Spatial working memory over long retention intervals: dependence on sustained cholinergic activation in the septohippocampal or nucleus basalis magnocellularis-cortical pathways?

Previous direct neurochemical studies of the temporal dynamics of cholinergic activation in the septohippocampal and nucleus basalis magnocellularis-cortical pathways at various stages during repeated testing of mice with selective spatial reference or working memory protocols [Durkin and Toumane (1992), Behav. Brain Res. 50, 43-52] showed that the post-test durations of cholinergic activation in each pathway varied as a function of the type of memory tested and the level of task mastery. Since (i) the hippocampal formation is considered to constitute a critical component of a temporary memory buffer, and (ii) working memory items are not thought to be submitted to consolidation and permanent storage, we postulated that the duration of testing-induced cholinergic activation in the septohippocampal pathway may govern the maintenance of the working memory trace over the retention interval. In order to test directly this hypothesis C57 B1/6 mice were extensively trained (one trial/day, 25-30 days) on an identical selective working memory task to attain high levels of retention (> 80% correct), but using either 5 min (Group 1), or 60 min (Group 2) retention intervals. At various times (30 s-75 min) following the initial acquisition phase of the test, cholinergic activity in the hippocampus and frontal cortex was quantified using measures of high-affinity choline uptake. Whereas cholinergic activation was observed in both pathways at 30 s post-acquisition and throughout the 5 min retention interval in Group 1, the situation in Group 2 is different, activation of the septohippocampal pathway being maintained for only 15 min, while activation in the nucleus basalis magnocellularis-cortical pathway is maintained for the totality of the 1 h retention interval. The nucleus basalis magnocellularis-cortical cholinergic pathway, in addition to its role in long-term reference memory storage processes may, thus, via an intervention in the temporal encoding of information, also subsume a complementary intermediate-term buffer storage role in working memory situations requiring retention intervals in excess of 15 min in mice. This secondary, "backup", function of the nucleus basalis magnocellularis-cortical pathway would thus liberate the septohippocampal complex from its primary active role in the temporary maintenance and/or accessibility of the working memory trace in these particular cases requiring long retention intervals.