基于脑功能连接探讨机器学习在针刺治疗失眠效果预测中的应用

Exploring the application of machine learning in predicting the effect of acupuncture for insomnia based on brain functional connectivity

针灸是治疗失眠的有效途径之一,但因基于临床经验,治疗方案尚待优化,且效果以主观评测为主,个体效果差异明显。因此,筛选有效人群的个体识别标记,预测疗效,优化针刺治疗方案十分重要。机器学习技术(ML)的分类和预测算法正迅速成为疾病诊断、预测及分类工具。而磁共振成像技术已揭示失眠个体的脑功能连接(FC)具有高度特异性,在针刺前后有特定的变化规律,可作为追踪治疗反应的标志物。本文拟基于FC技术,探析运用ML中的分类算法模型预测针刺治疗失眠疗效的可行性,为突破“经验治疗”瓶颈、优化针刺治疗方案提供借鉴。

Acupuncture is one of the effective ways to treat insomnia,but because it is based on clinical experience,the treatment protocol has yet to be optimized,and the efficacy is mainly subjective assessment,with significant differences in individual efficacy.Therefore,it is important to screen effective populations for individual identification markers to predict efficacy and optimize acupuncture treatment protocols.Machine learning(ML)for classification and prediction algorithms are rapidly becoming tools for disease diagnosis,prediction,and classification.In contrast,magnetic resonance imaging has revealed that the brain functional connectivity(FC)of insomniac individuals is highly specific,with specific patterns of change before and after acupuncture,and can be used as a marker to track treatment response.This paper proposes to explore the feasibility of using the classification algorithm model in ML to predict the efficacy of acupuncture treatment for insomnia based on FC technology,in order to break the bottleneck of“empirical treatment”and optimize the acupuncture treatment plan.