Partial reinforcement across trials impairs escape performance but spares place learning in the water maze

We studied the effects of partial reinforcement on escape performance and place learning in the water maze. Rats given 50% reinforcement across trials (i.e. the escape platform was present only on odd trials) were compared to controls given 100% reinforcement (platform present on all trials). Control groups either received 8 or 4 trials per day, which was equal to either the total number of trials (100%-8) or reinforced escapes (100%-4) of the 50% group. Analysis of escape performance (latency) revealed that the 50% group was impaired relative to the 100%-8 group, but not the 100%-4 group, during the first 5 days of acquisition. The 50% group was impaired relative to both control groups on days 6-10 of overtraining. However, analyses of within-trial behavior (target annulus preference and thigmotaxis) on nonreinforced trials suggest that the 50% group did learn the location of the hidden platform (place information), in addition to a wall-based thigmotactic response. By dividing the 60s nonreinforced trials into three 20-s time bins, we were able to detect a significant preference for the target annulus early in the trial (bin 1 of trial 40 and bins 1-2 of trial 80). Further, there was a significant increase in time spent in the periphery of the pool, near the wall, in the last time bin of trial 40. Because the platform was in the middle zone, this behavior competed with a place response. We conclude that across-trial partial reinforcement procedures may promote response competition and mask evidence of place learning in addition to weakening escape performance late in training.