A Network for Social Interaction Understanding in the Primate Brain

4223 

Abstract Submission 

Wednesday, June 28 & Thursday, June 29 

Julia Sliwa¹, Winrich Freiwald²

¹The Rockefeller University, New York, United States, ²The Rockefeller University, New York, NY

Julia Sliwa    -  Lecture Information | Contact Me
The Rockefeller University
New York, United States

Primates continuously decode complex visual scenes into material entities, such as agents, their movements, and their interactions (Grahe and Bernieri, 1999, Ambady et al., 2000). Social interactions and their meaning are quickly recognized by monkeys: they understand grooming, play, and fight, infer social rank from interactions, and utilize this knowledge to recruit allies (Cheney et al., 1986, Bergman et al., 2003). While this understanding is a core cognitive component in primates (Grahe and Bernieri, 1999, Spelke and Kinzler, 2007) and is particularly vulnerable to social pathologies (Kennedy and Adolphs, 2012), little is known about the neural circuitry that underlies it.

To chart the brain regions that process social interactions, we presented naturalistic videos during whole-brain fMRI to four rhesus monkeys. We used three main types of social videos: 1) social interactions between monkeys, 2) monkeys engaged in non-social goal-directed actions, and 3) non-acting monkeys, along with videos of objects' 1) physical interactions, 2) moving independently; and 3) stationary, as well as low-level motion control and natural complex scene videos. Real-world videos were chosen to maximize cognitive engagement. For data analysis, we controlled for eye-movements and visual motion energy, through nuisance regression in the Generalized Linear Model (GLM) and we inspected several other behavioral factors. We searched for areas more activated by social interactions than by any other video condition using conjunction analysis. We compared these areas to canonical face, body, and object patches we mapped with a standard localizer (Tsao et al., 2003), as well as to the Mirror Neuron System (MNS) we mapped using videos of humans grasping objects (Nelissen et al., 2005).

We found that a large portion of the monkey brain is engaged by social interactions (P-corrected < 0.01, False Discovery Rate (FDR) corrected for multiple comparisons, conjunction analysis). This set includes areas located in the STS known for its involvement in analysis of social cues such as faces and bodies, and areas in the parietal-premotor cortex, which houses the MNS. But also other areas, we further refer to as the Social Interaction Network (SIN). SIN areas partially overlapped with areas described as the monkey default-mode network (DMN)(Mantini et al., 2011). Because of their anatomical location, we wondered if any of the SIN areas would display a social functional signature akin to the human DMN. For videos of social interactions only, we found a set of brain regions showing exclusive activation compared to baseline, further termed ESIN, comprised of the mPFC, ACC and dmPFC, area 7a in the inferior parietal lobe, areas of the vlPFC, and areas 10o and 14r within the OFC. We wondered which brain networks support the SIN and the ESIN. ROI analysis indicates that interactions engage face, body and object patches of the STS differentially in an interaction specific manner (Student t-tests, Holm-Bonferroni corrected for multiple comparisons using family wise error (FWE) rate at P-corrected < 0.05). Further a parieto-premotor network, co-localizing with the MNS, was activated by agent-object interactions but also by social interactions and more surprisingly by physical interactions (Student t-tests, P-corrected < 0.01 FDR-voxel wise). Functional similarity measured with PCA and correlation distances showed face patches as putative entry points to the SIN; while the MNS appeared more similar to object and body patches and differed substantially from the (E)SIN.

These results uncover the existence of a social cognition network with deep evolutionary heritage, they offer a new view on the function of the STS, and re-define the role of the mirror neuron system.

Higher Cognitive Functions Other

Non-BOLD fMRI

Mirror System

Social Cognition ¹

Social Interaction ²

FUNCTIONAL MRI

Social Interactions

Systems