Item

Abstract

In the IBL task, a head-fixed mouse is trained to move a wheel to the left or right, depending on the side of a Gabor patch shown on a screen in front of the animal. The probability of a stimulus being on one side is 80% (20%) in blocks of consecutive trials before switching in an unsignalled manner. The animals’ action selection (including on trials with 0% contrast input which are directly revealing of any bias) indicates that they are sensitive to this preponderance. We are currently examining neural correlates of this partly persistent sensitivity.
Here we explore aspects of the pre-stimulus behavior of the subjects to see if information about the currently favored side (crudely a ‘prior’) is present in such a way that might confound or at least complicate our neural analysis. We investigated kinematic correlates, such as body posture or paw position on the wheel, and dynamic correlates such as laterally-dependent whisking, tongue movements/licking, or motion of the wheel. To do this, we had to control for the substantial autocorrelation in the favored side (and the aspects of this prior to which the animals’ behavior shows to be sensitive).
We found that in a measurable fraction of animals, whisker movement between trials is modulated by block type, meaning that there is for example more whisking when a right stimulus is more likely to appear. Here we report the prevalence of such kinematic and dynamic correlates of the prior.