Ballachey & Buel (1934) https://psycnet.apa.org/doi/10.1037/h0075474 reported 'centrifugal swing' in rat trajectories. Basically rats running round a corner often swing towards the outer edge like a racecar. Ballachey & Buel hypothesized this might influence subsequent turn choices. #animalcognition

They designed an elaborate #experiment to show and examine this effect; using 3 mazes (also their mirrored setups). Pressure plates before the choice points were connected to vacuum tube relays & a polygraph machine to record the rat's trajectory in detail.

They found that rats were 4x as likely to chose arms predicted by this centrifugal swing effect. Thus, the choices made by rats in a maze may not be independent, but can be strongly effected by previous #behaviour

Separating trials based on the rat's choice and looking at trajectories just before the choice point we can see a clear divergence in paths - choice could be predicted with 74% accuracy based on the paths immediately before it.

There may be a chicken<>egg problem with this last point: maybe the rats allow their path to swing out more when they are planning to turn in that direction, rather than the other way around? The general preference for the centrifugal choice suggests otherwise though.

In any case, it is curious I have never heard of this effect before. I have never seen it taken into account in models or statistical measures.

It demonstrates why we must control for path deviation when looking at trajectory-dependent activity/theta sequences/replay though.

@Sci_Hub link for this paper: https://sci-hub.se/10.1037/h0075474 because I don't think it is available anywhere else.