Figure 2: Basolateral to central amygdala shift in the control over aDLS dopamine-dependent cocaine seeking habits.

(a) Active lever presses during early stages cocaine seeking were decreased by B/M infusions into the BLA, but not the CeN (main effect of lever: F_1,21=138.55, P<0.001, partial η²=0.86; dose × structure interaction: F_1,21=7.54, P<0.013, partial η²=0.26 and dose × lever × structure interaction: F_1,21=7.01, P<0.015, partial η²=0.25). When cocaine-seeking behaviour was well-established, bilateral infusions of α-flupenthixol (a) into the aDLS decreased active lever presses both in the BLA (b) and CeN (c) groups (main effect of lever: F_1,21=173.82, P<0.001, partial η²=0.89; dose: F_1,21=34.19, P<0.001, partial η²=0.61; dose × lever interaction: F_1,21=9.07, P<0.01, partial η²=0.30 but no dose × lever × structure interaction: F_1,21<1). However, active lever presses were decreased both by bilateral inactivation of the CeN and functional disconnection of the CeN and aDLS (main effect of lever: F_1,10=148.49, P<0.001, partial η²=0.93; manipulation: F_4,40=3.32, P<0.02, partial η²=0.25 and lever × manipulation interaction: F_4,40=3.045, P<0.05, partial η²=0.23) to the same extent as by bilateral aDLS α-flupenthixol infusions (Fs_1,10<1) (c). In marked contrast, bilateral inactivation of the BLA or functional disconnections of the BLA and the aDLS had no influence on active lever presses when cocaine seeking was established as a habit (main effect of lever: F_1,11=357.80, P<0.001, partial η²=0.97; manipulation: F_4,44=1.17, P>0.16, partial η²=0.13 and lever × manipulation interaction: F_4,44=1.67, P>17, partial η²=0.13). *: different from vehicle, P<0.05; **: different from vehicle, P<0.01, Newman–Keuls post hoc test. Pink fills represent intra-amygdala B/M infusion. Blue fills represent intra-aDLS infusions of α-flupenthixol. B/M, baclofen/muscimol mixture. Inactive lever presses are represented by black shaded fills.