Behavioral Training of Mice to Recognize Violations of Temporal Visual Sequences

Abstract

The recognition of temporal patterns is an essential part of visual processing, and can be used to predict upcoming visual stimuli. Predictive computation is often associated with higher-order brain areas, but it has also been found in areas as early as the retina and primary visual cortex (V1). Studies in mice have shown that when the pattern of a temporal sequence is interrupted, there is a large novelty response by the neurons in V1, however it is unknown if or how the rest of the brain utilizes this response. In order to investigate the implications of the V1 neural novelty response, this study sought to train mice in a task to behaviorally respond to violations in temporal visual sequences. It was found that mice are able to learn the task, and perform it very well. Parameters such as the number of frames in the sequence and duration time of those frames affected the difficulty of the task and the performance of the mouse. Generally mice performed best with shorter sequences that were displayed quickly (~200ms). These findings emphasize the importance of the primary visual cortex in the predictive computation of temporal sequences, and indicate that studying lower-order brain regions may aid in the understanding of higher-order cognitive functions.