The Interplay Between Experimental Heat Pain and Noninvasive Stimulation of the Medial Prefrontal Cortex on Reinforcement Learning With Manipulated Outcome Contingencies

Abstract

Pain negatively affects several cognitive abilities, but knowledge about its effect on reinforcement learning (RL) is limited. During RL, instrumental choices can be influenced by heuristic tendencies to approach rewards or inhibit actions when facing potentially aversive events, introducing “Pavlovian bias” in behavior. Recent studies suggest that compromised outcome controllability enhances Pavlovian bias, a phenomenon that may be linked to suboptimal decision-making in learned helplessness (LH). Since LH is common in chronic pain syndromes, this study sought to establish a link between experimental heat pain (EHP), disrupted reward/loss contingencies, and RL performance in healthy adults. In addition, we investigated if intermittent theta burst stimulation (iTBS) above the medial prefrontal/dorsal anterior cingulate cortex (mPFC/dACC) alleviates the deleterious effects of EHP on choice behavior. In a preregistered, 2 × 2 between-group, double-blind study (N = 100), healthy adult participants underwent three blocks of an orthogonalized Go/NoGo task with two interleaved bouts of active or sham iTBS, and either EHP or warm skin stimulation combined with manipulated response–outcome contingency during the task. Although EHP did not impact response accuracy, it invigorated actions for rewards, reflecting enhanced Pavlovian bias. Whereas two bouts of iTBS attenuated Pavlovian tendencies, this effect was counteracted by EHP, indicating antagonistic effects of pain and iTBS-modulated mPFC activity on Pavlovian–instrumental interactions. Surprisingly, EHP and iTBS exerted largely similar effects on other latent parameters of RL (go-bias, learning rate, and exploration) in a manner that resembled LH. These findings shed light on the role of experimental pain and mPFC/dACC activity in LH-like choice behavior.