Sorry, you need to enable JavaScript to visit this website.

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

DOI:
10.60864/4466-k391
Citation Author(s):
Alice Hein, Sven Gronauer and Klaus Diepold
Submitted by:
Sven Gronauer
Last updated:
6 June 2024 - 10:27am
Document Type:
Presentation Slides
Document Year:
2024
Event:
Presenters:
Sven Gronauer
Paper Code:
GC-L2.2
 

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 indica- tors of atypical behavior and, thus, risk of relapse. Our final submission ranked 3rd on detecting non-psychotic relapses (Track 1) and 1st on detecting psychotic relapses (Track 2).

up
0 users have voted: