aalto1 untyped-item.component.html
Handler for synthetic data repeated analysis of EEG artifact-detection routines
Loading...
URL
Journal Title
Journal ISSN
Volume Title
Perustieteiden korkeakoulu |
Master's thesis
Electronic archive copy is available via Aalto Thesis Database.
Authors
Date
Department
Major/Subject
Mcode
SCI3102
Degree programme
Language
en
Pages
61+2
Series
Abstract
Existing tools for the preprocessing of EEG data provide a large choice of methods to suitably prepare and analyze a given dataset. Computational Testing for Automated Preprocessing (CTAP) provides batching and well-designed automation standardized preprocessing methods following Scientific Workflow Management System (SWMS) methodology to reduce human effort, subjectivity, and consequent error. Nonetheless, default parameters might not produce as good artifact removal performance as expected while preprocessing datasets with different features. Of course, users can modify parameters and test them to gain better performance, however, human effort, subjectivity, and time cost increase. This conflicts with our goal to reduce them. Moreover, the performance of the current pipe highly relies on how well the previous operations worked. Hence, in this thesis, to optimize the parameter used for the target function to reach a better performance on artifacts detection, CTAP introduces the Handler for Synthetic Data Repeated Analysis of EEG artifact-detection routines (HYDRA) method, a data-driven approach for EEG preprocessing optimum. The HYDRA generates synthetic EEG data as ground-truth, furnishes a fully automated method to evaluate the artifact detection performance under different parameters by numerical analysis. The thesis performed tests on several datasets by different branch pipe settings, which demonstrated EEG artifacts removal performance of the HYDRA method as well as the replicability of the results