Patient-Specific Modeling of Daily Activity Patterns for Unsupervised Detection of Psychotic and Non-Psychotic Relapses

In this paper, we present our submission to the 2nd e-Prevention Grand Challenge hosted at ICASSP 2024. The objective posed in the challenge was to identify psychotic and non-psychotic relapses in patients using biosignals captured by wearable sensors. Our proposed solution is an unsupervised anomaly detection approach based on Transformers. We train individual models for each patient to predict the timestamps of biosignal measurements on non-relapse days, implicitly modeling normal daily routines. The models' mean-normalized prediction errors are then used as indicators of atypical behavior and, thus, risk of relapse. Our final submission ranked 3^rd on detecting non-psychotic relapses (Track 1) and 1^st on detecting psychotic relapses (Track 2).