Fig. 2: Biologically inspired reinforcement learning model.
a Schematic of the basal ganglia circuitry. Dopaminergic neurons in the VTA modulate plasticity at the level of the cortico-striatal synapses on SPNs in the NAc. The SPNs are subdivided depending on the dopamine receptor type they express (D1R or D2R). b Schematic of the plasticity rules of VTA-NAc circuitry1,2,3. Transient increases in dopamine, caused by bursts in firing rate of dopamine neurons, generate increases in PKA activity in D1R-expressing SPNs, leading to cortico-striatal LTP. Transient decreases in dopamine, caused by pauses in the firing rate of dopamine neurons, generate increases in PKA activity in D2R-expressing SPNs, leading to cortico-striatal LTP. c Dose-occupancy curves for the D1R and D2R describing receptor occupancies as a function of dopamine concentrations. The curves are shifted between each other due to the different affinities of the receptors. The arrows represent a 3-fold increase (burst) and decrease (pause) in dopamine concentrations, which causes left-ward or right-ward shifts of the same magnitudes in the log-scale. d Schematic and equations of a biologically inspired reinforcement learning model based on ref. 4. VTA, ventral tegmental area; NAc, nucleus accumbens; SPN, spiny projection neurons; D1R, D1-type dopamine receptor; D2R, D2-type dopamine receptor; PKA, protein kinase A; LTP, long-term potentiation. Source data provided in ‘source_data/figure_2’.
