Danelakis Antonios, Stubberud Anker, Tronvik Erling, Matharu Manjit
Department of Computer Science, NTNU Norwegian University of Science and Technology, 7030 Trondheim, Norway.
NorHead Norwegian Centre for Headache Research, 7030 Trondheim, Norway.
Life (Basel). 2025 Jun 4;15(6):909. doi: 10.3390/life15060909.
This narrative review introduces key concepts in artificial intelligence (AI), data science, and wearable devices aimed at headache clinicians and researchers. PubMed and IEEEXplore were searched to identify relevant studies, and these were reviewed systematically. We identified six primary research topics. First, the most common application of AI and data science is in the diagnosis of headache disorders, with reported accuracies of up to 90%. Second, AI and data science are used for predicting headache disease trajectories and forecasting attacks. Third, prediction of treatment effects and data-driven individualization of treatment prescription demonstrate promising results, with accuracies ranging from 40% to 83%. Fourth, AI and data science can uncover hidden information within headache datasets, offering clinicians deeper insights. Fifth, wearables, combined with AI and data science, can improve remote monitoring and migraine management. Lastly, user experience studies indicate strong interest from both clinicians and patients in adopting these technologies. The potential applications of AI, data science, and wearable device technologies in headache research are vast. However, many studies are small pilot studies, and models often suffer from poor performance, limited reporting, and lack of external validation, which impede generalizability and clinical implementation.
这篇叙述性综述向头痛领域的临床医生和研究人员介绍了人工智能(AI)、数据科学和可穿戴设备的关键概念。通过检索PubMed和IEEEXplore来识别相关研究,并对这些研究进行系统回顾。我们确定了六个主要研究主题。第一,人工智能和数据科学最常见的应用是头痛疾病的诊断,报告的准确率高达90%。第二,人工智能和数据科学用于预测头痛疾病轨迹和发作预测。第三,治疗效果预测和基于数据驱动的治疗方案个体化显示出有前景的结果,准确率在40%至83%之间。第四,人工智能和数据科学可以揭示头痛数据集中隐藏的信息,为临床医生提供更深入的见解。第五,可穿戴设备与人工智能和数据科学相结合,可以改善远程监测和偏头痛管理。最后一项,用户体验研究表明临床医生和患者对采用这些技术都有浓厚兴趣。人工智能、数据科学和可穿戴设备技术在头痛研究中的潜在应用非常广泛。然而,许多研究都是小型试点研究,模型往往存在性能不佳、报告有限以及缺乏外部验证等问题,这阻碍了其推广应用和临床实施。