- BGOOSE stands for Basal Ganglia Oscillation-based mOdel for Sleep-stage Estimation. [1]
- BGOOSE 🦢 contains a sleep decoding model with an input of local field potential signals recorded from basal ganglia structures (i.e., subthalamic nucleus and globus pallidus internus), and an output of three-class sleep stage labels (i.e., NREM, REM and wakefulness).
- BGOOSE 🦢 is trained on the largest to date ⭐ synchronized basal ganglia LFP - PSG dataset in a cohort of 141 patients with movement disorders including Parkinson’s disease, Essential Tremor, Dystonia, Essential Tremor, Huntington’s disease and Tourette’s syndrome.
- Within-cohortly, the generalized BGOOSE 🦢 model achieved over 80% average accuracy across patients and across disease conditions, even in the presence of recordings from different basal ganglia targets. 😮
- In out-of-cohort validations, BGOOSE 🦢 still achieves around 75% sleep staging accuracy 🎯 on two external datasets (the Tsinghua dataset [2]) and the UCSF dataset [3]) recorded at different postoperative time points and using different DBS devices.
- Sorry that due to the limited time I currently have 😞, it’s not possible for me to organize BGOOSE into a neat python package. Having that said, the main function of the BGOOSE can still be utilized by downloading the pretrained BGOOSE 🦢 model (you can find it at https://osf.io/mt72e/). A demo data can also be downloaded at the same website.
- Please check the jupyter notebook (BGOOSE - demo.ipynb) for a demo usage of BGOOSE. For any question or suggestion please contact me directly at [email protected] 😃
- [1] Yin, Z. X. et al. Generalized sleep decoding with basal ganglia signals in multiple movement disorders. Not yet published (2024)
- [2] Chen, Y. et al. Automatic Sleep Stage Classification Based on Subthalamic Local Field Potentials. IEEE Trans Neural Syst Rehabil Eng 27, 118–128 (2019).
- [3] Anjum, M. F. et al. Multi-night naturalistic cortico-basal recordings reveal mechanisms of NREM slow wave suppression and spontaneous awakenings in Parkinson’s disease. bioRxiv (2023)