Skip to Main content Skip to Navigation
Theses

User Adapted Brain-Computer Interface

Abstract : Brain-Computer Interface (BCI) allows communication between a user and a machine, by converting the user's brain activity into commands that control external devices. Many limitations prevent the diffusion of BCI systems in real applications, such as the calibration phase that is a consequence of the issue of variability across sessions and among users. The calibration phase is fundamental because it allows to set the main parameters to extract the relevant information from the electroencephalograpy (EEG) signal of the subject, but it is considered time consuming and tedious for the user.The objective of this thesis is to overcome these limitations by novel methods based on the improvement or even replacement of the traditional calibration phase, proposing the development of a user-centered BCI system.Firstly, we present a design to develop an adaptive BCI system for two different applications. The former deals with a code-modulated Visual Evoked Potential (c-VEP) speller where an adaptive parameter setting phase is proposed to replace the standard calibration phase. The latter application concerns the development of a Mental Imagery (MI) BCI for a disabled user, characterized by a long user-centered multi-stage training phase, in the context of a international BCI competition.Secondly, we propose an auto-calibration c-VEP BCI system exploiting the language information. In our model the fundamental properties that characterize the VEP response are used to predict the full word using a dictionary, eliminating the traditional calibration phase.The proposed methods showed promising results and open new perspectives to the diffusion of BCI
Document type :
Theses
Complete list of metadata

https://hal.inria.fr/tel-03149221
Contributor : Abes Star :  Contact
Submitted on : Tuesday, March 23, 2021 - 11:17:10 AM
Last modification on : Wednesday, March 24, 2021 - 3:32:05 AM

File

2020COAZ4043.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03149221, version 2

Collections

Citation

Federica Turi. User Adapted Brain-Computer Interface. Human-Computer Interaction [cs.HC]. Université Côte d'Azur, 2020. English. ⟨NNT : 2020COAZ4043⟩. ⟨tel-03149221v2⟩

Share

Metrics

Record views

30

Files downloads

197