Eliciting code modulated visual evoked potentials by non-recognizable presentation of m-sequences

Master’s Thesis

Abstract

A brain-computer interface (BCI) allows it’s user to control their environment without muscles. Currently, the majority of non-invasive interfaces uses visual evoked potentials (VEPs) as control signal [0]. This is due to the high performance achieved within this methodological direction. These potentials are elicit by visual stimuli which are associated to the action the user wishes to execute. Most of the time these stimuli consist of a perceivable flickering sequence, thus are not end-user friendly and may lead to problems such as visual fatigue or migraine at best and to an epileptical attack at worst. Such an exhausting BCI is the c-VEP BCI [1] which uses a m-sequence as stimulus pattern.

Task

This thesis aims to research different encoding schemes for m-sequences, such that the m-sequence can be used as control signal while it’s perceived comfortable by the system’s user. The more practical part of the project is split into three phases:

  1. Research into the topic and subsequent design of the experiment and the data analysis.
  2. Implementation of an experiment collecting data suitable for an offline analysis.
  3. Offline analysis of the recorded data.

References

Requirements

  • Interest in EEG experiments.
  • Experience in MATLAB/Python is beneficial (for EEG data analysis).
  • Adequate knowledge of statistics and machine learning.

Contact

Blöck, Alexander