neural trajectories over time for each trial. During alcohol sessions, alcohol-associated cues elicited neural activity patterns that diverged before the availability of alcohol when drinking versus non-drinking trials were compared in Wistar (FDR-corrected rank-sum tests; p < 0.05; Fig. 5B), but not P rats (FDR-corrected rank-sum tests; p < 0.05; Fig. 5C). In other words, the temporal evolution of neural activity patterns in Wistar rats in response to alcohol-associated cues were predictive of future drinking/non-drinking trials, whereas the neural activity patterns in P rats were not. Additionally, during water sessions, population activity only briefly differentiated drinking trials from non-drinking trials in Wistar (FDR-corrected rank-sum tests; p < 0.05; Fig. 5D), and failed entirely in P rats (FDR-corrected rank-sum tests; p < 0.05; Fig. 5E). In contrast, there were large differences between P and Wistar in cue/task-elicited population activity. Specifically, on water drinking trials, P rats displayed greater alterations in neural activity patterns versus Wistar rats (FDR-corrected rank-sum tests; p < 0.05; Fig. 6A). Thus, in P rats, the mPFC was biased toward encoding alcohol drinking during alcohol consumption, whereas in Wistar rats, encoding of the intention to drink alcohol and alcohol drinking was present. Therefore, converging evidence suggests that the
