ICA of Auditory Evoked Potentials in Children with Cochlear Implants: Component Selection

Multi-channel recordings of scalp electroencephalogram (EEG) for auditory evoked potentials (AEPs) can be used to objectively evaluate hearing sensitivity in both adult and children. This paper uses Independent Component Analysis (ICA) to decompose ongoing (i.e. unaveraged) AEPs recorded from normal hearing children and children with cochlear implants (CIs). It is seen that ICA can decompose the ongoing EEG into independent components which appear to exhibit the expected characteristics of AEPs, along with other components such as artifacts and ongoing brain activity. This is particularly useful for children with CIs as there is significant CI artifact present in the recordings. We depict here some preliminary results of applying ICA to ongoing multi-channel EEG for both a normal hearing child as well as a child with a CI.