The technique of multireference adaptive noise canceling (MRANC) is ap
plied to enhance transient nonstationarities in the electroencephalogr
am (EEG), with the adaptation implemented by means of a multilayer-per
ceptron artificial neural network (ANN). The method was applied to rec
orded EEG segments and the performance on documented nonstationarities
recorded. The results show that the neural network (nonlinear) gives
an improvement in performance (i.e., signal-to-noise ratio (SNR) of th
e nonstationarities) compared to a linear implementation of MRANC. In
both cases an improvement in the SNR was obtained. The advantage of th
e spatial filtering aspect of MRANC is highlighted when the performanc
e of MRANC is compared to that of the inverse auto-regressive filterin
g of the EEG, a purely temporal filter.