input from more dorsal areas of the medial prefrontal cortex, while the shell receives input from more ventral regions (Groenewegen et al. 1999). These anatomical differences may contribute to behavioral differences. Whereas the core is hypothesized to contribute to stimulus-appropriate instrumental responding (Ikemoto 2007), the shell is hypothesized to contribute to stimulus-outcome learning (Ikemoto 2007) and the direct reinforcing effects drugs (Everitt & Robbins 2005). Prior studies have found that cocaine (Rodd-Henricks et al. 2002; Ikemoto 2003), amphetamine (Ikemoto, Qin & Liu 2005) and Δ9THC (Zangen et al. 2006) are self-administered directly into the medial shell. Most relevant to the present report, a recent paper by Engelman and colleagues found that ethanol is also self-administered directly into the medial shell (Engleman et al. 2009). Therefore, the shell and not the core may be a region in which drugs, including ethanol, produce (or contribute to) reinforcing effects by engaging a naturally occurring circuit that promotes contact and ingestion of substances. It is tempting to speculate that pharmacological interactions with α4-containing GABAARs may be the initial and specific means by which ethanol interacts with that circuit, although this remains to be determined.
