X syndrome (Fmr1 -/y ), and SYNGAP happloinsufficiency (Syngap1 +/- ).

Aim 2: To establish the contribution of measured top-down inputs to V1 activity through the development of a computational model of the V1 circuit. Aim 3. To test pharmacological treatments that rescue cortical dysfunction in Fmr1 -/y and Syngap1 +/- mice. References

Leinweber, Marcus, Daniel R. Ward, Jan M. Sobczak, Alexander Attinger, and Georg B. Keller. 2017. "A Sensorimotor Circuit in Mouse Cortex for Visual Flow Predictions." Neuron 95 (6): 1420-1432.e5. 2017.