Below is a brief overview of shortcut experiment task and analysis work flow.
The main objective of this task was to determine whether hippocampal place cells during replay events represent novel behaviorally-relevant shortcut trajectories that the subject has not yet physically encountered. This was accomplished with the following analysis:
- Behavior-only analyses
- Place cell dynamics
- Sharp-wave ripple statistics
- Replay statistics
- Replay content via Bayesian decoding
Four male Long-Evans rats (Harlan; Mississauga, Canada), 5-12 months old, whose identifiers are R063, R066, R067, and R068.
Each recording session had the structure outlined in the above figure. Briefly, a recording session consisted of three main running blocks (Phase 1, Phase 2 and Phase 3), each bracketed by a rest period on a nearby pedestal (Pre-task, Pause A, Pause B and Post-task). The task phases started with Phase 1 in which only the familiar common core route was available. In Phase 2 the novel shortcut and dead-end trajectories were present but unavailable as their entrances were blocked by transparent barriers. Finally, Phase 3 allowed access to the full maze. In total, there were 8 recording sessions, each with a unique shortcut and dead-end path.
Each recording session contains a unique json file that has the meta data for that session, housed in the info directory. Meta data including commonly used variables is in meta.py. meta_session.py contains session loading and grouping info. Functions are set up with doit dependencies, such that the dependencies can be forgotten and downstream analyses will rerun.
For emi thesis: The thesis repository should be in the same folder as emi_shortcut (eg. code) and contain: thesis>images and thesis>generated.
- Start from a fresh
doit forgetstate (or forget specific tasks) - Figure generation:
doit copy_figures - Variable generation:
doit copy_tex