Patients with frontal lobe damage lose motivated behaviour, and become apathetic. At first glance, this fits with classical across-species neurophysiology and human fMRI findings that ventromedial prefrontal cortex (vmPFC) represents rewards. However direct human evidence for a causal role in value representation is weak.

I will present some of my studies on patients with damage to the brain area, and link the findings to computational processes this brain area might be performing. Paradoxically, vmPFC lesions amplified incentive effects, whereas striatal damage blunted them. Moreover, vmPFC damage attenuates contextual value biases, such as history effects, and biases induced by counterfactual rewards.

I will argue that vmPFC moderates striatal value representations, modulating them according to context. I will sketch a simple mathematical model of vmPFC to explain why patients might have difficulty in representing goals, and also why, counterintuitively, apathy correlates with impulsivity. Intriguingly, preliminary data seems not to align with theories where vmPFC encodes the values of future states for planning. This is consistent with the model, and also with the idea that vmPFC is essential for goal-directed behaviour.