Adaptive stimulation pattern optimization of a VEP-based Brain-Computer Interface
Master’s Thesis, assigned to Ulzii-Utas Narantsatsralt
Lab: Neural Interfaces and Brain Signal DecodingUlzii-Utas Narantsatsralt
A Brain-Computer Interface (BCI) is an interface between the human brain and a computer. BCIs enable physically paralyzed patients, for example, to control a computer via pure brain activity. The EEG2Code BCI developed in Tübingen is currently the fastest non-invasive BCI system. The schematic flow of the method is shown in the figure. Each letter of the virtual keyboard flickers with a random stimulation pattern. This flickering generates so-called visually evoked potentials (VEPs), which can be recorded by electroencephalography (EEG). The EEG is then classified using a previously trained neural network, whereby the output corresponds to the stimulation pattern of the letter the user has been looking at (with a certain accuracy). The developed BCI already enables fast and asynchronous control, but tests have shown that certain stimulation patterns lead to a better/faster classification.
The aim of this work is to develop a method that adaptively optimizes the stimulation patterns. The method should “rate” the stimulation patterns, which are random at the beginning. The better rated stimulation patterns should then be used more frequently than the worse ones. In order to evaluate this, it is necessary to test the system on itself as well as with other test persons. In this respect there will be an introduction on how to prepare an EEG. In addition, the stimulation patterns will then be analyzed with the aim of identifying certain characteristics that lead to a better classification
- Interest in neuroscience
- good programming skills and experience in MATLAB are beneficial