The main goal of the Project was to uncover how the medial septum modulates the hippocampal spatial code. Because of unexpected findings we aimed to pursue another objective: characterization of the context-dependent activity of hippocampal place cells. The experimental work was carried out during the Outgoing Phase which was extended by 6 months. During this period the grant was suspended. The collected data was analyzed during the return phase.
Regarding the experimental part of the Outgoing Phase, we implemented a location-contingent closed-loop optogenetic stimulation strategy: the medial septum was light-activated when the animal entered a pre-defined location monitored by distance sensors.
We injected 36 mice, of which 19 were implanted and 11 generated data. First, we observed a reduction in the animals’ running speed while traversing the stimulated location, paralleled by the transient disruption of theta oscillation. About 50% of place cells changed their spatial firing. The most common effect was the emergence of a new place field. The disappearance of existing place fields was also observed. Notably, in control sessions without delivering any light about 30% of place cells exhibited spontaneous remapping. Sham stimulation through an optic fiber terminating outside the brain did not cause alteration of spatial activity different than that observed in control, non-manipulated sessions. Thus, our results indicate that the medial septum can indeed, change the spatial code carried by the location-coupled activity of place cells.
As mentioned earlier the return phase was devoted to the analysis of the collected data. Characterization of control sessions unraveled an unexpectedly high prevalence of context-dependent activity of place cells in the stem part (central corridor) of the maze. Analysis of 182 place cells from 22 control sessions (unperturbed or stimulated without light responsive opsin) of 5 mice revealed that, 53.5%, an unexpectedly high proportion of place cells, exhibited multiple forms of context-correlated activity: i) one place field selective for either left or right trials; ii) two place fields (38.2% of all stem place cells) present in opposite trials; iii) two place fields had varying level of context-dependence; iv) non-splitter place cells with statistically non-distinguishable left and right activity. Strikingly, the proportion of splitter cells increased as the learning proceeded. Comparison of spatial activity in correct and error trials revealed, unexpectedly, a high correlation between trials with common origin (past choice i.e. where the mouse came from) but opposite upcoming choice (which side the mouse entered). This latter finding implies that splitter cells’ contextual code is retrospective rather than prospective.
I presented our findings on one local (restricted to the NYU Neuroscience Institute) and three international conferences and on a seminar organized in my home institute. I also gave a talk as part of a series of group reports in my home institute. Inspired by the approaches I learned during the Outgoing Phase, I had formulated a research program and successfully applied for a national grant. I was also awarded by a mentoring grant that enabled the training of a summer intern.
