【目的】通过数据挖掘技术探讨针灸治疗帕金森睡眠障碍的选穴规律。【方法】计算机检索中国知网期刊全文数据库 （CNKI）、万方学术期刊全文数据库（Wanfang）、维普中文科技期刊数据库（VIP）、中国生物医学文献服务系统（SinoMed）、美 国生物医学信息检索系统（PubMed）、科学网（web of science）以及荷兰医学文摘（Embase）等各大数据库。检索时间为各数据 库自建库以来至2024 年12 月10 日收录的有关针灸治疗帕金森睡眠障碍的临床研究类文献。利用Excel 2021 构建数据库，采 用SPSS 27.0 软件对高频腧穴进行系统聚类分析，通过SPSS Modeler 18.0 软件结合Apriori 算法对腧穴进行关联规则分析。 【结果】共纳入26 篇文献，提取26 条针灸处方，涉及56 个穴位，总使用频率为200 次。针灸治疗帕金森睡眠障碍中使用频 率最高的前3 位腧穴是百会、三阴交、神门穴，总使用频次排前3 位的经脉分别是督脉、非经脉、足少阳胆经，腧穴主要的 分布部位依次为头颈部、下肢部和上肢部。所选处方涉及38 个特定穴，总使用频率达141 次，占所有穴位使用频率的70.5% （141/200），使用频率前3 位依次为交会穴、非特定穴、原穴和五输穴（并列第3 位）。高频腧穴（使用频次≥ 4 次）的系统聚类 分析表明，当冰柱图的群集数尺度为4、树状图的标度为10 时，得到6 个有效聚类群组，分别为“太冲-合谷-百会-四神聪- 三阴交-神门-印堂”“照海-申脉”“安眠”“神庭”“风池”“内关”。高频腧穴的关联规则分析显示，在设定支持度阈值为 25%、置信度阈值为80%的参数条件下，共研究挖掘出26 组具有显著关联性的穴位配伍组合，支持度最高的穴组为“百会- 三阴交”。【结论】针灸治疗帕金森睡眠障碍选穴以督脉为主，主要分布在头颈和四肢部，注重特定穴的使用，重视“上病下 取，下病上取”“远近配穴”“上下配穴”原则，重视从头论治。

Objective To investigate the acupoint selection rules in acupuncture for Parkinson’s disease sleep disorders using data mining technology. Methods Clinical studies on acupuncture for Parkinson’s disease sleep disorders were retrieved from major databases including CNKI，Wanfang，VIP，SinoMed，PubMed，Web of Science，and Embase，from database inception to December 10，2024. Excel 2021 was used to construct a database. SPSS 27.0 software was employed for systematic cluster analysis of high-frequency acupoints，and SPSS Modeler 18.0 software combined with the Apriori algorithm was used for association rule analysis of acupoints. Results A total of 26 articles were included，from which 26 acupuncture prescriptions were extracted，involving 56 acupoints with a total application frequency of 200 times. The top three acupoints with the highest frequency in acupuncture for Parkinson’s disease sleep disorders were Baihui（GV20），Sanyinjiao（SP6），and Shenmen （HT7）. The top three meridians in terms of total application frequency were the governor vessel，non-meridian points，and the gallbladder meridian of foot-shaoyang. The acupoints were primarily distributed in the head and neck region，followed by the lower limbs and upper limbs. The selected prescriptions involved 38 specific acupoints，with a total application frequency of 141 times，accounting for 70.5%（141/200）of the total acupoint application frequency. The top three categories in terms of frequency were crossing points，non-specific points， and Yuan-Source points and Five-Transport points（tied for third）. Systematic cluster analysis of high-frequency acupoints（frequency ≥ 4）indicated that when the cluster number scale of the icicle plot was 4 and the scale of the dendrogram was 10，six effective cluster groups were formed，namely：“Taichong（LR3）-Hegu（LI4）-Baihui Sishencong（EX-HN1）-Sanyinjiao-Shenmen-Yintang（GV29）”，“Zhaohai（KI6）-Shenmai（BL62）”，“Anmian （EX-HN22）”，“Shenting（GV24）”，“Fengchi（GB20）”，and“Neiguan（PC6）”. Association rule analysis of high-frequency acupoints revealed that，under the parameter conditions with a support threshold set at 25% and a confidence threshold at 80%，a total of 26 significantly associated acupoint combinations were identified. The acupoint pair with the highest support was“Baihui-Sanyinjiao”. Conclusion Acupuncture for Parkinson’s disease sleep disorders primarily selects acupoints from the governor vessel，mainly distributed in the head，neck，and limbs. Emphasis is placed on the use of specific points，adhering to the principles of“treating upper disorders by managing lower body parts and treating lower disorders by managing upper body parts”，“combining distal and proximal points”，and“combining upper and lower points”，with considerable attention given to treatment strategies targeting the head.

R246.9

北京市重大疑难疾病中西医协同攻关项目（编号：京中医政函[2023]55 号）；北京市医院科研培育计划资助项目（编号：PZ2023014）