Automated Analysis of Stroke Mouse Trajectory Data With Traja

Abstract

Quantitative characterization of mouse activity, locomotion and walking patterns requires the monitoring of position and activity over long periods of time. Manual behavioral phenotyping, however, is time and skill-intensive, vulnerable to researcher bias and often stressful for the animals. We present examples for using a platform-independent open source trajectory analysis software, Traja, for semi-automated analysis of high throughput mouse home-cage data for neurobehavioral research. Our software quantifies numerous parameters of movement including traveled distance, velocity, turnings, and laterality which are demonstrated for application to neurobehavioral analysis. In this study, the open source software for trajectory analysis Traja is applied to movement and walking pattern observations of transient stroke induced female C57BL/6 mice (30 min middle cerebral artery occlusion) on an acute multinutrient diet intervention (Fortasyn). After stroke induction mice were single housed in Digital Ventilated Cages [DVC, GM500, Tecniplast S.p.A., Buguggiate (VA), Italy] and activity was recorded 24/7, every 250 ms using a DVC board. Significant changes in activity, velocity, and distance walked are computed with Traja. Traja identified increased walked distance and velocity in Control and Fortasyn animals over time. No diet effect was found in preference of turning direction (laterality) and distance traveled. As open source software for trajectory analysis, Traja supports independent development and validation of numerical methods and provides a useful tool for computational analysis of 24/7 mouse locomotion in home-cage environment for application in behavioral research or movement disorders.

Keywords: animal tracking; home-cage; machine learning; mouse; neuropsychiatric disorders; stroke.