Selective dopaminergic reward signals in the visual cortex of monkeys, humans en Parkinson patients.

We recently observed a remarkably robust but surprising reward effect in early visual cortex of monkeys in the absence of a visual stimulus. In particular, we found that rewards without visual stimulation selectively decrease fMRI activity within the representation of a stimulus that was paired with the reward during entirely different trials. This result suggests that rewards can ‘tag’ stimulus representations in sensory cortex during periods surrounding stimulus-reward association events. This ‘tagging’ may be an attractive mechanism by which plasticity in visual cortex, as observed during perceptual learning, can be controlled. Given these unexpected but exciting results, we predict that reward signals can modulate (reduce) activity in visual cortex in a spatially-specific manner. We further hypothesize that these signals are dopaminergic-dependent, originate in the ventral midbrain and may be a correlate of a prediction error signal –hence show very similar properties as dopamine midbrain neurons as described by Schultz, Hikosaka and others. Moreover, we speculate that hypoactivity of visual cortex observed in Parkinson’s patients and the frequently observed hallucinations of these patients are related to this dopaminergic signal originating in the ventral midbrain which reduces activity in visual cortex. We here propose a number of experiments aimed to test the specificity of this reward signal, whether it can be causally linked to midbrain activity and whether it correlates with hypoactivity and the presence of hallucinations in Parkinson’s patients.