Discriminative channel selection method for the recognition of anticipation related potentials from CCD estimated cortical activity

Abstract

Recognition of brain states and subject's intention from electroencephalogram (EEG) is a challenging problem for braincomputer interaction. Signals recorded from each of EEG electrodes represent noisy spatio-temporal overlapping of activity arising from very diverse brain regions. However, un-mixing methods such as Cortical Current Density (CCD) can be used for estimating activity of different brain regions. These methods not only improve spatial resolution but also signal to noise ratio, hence the classifiers computed using this activity may ameliorate recognition performances. However, these methods lead to a multiplied number of channels, leading to the question -"How to choose relevant and discriminant channels from a large number of channels?". In the current paper we present a channel selection method and discuss its application to the recognition of anticipation related potentials from surface EEG channels and CCD estimated cortical potentials. We compare the classification accuracies with previously reported performances obtained using Cz electrode potentials of 9 subjects (3 experienced + 6 naïve). As hypothesised, we observed improvements for most subjects with channel selection method applied to CCD activity as compared to surface-EEG channels and baseline performances. This improvement is particularly significant for subjects who are naïve and did not show a clear pattern on ERP grand averages.

Cortical current density vs. surface EEG for event-related potential-based Brain-Computer Interface

“…A preliminary study on the feasibility of using CCD inverse solution for ERPs is reported in [7]. In our study, we found that the data from cortical sources not only increases classification accuracy for some subjects, but it also provides us an opportunity for neurophysiological interpretation of the ERP data in terms of active cortical sources.…”

Single trial analysis of slow cortical potentials: a study on anticipation related potentials

“…We have previously explored its single trial recognition for BCI, designing methods for fast and online classification [22][23][24]. However, classification performance was largely varied across runs and subjects, sometimes staying at levels close to chance.…”

Online EEG channel weighting for detection of seizures in the neonate

“…On the other hand, it becomes more difficult to automatically find what information is useful in the available channels. Channel selection has been widely used in Brain Computer Interfaces [4][5][6][7][8][9][10] mainly for the purpose of computational load reduction. Constant (in time) importance of information captured by a channel can be assumed for these tasks, especially when classification systems targeted are patientspecific, that is, some representation of testing patient data is available beforehand [4], [7][8][9][10].…”