Automatic nuclei detection in DBS microelectrode recordings
The deep brain stimulation (DBS) is a modern technique for treatment of late-stage movement disorders, consisting of permanent electrical stimulation of deep brain structures. In order to achieve good clinical outcome with low side effects, accurate positioning of the stimulating contacts in the target structure (e.g. subthalamic nucleus in Parkinson's disease) is necessary. The most commonly used method of accurate electrode placement consists of 1) preoperative MRI and CT imaging, 2) intra-operative micro-EEG recording around the presumed target position using a set of micro-electrodes 3) post-operative verification of electrode position using MRI and/or CT.
Analysis and modelling of micro-EEG signals from areas around the STN nucleus - which is the core part of this project - may help in accurate electrode positioning, as well as understanding of correspondences between recorded microEEG and pre/post operative MRI.
Bakstein, E.; Sieger, T.; Novák, D.; Jech, R. (2016) Probabilistic model of neuronal background activity in deep brain stimulation trajectories In: Information Technology in Bio- and Medical Informatics. Basel: Springer, 2016, pp. 97-111. LNCS 9832 [PDF preprint]
Bakštein, E., Sieger, T., Novák, D., Růžička, F., & Jech, R. (2018). Automated Atlas Fitting for Deep Brain Stimulation Surgery Based on Microelectrode Neuronal Recordings. In Proceedings of the World Congress on Medical Physics and Biomedical Engineering 2018 (pp. 105–111). [PDF preprint]
Automatic artifact detection in microelectrode recordings
Micro-EEG recordings are very susceptible to motion-induced and other types of technical artifacts. As the mEEG signals are often used not only in DBS targetting and nuclei identification process but also in all sorts of neuroscience experiments and unit activity evaluation, it is crucial to identify artifact-free segments.
This is why we developed the sigInspect: a graphical user interface (GUI) application for Matlab, developed for inspection and annotation of extracellular microelectrode recordings (MER), which allows also automatic artifact detection, using the algorighms presented below.
Bakštein, E., Sieger, T., Wild, J., Novák, D., Schneider, J., Vostatek, P., Urgošík, D., Jech, R. (2017). Methods for automatic detection of artifacts in microelectrode recordings. In: Journal of Neuroscience Methods, 290, 39–51. [PDF, supplementary: PDF, matlab codes, data (18MB)]
Automatic relapse prevention in psychiatric patients
In diagnoses such as schizophrenia or bipolar disorder (BD), the treatment objective is to keep patients in "remission" - a safe state with low presence of disease symptoms. Our goal is to develop a n automatic e-health system for automatic relapse prediction in schizophrenia and BD, based o remote monitoring using behavioral data (wrist-worn actigraphy) and self-reports (mobile app questionnaires).
Spaniel, F., Bakstein, E., Anyz, J., Hlinka, J., Sieger, T., Hrdlicka, J., Gornerova, N., Hoschl, C. (2016) Relapse in schizophrenia: definitively not a bolt from the blue. Neuroscience Letters, S0304-3940(16), 30265–8. [PDF preprint at researchgate]