An automatic classification method of mouse self-grooming behavior based on machine learning
Journal Title: Journal of Air Force Medical University - Year 2023, Vol 44, Issue 11
Abstract
Objective To achieve the automatic analysis of mouse self-grooming behavior using open-source software based on a machine learning algorithm. Methods The self-grooming behavior of mice was recorded by video and analyzed manually by two observers. At the same time, samples of self-grooming behavior were selected from the video to train a machine learning classifier, and the classifier was used to automatically analyze the self-grooming behavior of mice. The results of manual and automatic analysis conducted by the classifier were compared. Results There was no difference in grooming duration between manual analysis and automatic classifier analysis. However, the classifier recorded a higher number of grooming bouts compared to the observers, with no difference between the observers. The mean grooming duration was shorter for the classifier compared to the observers, with no difference between the observers. In addition, there was no difference in the proportion of marks reported by the observers and the classifier. The Youdens index was also not different between the observers and the classifier. Conclusion The classifier trained and corrected with small samples can accurately identify the self-grooming behavior of mice, reduce the workload of manual labeling and improve the efficiency of behavior analysis.
Authors and Affiliations
QI Yuxin, ZHENG Yuanjing, TAN Zehua, YANG Zile, WANG Wenting
A cross-temporal meta-analysis on test results of SCL-90 in Chinese military personnel
Objective To explore the changes of mental health of Chinese military personnel over time and its relation with social changes. Methods A total of 390 original studies (n = 310 436) reported Chinese military personnels...
Morphological study of GABAergic neurons involved in pulpitis pain and related anxiety
Objective To establish a mouse model of experimental pulpitis, detect pain and anxiety-like behaviors, and observe the activation of neurons in insular cortex (IC) and medial prefrontal cortex (mPFC) and the activity ch...
Research progress on potential brain targets of physical therapy for post-traumatic stress disorder
Post-traumatic stress disorder ( PTSD) is a kind of mental illness that seriously threatens individual health and survival. A large number of studies have found that the abnormal state of specific brain regions is an im...
Development of an integrated measuring standard device for “ temperature-blood pressureoxygen saturation”
Objective To design and develop a portable, integrated and efficient measuring standard device of “ temperature-blood pressure-oxygen saturation” under the guidance of the thought of accompanying measurement guarantee,...
Construction of ubiquitination risk model of COVID-19 based on LASSO regression
Objective To explore the relationship between the activation of ubiquitin proteasome function and coronavirus disease 19 ( COVID-19) and construct a risk model, so as to provide potential biomarkers for the prognosis e...