Primary somatosensory cortical processing in tactile communication

Touch is an essential form of non-verbal communication. While language and its neural basis are widely studied, tactile communication is less well understood. We used fMRI and multivariate pattern analyses in pairs of emotionally close adults to examine the neural basis of human-to-human tactile communication. In each pair, a participant was designated either as sender or as receiver. The sender was instructed to communicate specific messages by touching only the arm of the receiver, who was inside the scanner. The receiver then identified the message based on the touch expression alone. We designed two multivariate decoder algorithms – one based on the sender’s intent (sender-decoder), and another based on the receiver’s response (receiver-decoder). We identified several brain areas that significantly predicted behavioral accuracy of the receiver. Regarding our a priori region of interest, the receiver’s primary somatosensory cortex (S1), both decoders were able to accurately differentiate the messages based on neural activity patterns here. The receiver-decoder, which relied on the receivers’ interpretations of the touch expressions, outperformed the sender-decoder, which relied on the sender’s intent. Our results identified a network of brain areas involved in human-to-human tactile communication and supported the notion of non-sensory factors being represented in S1.

Log files per subject and run (end of file name: subID_run_log.csv)

response mat file per subject receiver_only and sender_only

readresponsemat_SVM_S1...m = reads in response mat files and performs SVM classification

ECOC_S1.m = decoding code

Normalized brain scan data can be found here: https://zenodo.org/records/4925648