Cholinergic and noradrenergic modulation of perceptual decision making and learning

Abstract

Decision making often requires the integration of noisy bottom-up sensory evidence with top-down prior expectations learned from past experiences. Acetylcholine and noradrenaline have been proposed to shape these computations, yet it remains unclear whether they directly influence the weighting of sensory evidence and prior expectations during decision making or instead shape the learning process that gives rise to these expectations. To disentangle this, healthy participants (n = 62) completed a perceptual decision-making task under the influence of either the muscarinic acetylcholine receptor antagonist biperiden, the {beta}-adrenoceptor antagonist propranolol, or placebo, in a within-subject crossover design. The task required the integration of sensory evidence with learned expectations. The reliability of these two sources of information varied independently. We show that both drugs led to a faster updating of current beliefs in response to new evidence. Under propranolol, these effects were specific to reward outcomes, whereas under biperiden, faster updating in response to both reward and non-reward outcomes resulted in less stable beliefs. Notably, neither drug affected the degree to which choices were governed by the strength of the sensory evidence. Together, these findings suggest that acetylcholine and noradrenaline guide perceptual decision making by modulating the updating of top-down prior expectations in response to new bottom-up evidence during learning.