Answers are grounded in this paper and cite their sources.
Demonstrates the potential of wearable device data to predict migraine occurrence, suggesting improved management strategies.
Migraine is a chronic neurological disorder that significantly affects quality of life due to its recurrent and unpredictable nature. Early prediction of migraine attacks has the potential to improve disease management by enabling timely preventive interventions. With the increasing availability of wearable devices, continuous monitoring of daily physiological and behavioral patterns has become feasible in real-world settings. This study proposes a machine learning–based framework for predicting migraine occurrence using wearable-derived daily data. The dataset consists of longitudinal records collected from multiple individuals, capturing behavioral and lifestyle indicators such as sleep duration, stress level, hydration, mood, and screen time. Data preprocessing was performed to ensure temporal consistency and user-specific analysis, followed by feature engineering using short-term rolling statistics to capture recent trends preceding migraine attacks. Three classification approaches were evaluated: Logistic Regression, Extreme Gradient Boosting (XGBoost), and a stacked ensemble model combining linear and tree-based learners. Model performance was assessed using standard evaluation metrics, including accuracy, precision, recall, F1-score, and ROC-AUC. The ensemble-based model achieved slightly higher performance compared to the individual models, while Logistic Regression offers strong interpretability and XGBoost effectively captures non-linear relationships. The findings indicate that wearable-based daily indicators, when combined with appropriate feature engineering and machine learning models, can provide meaningful predictive signals for migraine occurrence. The results suggest that wearable-based data, combined with classical machine learning methods, may be useful for supporting migraine monitoring, in line with previous research on wearable technologies in healthcare [9].
No takes yet. Share an insight, caveat, or question.
Yazan Y. Hasan (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: