Pan: Drag the map with the mouse. Zoom: Press the ALT-key and turn the mouse-wheel.

Subsections


3 Nucleus Accumbens (NAcc)

The Nacc is the ventral extension of the striatum, and therefore also called the ventral striatum. The ventral striatum contains neurons known as medium spiny neurons (MSN's).

The accumbens has two major subterritories (Heimer et al., 1991): the shell (section 3.5) and the core (section refcore) where the shell can be further subdivided (Usuda et al., 1998) in lateral and medial parts and the core rather depending on its D1 or D2 receptors.

3.1 Behavioural experiments

The Nacc core and shell have distinct roles controlling reward based learning:

In reversal learning (Dalton et al., 2014) the shell seems to control switching contingencies (i.e. the reward is moved from one site to another) whereas the core controls the actual approach behaviour towards the rewarding site.

Impulsivity is altered by 5HT antagonists which point towards an important role of 5HT in the retrieval of delayed rewards. 5-HT(2A) antagonists reduced decreased impulsive responding and the 5-HT(2C) antagonist increased impulsivity (Robinson et al., 2008).

3.2 Signals

By measuring the DA concentration in both the core and shell (Saddoris et al., 2015) it turns out that Core DA follows the classical prediction error signals were it spikes most to the predicting cue whereas the shell responds to all reward related events during the experiment.

3.3 Plasticity

It's well known the bursts of dopamine cause LTP in conjunction with pre- and (possible) postsynaptic activity (so called 3 factor Hebbian rule) or heterosynaptic LTP. However Goto and Grace (2005) showed that D1 receptors cause LTP on hippocampal fibres whereas D2 receptors control the cortical inputs in the opposite way. It appears to be that 5HT causes retrograde cannabinoid (CB) release which inhibits pre-synaptic GLU release (Burattini et al., 2014). However, also postsynaptic HFS causes CB1 mediated depression. Mathur et al. (2011) tested this more thoroughly in that they conclude that 5HT actually causes presynaptic inhibition via the 5HT1B receptor, with that LTD and that HFS can have a similar effect by releasing CB.


3.4 Nucleus Accumbens Core

3.4.1 Afferents

The NAcc core receives inputs from the dorsal-medial prefrontal cortex and the hippocampus (Brog et al., 1993a).

3.4.2 Efferents

There are two distinct output pathways from the NAcc core which have its origins from the two sub-populations of neurons in the NAcc core. The one sub-population carries mainly D1 receptors and the other one carries mainly D2 receptors (Humphries and Prescott, 2010; Kelley, 2004).

3.4.2.1 Direct pathway

The D1 receptor carrying neurons feed directly into the SNr and are able to inhibit tonically active SNr neurons, thus the NAcc core is able to disinhibit motor programs.

3.4.2.2 Indirect pathway

There is an indirect pathway via the VP to the SNr originating from the NAcc core. In contrast to the direct pathway these neurons in the indirect pathway carry mainly D2 receptors which are inhibitory in nature. D2 receptors are very sensitive to low DA concentrations and will react to the tonic DA concentrations.


3.5 Nucleus Accumbens Shell

The shell can be further divided into the medial and lateral shell (Ikemoto, 2007; Usuda et al., 1998) used in the model by (Humphries and Prescott, 2010).

3.5.1 Medial Shell

The medial Shell projects to the medial Ventral Pallidum (VP) (Ikemoto, 2007).

3.5.2 Lateral Shell

The lateral shall projects to the ventrolateral Ventral Pallium (VP) (Ikemoto, 2007).

3.5.3 Behavioural experiments

The shell seems to be responsible for behavioural flexibility meaning that it controls the animal's ability to shift to another target when the reward is lost (Aquili et al., 2014)

3.5.4 Function

Overall the Shell seems to learn the stimuli which are associated with a reward, and thus enhances the future salience of those stimuli (Cassidy and Tong, 2017).

In this context it is also interesting that shell DA also tracks rather the inventive value than the reward prediction error (Sackett et al., 2017) which probably means that the VTA has regions which work with the shell and that they are distinct from the core.