A neuro-computational model showing the effects of ventral striatum lesion on the computation of reward prediction error in VTA

International audienceOne of the earliest attempts to understand how animals learn involved pairing anunconditioned stimulus (US) with a cue or conditioned stimulus (CS) and observing thatanimals start responding to the CS after some point in time. This is the basis of pavlovianlearning, a fundamental learning mechanism in animals, which has been addressed byseveral models of neural networks. We have developed a model focusing on the mechanismof reward prediction error within pavlovian learning and studied the effects of VentralStriatum (VS) lesions to illustrate a fundamental dissociation of magnitude and timingreplicating experimental studies.The paradigm used to evaluate the model is a simple CS-US associative learning task andconsiders also how the expectation cancels out the dopamine peak at the time of thereward. The trial duration is 500 time steps with each time step corresponding to 1ms. Thestimulus is presented at the 10th time step and is kept switched on till the arrival of thereward at the 400th time step (400ms). The reward and the stimulus have by default amagnitude of 1. The number of trials for the entire conditioning to happen was 16 trials.Virtual lesions of VS to VTA GABA was made by disconnecting the link between them. Twoexperiments were conducted where the time and magnitude were varied. In the firstexperiment, a reward magnitude of 2 is given instead of 1. The reward prediction error inVTA still shows a firing of 1 indicating the magnitude is conserved even after the VS lesion. Inthe second experiment, an early reward is delivered which doesn’t show firing compared tothe control scenario where firing exists thus replicating the studies done by Takahashi(2016).The results show that there exists fundamental dissociation of magnitude and temporalwhen calculating reward prediction error (RPE) in the VTA. We propose this is achievedthrough magnitude being computed in the PPN FT neurons and time being computed in theventral striatum (VS) respectively. The implications of this model bring into light newinterpretations of dopamine firing extending to state prediction errors and creating asensory representation before learning the reward fully