Abstract: A tail suspension analysis system has been established based on DeepLabCut (DLC) algorithm for evaluating depressive behavior in mice.Based on the DLC algorithm in deep learning technology,multi-path,semi-closed tail suspension hardware,the systems of software and hardware were constructed;this system and open field analysis system were applied to evaluate the tail suspension and open field behavioral changes in mice with reserpine depression model,intermittent alcohol consumption injury model,and periodontal silk ligation model.This system realized the synchronous analysis of head movement and limb movement in tail suspension,and automatically quantified 50 tail suspension indexes of dynamic and static type,posture type,and intensity type.It was found that in the reserpine depression model,compared with the control group,the movement distance,wall climbing times,and head swing times were reduced in the low-,medium-,and high-dose reserpine groups in 2 days,and the head swing times were reduced in the medium- and high-dose reserpine groups in 3 days and 4 days.In the intermittent alcohol consumption injury model,compared with the control group,the immobility time was reduced,the limb struggling times,movement distance,and movement time were increased in the low- and medium-dose ethanol groups,and the forelimb,hindlimb,and absolute struggling times were reduced in the high-dose ethanol group.In the periodontal silk ligation model,the movement distance and length,limb and absolute struggling times,and the total value of forelimb,hindlimb,and limb angles were reduced in the ligation group.Clearly,the DLC algorithm-based mouse tail suspension analysis system can accurately distinguish the behavioral characteristics of different mechanistic depression models.

Key words: tail suspension analysis system, artificial intelligence, DeepLabCut, depression-like behavior