My collaborator Alan Anticevic is presenting a poster on our work at 2013 Society for Neuroscience meeting, on Monday, November 11, in the morning session. Our poster is titled: Neural circuit model for cognition-emotion interactions: Affective and task-relevant interference differentially impact working memory. This work extends our collaboration in the area of “computational neuropsychiatry” to examine interactions among brain systems during cognition, and their potential breakdown in disorders. Our abstract is below:
Neural circuit model for cognition-emotion interactions: Affective and task-relevant interference differentially impact working memory
Alan Anticevic, John D. Murray, John H. Krystal, Xiao-Jing Wang
Optimal cognitive performance critically depends on resisting external interference, but incoming distracting information can vary substantially based on its content, thus requiring distinct filtering mechanisms. For instance, survival-relevant affective information has privileged access to our awareness even during active cognitive engagement. A number of studies demonstrated that during working memory (WM), negative distraction disrupts performance and activation in frontoparietal regions and engages bottom-up regions such as the amygdala. Conversely, distraction that matches features of the WM cue, i.e. task-relevant distraction, can still disrupt performance but has a markedly different effect on neural activity. Specifically, frontoparietal regions show increased signals when task-relevant distraction is presented during WM. Distinct bottom-up and top-down mechanisms could govern the impact of different distraction on cognitive performance. However, the neural mechanisms behind such top-down and bottom-up interactions remain unexplored.
We first replicated this core observation using BOLD fMRI across two studies in 45 healthy adults. We then implemented a hypothesized mechanism using a biophysically realistic computational WM model, which we extended to the level of neural systems. The model is comprised of two spiking circuit modules, cognitive and affective. The cognitive module is characterized by representation and WM-related maintenance of task-relevant stimuli. The affective module is characterized by activation by emotional stimuli. The cognitive and affective modules interact through long-range projections that are net inhibitory, implementing a functional antagonism.
We found that this architecture can reproduce the BOLD effects across brain areas; task-relevant distractors increase BOLD signal in the cognitive module, whereas emotional distractors decrease BOLD signal in the cognitive module and increase it in the affective module. The model predicts that these two types of distractors induce distinct behavioral error patterns. Furthermore, we tested the role of top-down inhibition from the cognitive module to the affective module. We found that the strength of this projection regulates the strength of activation in the affective module by emotional stimuli and the threshold at which an emotional distractor can impair behavior. The model makes specific predictions for neural activity and behavior. We hypothesize that dysregulation of the functional antagonism between cognitive and affective areas may contribute to post-traumatic stress disorder, anxiety, and depression.