Abstract
It has been established that aged rats are worse at learning the spatial version of the Morris watermaze task compared to their younger counterparts. It remains unclear, however, whether this poorer performance by the older rats can be attributed to the use of different behavioral strategies to solve the task. We trained young (6-9 months, n = 37) and old (19-22 months, n = 65) male Fischer 344 rats on the Morris watermaze for four consecutive days with six trials per day. Using Rtrack, an automated discriminant classifier, we analyzed each rat's swimming trajectory to estimate the probability of the rat using one of eight distinct navigation strategies on each given trial. Across acquisition, the behavioral strategy profiles diverged markedly by age. The use of both platform-independent and procedural strategy types declined across learning in young rats in favor of allocentric ones. In contrast, there was no decline in the old rats' use of procedural strategies across learning, although their use of allocentric strategies did increase across days. The allocentric strategies that older animals favored tended to be less accurate than those favored by young rats. To investigate the tendency to switch strategies between trials, the entropy of strategy transition matrices was calculated. Young rats exhibited lower entropy between trials, which reflects the fact that they tended to converge onto a restricted set of allocentric strategies by Day 4. Old rats, on the other hand, had higher entropy, reflecting the fact that they continued to interleave procedural and allocentric strategies throughout training. These results align with human literature that hypothesizes that aging shifts the balance of strategy selection from allocentric hippocampus-dependent circuits to procedural extra-hippocampal circuits.