Neural Signatures of Positive and Negative Emotional Facial Expressions in Mice
Yujia Chen¹, Ruiqing Hou¹, Zhinan Chen², Junli Lu¹, Si Chen³, Shisheng Xiong², Jianfeng Feng¹, Trevor Robbins¹˒⁴, Haitao Yan⁵, Xiao Xiao¹* ¹ Department of Anesthesiology, Huashan Hospital; Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence, Ministry of Education; Behavioral and Cognitive Neuroscience Center, Institute of Science and Technology for Brain-Inspired Intelligence, MOE Frontiers Center for Brain Science, Fudan University, Shanghai, China² School of Information Science and Technology Micro Nano System Center, Fudan University, Shanghai, China³ Department of Educational Psychology, Faculty of Education, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, China⁴ Department of Psychology and Behavioural and Clinical Neuroscience Institute, University of Cambridge, Cambridge, UK⁵ State Key Laboratory of Toxicology and Medical Countermeasures, Beijing Key Laboratory of Neuropsychopharmacology, Beijing Institute of Pharmacology and Toxicology, Beijing, China Email: xiaoxiao@fudan.edu.cn IntroductionUnderstanding how emotional states are encoded and expressed in animals is a central goal in affective neuroscience. while mice are widely used as models for studying emotion, the precise facial and neural signatures of emotional states in mice remain poorly defined.
MethodsIn this study, we investigated mouse facial expressions as reliable indicators of emotional responses elicited by both transient sensory stimuli and pharmacological interventions. By combining high-resolution facial imaging with advanced machine learning, we developed and validated data-driven prototypes of facial expressions. Our AI-driven model efficiently extracted emotion-related features, reducing the subjectivity associated with traditional behavioral assessments. In addition, dynamic keypoint tracking enabled the quantification of specific facial action units associated with emotional valence.ResultsOur AI-driven system revealed consistent valence and arousal dimensions in mouse facial expressions, paralleling emotion models in humans. Notably, ear-related movements were particularly informative of emotional valence. By integrating neural recordings from the ventral tegmental area (VTA), we uncovered cell-type-specific tuning of dopaminergic and GABAergic neurons to facial expressions of valence. Furthermore, optogenetic inhibition of these neurons selectively disrupted valence-specific facial patterns, pointing to a potential role in modulating facial expressions associated with emotion.DiscussionWe present a real-time, high-resolution platform for decoding emotional states in mice based on facial expression analysis. Our findings demonstrate that mouse facial expressions are robust indicators of affective states and are tightly linked to VTA neural dynamics. This work advances our understanding of the neural basis of emotion and provides a foundation for developing more targeted approaches to studying and modulating emotional processes in both animals and humans.
This work was supported by the National Key R&D Program of China (2021ZD0202805 and 2019YFA0709504);the National Natural Science Foundation of China (32471083);the Innovative Research Team of High-level Local Universities in Shanghai, 111 Project (B18015);Shanghai Municipal Science and Technology Major Project (2018SHZDZX01);Shanghai Center for Brain Science and Brain-Inspired Technology. N/A
