European Commission logo
English English
CORDIS - EU research results
Content archived on 2024-05-29

On-line adaptive classification for multi-class EEG-based brain-computer interfaces

Final Activity Report Summary - MULTI-ADAPTIVE BCI (On-line adaptive classification for multiclass EEG-based Brain-Computer Interfaces)

A Brain Computer Interface (BCI) is a system that translates human intentions into a technical control signal bypassing the classical neuromuscular output pathways. In this project electroencephalographic brain signals (EEG) were used. The EEG is known to be non-stationary and therefore, EEG based-BCI systems can benefit from adaptation of its components. Furthermore, acquiring knowledge of the changes in brain signals allows the design of tools invariant to these and make more robust BCI systems.

This project has pursued the adaptation of a BCI system with the goal of training inexperienced or illiterate BCI subjects. It has been found that many illiterate subjects cannot produce stable mental patterns for the control of the system, thus, adaptation to the spatial distribution of their brain activity has been decisive for the successful completion of the study. The adaptation to this changing brain activity made necessary the design of new adaptive classifiers and both ingredients jointly have provided high quality feedback that allowed our illiterate subjects take control of the BCI system. Finally, new invariant techniques are developed for the design of a robust system partly invariant partly adaptive whose quality could bring BCI technology to society.