Towards noise-enhanced augmented cognition

Workload classification Augmented Cognition systems aim to detect when an operator is in a high or low workload state, and then to modify their work flow and operating environment based upon this knowledge. This paper reviews state-of-the-art electroencephalography (EEG) recorders for use in such systems and investigates the impact of EEG noise on an example system performance. It is found that adding up to 15 μV of artificially generated noise still leaves EEG signals that have correlations in-line with the correlations found between conventional wet EEG electrodes and new dry electrodes. The workload classification system is found to be robust in the presence of small amounts of noise, and there is initial evidence of small stochastic resonance effects whereby better performance can actually be obtained in the noisy case compared to the traditional noise-less case. © 2013 Springer-Verlag Berlin Heidelberg.