Regulation of amygdala and dopamine system in naïve and chronic alcohol states.

Chronic alcohol use disorder (AUD) is associated with a transition from reward-driven to negative affect-driven alcohol seeking. However, the circuit-level mechanisms linking motivational, affective, and stress systems remain unclear. We develop a reduced computational model that integrates the basolateral amygdala (BLA), bed nucleus of the stria terminalis (BNST), and ventral tegmental area (VTA) into a closed feedback loop to examine how chronic alcohol exposure reshapes circuit dynamics. The model includes antagonistic BLA reward- and aversion-encoding populations, CRF-positive and CRF-negative BNST interneurons and projection neurons, and VTA GABA and dopamine (DA) neurons. Simulated DA modulation of recurrent inhibition in the BLA completes the feedback loop, with core behaviors governed by BLA competition, BNST-mediated control of DA levels, and DA-modulated BLA inhibition. Parameter changes mimicking chronic alcohol exposure are constrained by experimental data and implemented as alterations in BLA→BNST connectivity, BNST excitability, and paraventricular thalamus inputs. Simulations show that these adaptations (i) shift the system from a moderately high, "safe" reward/anxiolytic state to a hypodopaminergic, anxiety-prone state with reduced BNST CRF+ projection activity and elevated BNST anxiogenic interneuron activity, and (ii) concurrently amplify cue-evoked DA transients in response to alcohol-paired stimuli while blunting responses to natural rewards. A key mechanism is the weakening of BLA reward→BNST CRF+ projections and strengthening of BLA aversion→BNST CRF- projections, which drive a sharp decrease in DA tone and collapse of differentiation between BLA reward and aversion populations into a "mixed" state. In this regime, salient cues can evoke stochastic, exaggerated reward or aversion responses, providing a mechanism for maladaptive motivation, negative affect, and potential exacerbation of mood pathology in chronic AUD.