Beyond the Signal: Medication State Effect on EEG-Based AI models for Parkinson’s Disease
Anna Kurbatskaya 1, Fredrik Nilsen Låder 1, Andreas Solvang Nese 1, Kolbjørn Brønnick 2,3, Alvaro Fernandez-Quilez 1 · 2025
Read Beyond the Signal: Medication State Effect on EEG-Based AI models for Parkinson’s Disease by Anna Kurbatskaya 1, Fredrik Nilsen Låder 1, Andreas Solvang Nese 1, Kolbjørn Brønnick 2,3, Alvaro Fernandez-Quilez 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-21992.txt.
Description
Anna Kurbatskaya and four co-authors tested in 2025 whether a Parkinson's patient's medication state skews AI models trained on brain wave data. Brain waves plus AI may track the disease and guide care, but drug state at recording time is rarely studied. The team used paired recordings from the same patients, taken on and off medication. Models were trained on each state and tested on both.
Models trained off medication scored modestly and evenly across states. Models trained on medication scored far higher on medicated tests, and still well on unmedicated ones. In short, training state shaped what the models learned more than test state did. The authors say drug state must be handled openly before such tools enter clinics.
Catalog record
| CREATOR | Anna Kurbatskaya 1, Fredrik Nilsen Låder 1, Andreas Solvang Nese 1, Kolbjørn Brønnick 2,3, Alvaro Fernandez-Quilez 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-21992