A1071
Title: A user-friendly EEGLAB plug-in integrating functional data analysis to advance EEG research
Authors: Mohammad Fayaz - Shahed University (Iran) [presenting]
Abstract: EEG datasets pose complex challenges requiring advanced statistical methods. Functional data analysis (FDA) and machine learning offer powerful frameworks to model EEG signals over continuous domains. EEGLAB is a leading open-source EEG platform, widely cited in Web of Science and SCOPUS, and used globally, confirmed by bibliometric analyses. Despite FDAs potential, programming complexity limits its use among many researchers. The EEGLAB-FDA plug-in addresses this by integrating FDA methods into EEGLAB with a user-friendly graphical interface, enabling computational neuroscientists and statisticians to apply advanced FDA tools without coding. It includes modules for functional principal component analysis (FPCA), functional canonical correlation analysis (FCCA), and event-related potential (ERP) analysis, including smoothing, derivative estimation, and phase-plane visualization. Supporting dense and sparse data, recent updates add functional regression and enhanced sparse data handling. Bug fixes and interface improvements improve usability. Applications with visual and auditory oddball tasks show the plug-in's capacity to reveal insights, especially when analyzing the P300 component. Available free on GitHub and the EEGLAB plug-in list, it is supported by a YouTube channel with tutorials. Ongoing development aims to expand its features and impact in computational neuroscience and statistics.
