Recognizing others’ moods, emotions and intentions from facial expressions lies at the core of human interpersonal communication and social engagement. This relatively automatic, visuocognitive feature that neurotypical human adults take for granted shows significant differences in children and adults with autism spectrum disorder (ASD).

In this project, Kohitij Kar plans to investigate neural circuit mechanisms that underlie such atypical behavior in autistic adults by developing a nonhuman primate (rhesus macaque) model of autism-relevant facial emotion processing. Previous research has demonstrated that state-of-the-art artificial neural network models of primate vision can be leveraged to provide neural-level hypotheses and design targeted experiments to probe brain mechanisms underlying atypical facial emotion processing in ASD^{^1}. This project combines expertise in computer vision and artificial intelligence (AI), nonhuman primate systems neuroscience, psychology and ASD research to generate knowledge highly relevant to health outcomes (specifically for autistic adults).

The proposed project has two main aims. First, Kar and colleagues plan to conduct large-scale, chronic multi-electrode recordings across the macaque’s ventral visual cortex and prefrontal cortex to identify candidate neural biomarkers of atypical facial emotion processing in ASD. They will simultaneously measure the monkey’s performance using a battery of human-testing compatible, ASD-relevant facial emotion judgment tasks. These experiments will test the explicit predictions provided by current artificial neural network models of ASD.

Second, Kar’s team plan to causally perturb the macaque inferior temporal cortex with inhibitory cell-type-specific chemogenetic probes (AAV9-hDIx-GiDREADD-dTomato-Fishell5) to test the excitatory-inhibitory imbalance theory implicated in ASD. In silico experiments have revealed that learning under noisy sensory representations could lead to ASD-like atypical facial emotion processing. Higher sensory variability in ASD is consistent with previous findings. Given that genetic mutations that impact the function of interneurons have been previously linked with ASD, Kar plans to test the hypothesis that lower inhibition (temporally induced by inhibitory interneuron-specific chemogenetic perturbation) in cortical networks leads to larger neural variability. Simultaneous behavioral and neural measurements in monkeys will test whether the increase in variability matches the ASD phenotype reported in human studies.

1. Wang S. and Adolphs R. Neurospychologia 99, 286–295 (2017) PubMed