JMIR Publications

ABSTRACT

Background:

The use of wearables facilitates data collection at a previously unobtainable scale, enabling the construction of complex predictive models with the potential to improve health. However, the highly personal nature of this data requires strong privacy protections against data breaches and against using data in a way the users did not intend. One method to protect user privacy while taking advantage of sharing data across users is federated learning, a technique which allows a machine learning model to be trained using data from all users, while only storing a user’s data on that user’s device. By keeping data on users’ devices, federated learning protects the user’s private data from data leaks and breaches on the researcher’s central server, and provides users with more control over how and when their data is used. However, there are few existing rigorous studies of the effectiveness of federated learning in the mobile health domain.

Objective:

We review federated learning and assess whether it can be useful in the mobile health field, especially for addressing common mobile health challenges such as privacy concerns and user heterogeneity. Our objectives were to (1) describe federated learning in a mobile health context, (2) apply a simulation of federated learning to a mobile health dataset, and (3) compare the performance of federated learning with other predictive models.

Methods:

We applied a simulation of federated learning to predict the affective state of 15 subjects using physiological and motion data collected from a chest-worn device for about 36 minutes. We compared the results from this federated model to those from a centralized (server) model and to results from training individual models for each subject.

Results:

In a three-class classification problem using physiological and motion data to predict whether the subject was undertaking a neutral, amusing, or stressful task, the federated model achieved 92.8% accuracy on average, the server model achieved 93.2% accuracy on average, and the individual model achieved 90.2% accuracy on average.

Conclusions:

Our findings support the potential for using federated learning in mobile health. The results showed that the federated model performed better than a model trained separately on each individual, and nearly as well as the server model. Since federated learning offers more privacy than a server model, it may be a valuable option for designing sensitive data collection.