Network level characteristics in the emotion recognition network after unilateral temporal lobe surgery.

Abstract

The human amygdala is considered a key region for successful emotion recognition. We recently reported that temporal lobe surgery (TLS), including resection of the amygdala, does not affect emotion recognition performance (Journal of Neuroscience, 2018, 38, 9263). In the present study, we investigate the neural basis of this preserved function at the network level. We use generalized psychophysiological interaction and graph theory indices to investigate network level characteristics of the emotion recognition network in TLS patients and healthy controls. Based on conflicting emotion processing theories, we anticipated two possible outcomes: a substantial increase of the non-amygdalar connections of the emotion recognition network to compensate functionally for the loss of the amygdala, in line with basic emotion theory versus only minor changes in network level properties as predicted by psychological construction theory. We defined the emotion recognition network in the total sample and investigated group differences on five network level indices (i.e. characteristic path length, global efficiency, clustering coefficient, local efficiency and small-worldness). The results did not reveal a significant increase in the left or right temporal lobectomy group (compared to the control group) in any of the graph measures, indicating that preserved behavioural emotion recognition in TLS is not associated with a massive connectivity increase between non-amygdalar nodes at network level. We conclude that the emotion recognition network is robust and functionally able to compensate for structural damage without substantial global reorganization, in line with a psychological construction theory.