Neural Substrates of Approach-Avoidance Control in Motivational Conflict

Adaptive behavior requires control over automatic tendencies to approach reward and avoid threat. Dysregulation of this control characterizes disorders such as anxiety or addiction, yet the underlying neural mechanisms remain unclear. Here, we investigated circuits engaged during approach-avoidance control using fMRI with eye-tracking and behavioral measures. Forty participants (22 females) completed an approach-avoidance conflict (AAC) task with free- and forced-choice trials. Each trial comprised anticipation, response (using a joystick to approach/obtain or avoid/forgo outcomes), and outcome phases. Before scanning, participants learned that four conditioned stimuli (CSs) predicted an aversive painful stimulation (avCS+), an appetitive monetary reward (appCS+), both outcomes (confCS+), or no outcome (neuCS-). Discordant responses (e.g., approaching avCS+) were slower than concordant responses (e.g., avoiding avCS+), confirming heightened control demands. Overcoming threat-driven avoidance specifically recruited the left inferior frontal gyrus (IFG), while suppressing reward-driven approach lacked distinct neural signatures. During anticipation, confCS+ and avCS+ showed overlapping activation in salience-control networks, including middle/anterior cingulate cortex (MCC/ACC), anterior insula, IFG, and ventral striatum (VS). Interestingly, confCS+ evoked threat-like anticipatory neurophysiological responses (pupil dilation; ACC/insula activation) but subsequently triggered reward-like approach behavior. Multivariate pattern and psychophysiological interaction analyses revealed that this dissociation was driven by differential encoding of upcoming responses in the VS during the anticipation phase and by altered functional coupling between the VS and the right temporoparietal junction (rTPJ). These findings indicate that motivational control prioritizes salience over valence and suggest a VS-rTPJ network for resolving AAC, offering novel insights into neural dynamics of flexible goal-directed behavior.