Abstract
AbstractBackgroundThe nervous system functions of an animal are predominantly reflected in the behaviour and the movement, therefore the movement-related data and measuring behavior quantitatively are crucial for behavioural analyses. The animal movement is traditionally recorded, and human observers follow the animal behaviours; if they recognize a certain behaviour pattern, they will note it manually, which may suffer from observer fatigue or drift.ObjectiveAutomating behavioural observations with computer-vision algorithms are becoming essential equipment to the brain function characterization in neuroscience trials. In this study, the proposed tracking module is eligible to measure the locomotor behaviour (such as speed, distance, turning) over longer time periods that the operator is unable to precisely evaluate. For this aim, a novel animal cage is designed and implemented to track the animal movement. The frames received from the camera are analyzed by the 2D bior 3.7 Wavelet transform and SURF feature points.ResultsImplemented video tracking device can report the location, duration, speed, frequency and latency of each behavior of an animal. Validation tests were conducted on the auditory stimulation trial and the magnetic stimulation treatment of hemi-Parkinsonian rats.Conclusion/ SignificanceThe proposed toolkit can provide qualitative and quantitative data on animal behaviour in an automated fashion, and precisely summarize an animal’s movement at an arbitrary time and allows operators to analyse movement patterns without requiring to check full records for every experiment.
Publisher
Cold Spring Harbor Laboratory