Predicting Migraine Attack Onset with Real-Time Individualized Autonomic Patterns

ABSTRACT Current migraine management lacks objective tools for predicting imminent migraine attacks, which could enable preemptive treatment and improved patient outcomes. This project aims to establish predictive measures of migraine onset through continuous monitoring of autonomic activity using wrist-worn devices in individuals with episodic migraine. Approximately 30-40% of migraine sufferers experience prodromal symptoms, yet the mechanisms and temporal dynamics of these symptoms are poorly understood. Advances in wearable sensor technology provide an opportunity to monitor physiological parameters correlated with these symptoms continuously. We will exploit these advancements to test specific hypotheses through the following aims: Aim 1: Identify predictive markers of migraine onset by analyzing continuous physiological data (e.g., skin conductance, blood volume pulse, motion, skin temperature) from 60 patients using wrist-worn sensors. Aim 2: Differentiate objective versus subjective evaluative measures to enhance prediction accuracy of migraine onset. We hypothesize that autonomic changes will be detectable hours before an attack and that combined objective and subjective data will yield robust predictive markers. The methodology includes noninvasive, continuous data collection and transmission, complemented by self-reported data on prodromal symptoms and migraine events via validated e-diaries. Utilizing machine learning, we will develop algorithms to forecast migraine onset, enhancing the timing and efficacy of prophylactic treatments. This research leverages existing technology validated for clinical use and promises to elucidate the pathophysiology of migraine prodromes, offering transformative potential for migraine management.