According to current concepts, motivated and addictive behaviors across species depend on the activity of the nucleus accumbens (NAc), and specifically on two groups of projection medium spiny neurons (MSNs). Those that directly inhibit the dopaminergic ventral mesencephalon (VM) (the ‘direct pathway’) express the D1 dopamine receptor (D1-MSNs), disinhibit the thalamus and promote motivated behavior. The others, considered to be solely D2-MSNs, inhibit the ventral pallidum (VP), which inhibits the VM (the ‘indirect pathway’), this results in the inhibition of the thalamus and motivated behavior (Figure 1). Importantly, these pathways were originally described for the dorsal striatum in times of relatively limited technology (Albin et al, 1989), and were assumed to be true also for the NAc.

Figure 1
figure 1

Past and novel concepts of the direct and indirect pathways of the nucleus accumbens. Left—canonical concept. D1-MSNs and D2-MSNs are completely segregated to direct and indirect pathways. Right—updated concept. Both D1-MSNs and D2-MSNs participate in direct and indirect pathways. Note that MSNs projecting to the VP may be part of a direct or indirect pathway, depending on the target of their downstream VP neuron.

PowerPoint slide

Full size image

Recent developments in cell-type-specific tools to investigate neural circuits have allowed researchers to evaluate the role of D1-MSNs and D2-MSNs in the expression of motivated and addictive behavior. As expected from the canonical circuitry, NAc D1-MSN activation promotes drug-seeking behavior while D2-MSNs had an inhibitory effect. However, other studies have shown that it is the input to the VP, rather than to the VM, which promotes reinstatement of drug-seeking behavior after withdrawal and that the VP itself, unlike its dorsal counterpart the globus pallidus, directly inhibits the thalamus (Root et al, 2015). Thus, the classic circuitry of the reward system was questioned.

In a recent study (Kupchik et al, 2015), we provided the first definitive evidence that the direct and indirect pathways of the NAc are not coded by MSN cell types (Figure 1). Using optogenetics in transgenic mice combined with neural tracing tools we showed that D1-MSNs comprise a significant portion of the classical indirect pathway by synapsing on VP neurons that project to the VM. Conversely, we showed that NAc D2-MSNs target VP neurons that innervate the thalamus directly. Thus, these D2-MSNs make a direct pathway through the VP that disinhibits the thalamus. These findings are not restricted to the ventral striatum as other emerging studies have observed similar results in the dorsal striatum (Cazorla et al, 2014; Saunders et al, 2015).

These novel findings in the ventral basal ganglia highlight the VP as a central node of the reward system. Other than the classical D1-MSN-to-VM pathway, whether a neuron is part of a direct or indirect pathway is determined in the VP. Curiously, some 15% of VP cells project to both the VM and the thalamus (Tripathi et al, 2013), converging the direct and indirect pathway on the same neuron. Thus, our perspective of the reward system needs to be more nuanced than simply through the prism of the direct and indirect pathways. Understanding this complexity requires deeper analysis of the reward system, and particularly of the understudied VP.

In conclusion, the traditional view of the reward system as being comprised of two distinct cell-type-specific pathways is an oversimplification. Emerging data requires developing a new perspective of the circuitry of the reward system that will guide future treatment strategies for addiction and other related mental health disorders.

Funding and disclosure

The authors declare no conflict of interest.