Institutional Repository of Vilnius University
EEG duomenų kokybės ir apdorojimo strategijų poveikio motorinės vaizduotės klasifikavimo tikslumui įvertinimas /
Abstract
dc:descriptionDue to various injuries and diseases, the connection between the brain and muscles can be disrupted, leading to a loss of motor functions. To restore these functions, brain-computer interfaces, which translate recorded brain activity into useful actions, have been rapidly developing in recent years. These systems typically rely on electroencephalography (EEG) to record brain activity. This method has good temporal resolution but is prone to noisy signals. Typically, signal cleaning is performed manually, but for real-time applications, automated signal cleaning methods are necessary. The aim of this study was to evaluate the impact of EEG signal preprocessing and classifica-tion methods on model accuracy and transferability across sessions. The analysis was performed using the publicly available Lee2019 EEG dataset. A set of classification models was created to assess the effectiveness of different feature extraction and cleaning strategies. The most accurate result was achieved using features based on Riemannian geometry combined with logistic regression, reaching an accuracy of 73.5 %. The study revealed that channel selection and frequency filtering are the most effective signal cleaning methods, and additional cleaning does not improve classification accuracy or model transfer across sessions. Additionally, the use of event-related synchronization and desynchronization responses for distinguishing between literate and illiterate users is proposed, along with neurofeedback as a potential solution for addressing illiteracy.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bertašius, Rokas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD239028142&prefLang=en_US