Pan: Drag the map with the mouse. Zoom: Press the ALT-key and turn the mouse-wheel.
Electrophysiology suggest that the NAcc projects mainly on the GABAergic neurons in the VTA.
The VTA projects to numerous targets which include (Beckstead et al., 1979):
Although it is principally cited for its DA output GABA also plays a major role in the activity of the VTA. GABA projections from the VTA to the NAcc are reciprocated with GABA projections to the VTA. There is also a large projection of GABA neurons from the VTA to the PFC (Carr and Sesack, 2000). Local GABA neurons can also inhibit their neighbouring dopamine neurons.(Sesack and Grace, 2010) and are a strong candidate to calculate the reward prediction error (Eshel et al., 2015).
In particular the pathway from the VTA to the NAcc core (NAcc) and NAcc shell (NAcSh) is instrumental here. VTA dopamine release in response to a rewarding stimulus induces goal-directed behaviour to acquire and consume it (Morales and Margolis, 2017).
In an experiment where DA release was artificially triggered via an optogenetic stimulation caused robust reward seeking behaviour (Steinberg et al., 2013).
DA in the VTA signals a reward prediction error resembling that of TD learning which has been first suggested by (Schultz et al., 1997) and then matched quantitatively by (Bayer and Glimcher, 2005).
DA VTA neurons react strongly to unexpected rewards, these responses diminish after repeated presentation of the reward but then rather spike when a CS is presented which predicts the reward.
During omission of the reward the DA activity supposed to experience a 'dip' in activity (Takahashi et al., 2017). However, except of a few examples it is usually a reduction of the DA response after omission.
Also the DA activity won't vanish completely after a reward is expected but is diminished. This behaviour can still be matched on TD learning when using long-lasting eligibility traces (Pan et al., 2005).
However, Sadacca et al. (2016) has recently challenged this view that DA neurons code simply a reward prediction error about an experienced reward but that they also respond to putative cached values of cues which have been previously paired with a reward.