Automated procedure to detect subtle motor alterations in the balance beam test in a mouse model of early Parkinson’s disease

Abstract

Parkinson’s disease (PD) is the most common motor neurodegenerative disorder, characterised by aggregated α-synuclein (α-syn) constituting Lewy bodies. We aimed to investigate temporal changes in motor impairments in a PD mouse model induced by overexpression of α-syn with the conventional manual analysis of the balance beam test and a novel approach using machine learning algorithms to automate behavioural analysis. We combined automated animal tracking using markerless pose estimation in DeepLabCut, with automated behavioural classification in Simple Behavior Analysis. Our automated procedure was able to detect subtle motor deficits in mouse performances in the balance beam test that the manual analysis approach could not assess. The automated model revealed time-course significant differences for the “walking” behaviour in the mean interval between each behavioural bout, the median event bout duration and the classifier probability of occurrence in male PD mice, even though no statistically significant loss of tyrosine hydroxylase in the nigrostriatal system was found in either sex. These findings are valuable for early detection of motor impairment in early PD animal models. We provide a user-friendly, step-by-step guide for automated assessment of mouse performances in the balance beam test, which aims to be replicable without any significant computational and programming knowledge.