Towards Robust Motor Attempt BCI control

Activity: Talk and presentation › Academic presentation › Academic

Description

Non-invasive motor intention based BCIs have been used to control various devices. We intend to address two challenges related to these BCIs in an upcoming experiment. Firstly, we attempt to replicate Motor Attempts in healthy subjects, as a more consistent alternative to Motor Imagery. Secondly, we propose the introduction of Uncertainty Quantification to identify when a model’s prediction is likely to be wrong due to artifacts, off-task thoughts or poor class separability. This should minimize the risk of misclassified Motor Attempts, improve usability, and be more generalizable to paralyzed patients.
Period17-Oct-2023
Event titleCutting Gardens 2023: EEG and MEG methods multi-hub meeting
Event typeConference
Degree of RecognitionLocal

Keywords

  • EEG
  • Brain Computer Interface
  • Uncertainty Quantification
  • Machine Learning