Multiple value signals in dopaminergic midbrain and their role in avoidance contexts

The role of dopaminergic brain regions in avoidance behaviour is unclear. Active avoidance requires motivation, and the latter is linked to increased activity in dopaminergic regions. However, avoidance is also often tethered to the prospect of punishment, a state typically characterized by below baseline levels of dopaminergic function. Avoidance has been considered from the perspective of two-factor theories where the prospect of safety is considered to act as a surrogate for reward, leading to dopamine release and enhanced motivational drive. Using fMRI we investigated predictions from two-factor theory by separating the neural representation of a conventional net expected value, which is negative in the case of avoidance, from an adjusted expected value which factors in a possibility of punishment and is larger for both big rewards and big (predictably avoidable) punishments. We show that neural responses in ventral striatum and ventral tegmental area/substantial nigra (VTA/SN) covaried with net expected value. Activity in VTA/SN also covaried with an adjusted expected value, as did activity in anterior insula. Consistent with two-factor theory models, the findings indicate that VTA/SN and insula process an adjusted expected value during avoidance behaviour.