Robust brain-computer interfaces


Share/Save/Bookmark

Reuderink, Boris (2011) Robust brain-computer interfaces. thesis.

[img]
Preview
PDF
3020Kb
Abstract:A brain-computer interface (BCI) enables direct communication from the brain to devices, bypassing the traditional pathway of peripheral nerves and muscles. Current BCIs aimed at patients require that the user invests weeks, or even months, to learn the skill to intentionally modify their brain signals. This can be reduced to a calibration session of about half an hour per session if machine learning (ML) methods are used. The laborious recalibration is still needed due to inter-session differences in the statistical properties of the electroencephalography (EEG) signal. Further, the natural variability in spontaneous EEG violates basic assumptions made by the ML methods used to train the BCI classifier, and causes the classification accuracy to fluctuate unpredictably. These fluctuations make the current generation of BCIs unreliable. In this dissertation,we will investigate the nature of these variations in the EEG distributions, and introduce two new, complementary methods to overcome these two key issues. To confirm the problem of non-stationary brain signals, we first show that BCIs based on commonly used signal features are sensitive to changes in the mental state of the user. We proceed by describing a method aimed at removing these changes in signal feature distributions. We have devised a method that uses a second-order baseline (SOB) to specifically isolate these relative changes in neuronal firing synchrony. To the best of our knowledge this is the first BCI classifier that works on out-of-sample subjects without any loss of performance. Still, the assumption made by ML methods that the training data consists of samples that are independent and identically distributed (iid) is violated, because EEG samples nearby in time are highly correlated. Therefore we derived a generalization of the well-known support vector machine (SVM) classifier, that takes the resulting chronological structure of classification errors into account. Both on artificial data and real BCI data, overfitting is reduced with this dependent samples support vector machine (dSVM), leading to BCIs with an increased information throughput.
Item Type:Thesis
Faculty:
Electrical Engineering, Mathematics and Computer Science (EEMCS)
Research Group:
Link to this item:http://purl.utwente.nl/publications/78274
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page

Metis ID: 284894