DeepLabCut™ is an efficient method for 3D markerless pose estimation based on transfer learning with deep neural networks that achieves excellent results across a broad collection of behaviours. This project will utilise the DeepLabCut package to analyse the behaviour of rats and mice as they are trained and tested on reward-based learning tasks designed to examine aspects of attention, memory and impulsive behaviour. These tasks are performed in touchscreen chambers that have on one side a monitor to display visual images in different arrangements and on the other side, a sugar pellet magazine for the dispensing of rewards. Importantly, these chambers are attached, via an automated sorting mechanism, to a home cage in which rats and mice live in social groups. They access the touchscreen tests in a self-directed manner, without experimenter intervention, so that their behaviour is as naturalistic as possible for a laboratory testing environment.
Video recordings of test sessions will be examined to determine patterns in behaviour that correlate with learning. These include how the speed of movement between the stimulus on screen and reward collection changes over time, side-preference patterns for turning, and types of movement in anticipation of an expected outcome (side-to-side head movements that scan for potential stimulus presentation location). Videos are recorded during different times of the day and night, therefore, information already collected about when learning is fastest will also inform the assessment of movement and behaviour. 

For more information contact the primary supervisor Dr Claire Foldi <claire.foldi@monash.edu>