Figure 10:

Simulation 2a: Extinction and rapid reacquisition. (a) Empirical learning curves for initial acquisition (lower curve) and reacquisition (upper), documenting rapid reacquisition, from Ricker & Bouton (1996), with permission from Springer Nature: Animal Learning & Behavior, copyright 1996. (b) Simulation results showing the evolution of dopamine signaling over a sequence of acquisition, extinction, and reacquisition; CS-onset dopamine = solid line; US-onset = dotted line; (c-e) Focus on network activity in the amygdalar layers after acquisition training (c), extinction (d), and reacquistion (e). Initial acquisition is mediated by BLAmygPosD1 and CElAcqPosD1 D1-dominant cells, while extinction drives opponent BLAmygPosD2 and CElExtPosD2 D2-dominant cells (learning via dopamine dips). Extinction takes longer due to the need for learning in extinction cells to out-compete the acquisition cells. Reacquisition is fast because the original acquisition weights are largely intact, and the relative balance can be rapidly shifted.