Patterning valence specific amygdala memory by dorsal tegmental area (DTA) and ventral tegmental area (VTA) dopamine (DA) subsystems

点击翻译按钮获取中文摘要

The amygdala is a key structure for the association of Pavlovian conditioned (CS) to unconditioned (US) stimuli. The basolateral complex of the amygdala (BLA) integrates CS information from the auditory cortex and aversive US information from thalamic andsensory cortical inputs. Signals are then relayed via an inhibitory network of primarily central lateral amygdala (CEl) SST+ and PKC delta+ neurons to basal forebrain and the brainstem nuclei, thereby controlling fear behaviors (Tovote, 2016). Dopamine (DA) neurons located in the dorsal tegmental area (DTA neurons), are interconnected with the basolateral (BLA)- central amygdala (CE) circuitry. The CE-projecting DTA neurons send a prediction error coupled DAergic reinforcement signal to the CE. Importantly, this signal rewires the BLA to CEl neuronal connectivity by shifting the weight from PKC delta+ to SST+ synapses. The amygdala has been mostly investigated in aversive fear learning, but there is increasing recognition that the BLA-CE network encodes also reward behaviors. However, the specific BLA-CE circuit rearrangements underlying discriminatory associative reinforcement learning related to negative or positive experiences are not resolved.The ventral tegmental area (VTA) and the mesolimbic reward system also project to the BLA/CE network. Therefore, we propose that BLACE circuitry unifies both negative and positive associative learning by DTA and VTA coupled reinforcement signals, respectively. In support of this view, VTA neuron activity and amygdala DA levels, likely originating from the VTA, increase during reward learning. Likewise, DTA neuron activity and amygdala DA levels, in part originating from DTA cells (unpublished), are strongly increased during exposure to aversive experience. Thus, these two circuits might represent two distinctly different midbrain systems recruited during positively and negatively rated learning paradigms. Moreover, D1 vs. D2 DA receptors are asymmetrically distributed in the genetically defined neuronal subtypes. Here, a simple assumption would imply that DTA and VTA differentially innervate SST+ and PKC delta+ cells. We therefore hypothesize that negatively valenced fear and positively valenced reward signals generate memory traces that differentially map on the genetic BLA to SST+ and BLA to PKC delta+ circuit architecture. We propose that DA originating from the DTA reinforces BLA to SST+ synapses during fear learning, while DA arising from the VTA enhances BLA to CEl PKC delta+ synapses during reward learning. If we will find that aversive and rewarding stimuli affect the network in the same direction, the simple hypothesis would have to be rejected in favor of control of synaptic transmission by DA along the anatomical rostro-caudal gradients rather than along genetically defined neuronal types.