Investigating the Impact of Weather on Headache Frequency Using Big Data and Deep Learning Techniques: Insights from the Zutool Smartphone Application

We investigated the relationship between weather and headache frequency using data from the “Zutool” smartphone app, which logs headaches and weather conditions. We employed deep learning (DL) techniques alongside statistical analysis to explore this association. Data from 4375 users, totaling 336,951 hourly headache events and weather records from December 2020 to November 2021, were analyzed. Statistical and DL models were developed to predict headache occurrences based on weather factors. Clustering was used to assess weather sensitivity. Lower barometric pressure, higher humidity, increased rainfall, significant barometric pressure drops 6 h prior, higher morning barometric pressure, lower barometric pressure the following day, and specific barometric pressure patterns over 6 days (types I and II) were significantly linked to headache occurrences in both models. Users were clustered into seven groups based on weather sensitivity. We found that low barometric pressure, barometric pressure changes, higher humidity, and rainfall correlated with increased headache frequency. These findings support personalized medicine and neurorehabilitation efforts.