抑郁症肝气郁证患者抑郁眼神、面部表情和姿态多模态识别及其脑区功能定位

The objective measurement of depression of depressive disorder is major problem for many years studies have not been solved by the medical profession at home and abroad.External expressions changes reflect the inherent physiological and pathological , is a basic understanding of emotional theory of traditional Chinese medicine, but no rigorous research.Our previous developed depressive eye single modal identification and key index of exploration research, find the absolute value on the basis of pupil inside canthus distance skewness canpreliminary identification of depression.This study on the basis of this, around the scientific issues of depressive disorder liver qi stagnation syndrome in patients depressive eye and facial expression,gesture multimodal recognition and brain functional location,General methods such as affective computing,pattern recognition and machine learning ,fMRI. Carry out the expert judge study ,Research the multimodal depressive recognition and relationship its central function in the brain.carry out the modern medical science, the psychology of emotion, exploratory research and the foundation of traditional Chinese medicine and clinical are not yet developed, expectation is expected for depression emotion of objective evaluation and TCM liver Shuxie emotion central mechanism was revealed to contribute, and enrich the development of TCM psychology theory, among the modern medical science, the psychology of emotion in the frontier.This study has become in the spirit of affective disorder in our country is serious public health and social harmony when the problem, has important scientific significance and social significance.

抑郁症抑郁情绪的客观测评是国内外医学界多年研究尚未解决的难题。外在表情变化反映内心情志体验及其病理生理，是中医情志学说的基本认识，但未见严谨研究。我们前期开展了抑郁眼神单模态识别及其关键指标探索研究，发现可以依据瞳孔内眼角距的绝对值偏度初步识别抑郁情绪，本研究拟在此基础上，围绕抑郁症肝气郁证患者抑郁眼神及面部表情、姿态多模态识别及其脑功能定位这一科学问题，综合情感计算、图像识别及机器学习、fMRI等方法，开展该病证患者抑郁眼神及面部表情、姿态专家判别，多模态识别及其与脑中枢功能定位研究。开展这一现代医学、情绪心理学，以及中医基础和临床均尚未开展的探索性研究，预期有望为抑郁情志的客观测评及中医肝主疏泄调畅情志中枢机制揭示做出贡献，丰富发展中医情志理论，跻身现代医学、情绪心理学在该研究领域前沿。该研究在精神情感障碍已成为我国严重公共卫生和社会和谐问题之时，具有重要科学意义和社会意义。

抑郁症是一种常见的精神疾病，严重时出现自杀观念和行为，WHO有关报道表明其将成为 21 世纪人类的主要杀手，但全球只有不足一半的患者（在许多国家中仅有不到 10%的患者）接受有效治疗。其中不能做出正确诊断是影响有效治疗的重要因素。而抑郁情志的客观测评在抑郁症的诊断和疗效评价中具有重要意义。本项目围绕这一关键科学问题，综合情感计算、图像识别及机器学习、fMRI等方法，开展抑郁症肝气郁证患者抑郁眼神及面部表情、姿态多模态识别及其脑中枢功能定位研究，研究结果揭示了抑郁症肝气郁证患者眼神特征为：平滑追踪眼动的均方根误差大、眨眼频率高，闭环追踪阶段和总体的眼跳频率低、增益值高，其追踪的平滑性差；固视眼动的均方根误差和眼跳幅度大、眨眼频率高、注视时间比和注视频率低、其注视的稳定性差；反向眼跳正确率差，平均速度位置主效应显著；眼睛开合度小、眉毛运动幅度小、眨眼频率高、眼球运动幅度小、眼睛颜色有差异。建立了其眼神、面部表情及姿态识别的支持向量机(Support Vector Machine，SVM）和梯度提升树（Gradient Boosting Decison Tree，GBDT）模型，其识别正确率分别为88.68%、84.06%、73.08%。识别出的眼神、面部表情、姿态的脑区功能定位研究结果表明枕叶距状裂上下周围的皮质、梭状回激活增强；额极区、眶额区、额下回前下部皮层、杏仁核、海马、左侧前扣带回、右侧前扣带回激活减弱。本项目为抑郁情志的客观测评提供新方法，同时结合前期研究部分揭示中医肝主疏泄调畅情志中枢机制，丰富发展中医情志理论，跻身现代医学、情绪心理学在该研究领域前沿，具有重要科学意义和社会意义。依据上述研究，发表论文 7 篇，其中中文核心 2 篇，SCIE2 篇，EI3 篇。获得软件著作权 2 项。培养研究生 6 名，其中博士研究生 2 名，硕士研究生 4 名。