Neural basis of social inference in nonhuman primates
Mehrdad Jazayeri, Ph.D. and Nancy Kanwisher, Ph.D.
Ruidong received a B.Eng in computer science from the University of Hong Kong and a Ph.D. in neurobiology and behavior from Cornell University. In his thesis, Ruidong discovered the role of ventral pallidum in song learning in the Zebra Finch. He also discovered mixed representation of vocal/non-vocal gestures in the cortical motor nucleus of the parrot budgerigar. Ruidong studied octopus behavior as a Grass Fellow in the Marine Biological Laboratory. As a postdoc fellow working with Dr. Jazayeri and with mentorship from Dr. Kanwisher, Ruidong is investigating the neural basis of social inference in non-human primates.
How do we learn from observing others’ actions? The ability to flexibly infer someone else’s intention from observations is an important aspect of our social life, and can be severely compromised in individuals with autism spectrum disorder (ASD)1–4. Yet the neural basis of rapid social inference and its dysfunction in ASD remain largely unknown. Genetic models of ASD in non-human primates (NHPs) can be the basis of novel diagnostic and therapeutic methods. To take full advantage of these models, quantitative readout of social inference is needed5,6. To uncover the neural substrate of flexible social inference, I’ve designed a two- player approach-or-avoid task where the subject is required to perform flexible one-shot learning from either experience or observation (Aim 1). In this task, the subject approaches or avoids objects that come their way. The stimulus corresponding to abstract identities (‘prey’ or ‘predator’) change every two trials. In self experiential trials, the subject can learn the identity of each new pair of stimuli on the first trial. In social observation trials, the subject sees the behavior from another animal for one trial and has one chance to perform correctly on the second trial. In isolated observation trials, the subject sees the trajectories of another agent’s performance on the screen. I will test whether animals learn from sensorimotor contingencies alone or instead learn from the intention of others (Aim 2). To distinguish between simulation and theory-based accounts of social inference, I will record from multiple brain areas related to social cognition while animals perform the task (Aim 3).
Chen, R., Puzerey, P., Roeser, A., Riccelli, T., Podury, A., Maher, K., Farhang, A., & Goldberg, J. (2019). Songbird Ventral Pallidum Sends Diverse Performance Error Signals to Dopaminergic Midbrain. Neuron
Chen, R., & Goldberg, J. H. (2020). Actor-critic reinforcement learning in the songbird. Current Opinion in Neurobiology.
Chen, R., Gadagkar, V., Roeser, A. C., Puzerey, P. A., & Goldberg, J. H. (2021). Movement signaling in ventral pallidum and dopaminergic midbrain is gated by behavioral state in singing birds. Journal of Neurophysiology
social inference, one-shot learning, non-human primates, electrophysiology