Electroencephalography (EEG) research is challenged by the prevalence of artifacts, i.e., non-brain contributions to the EEG signal. Brain electrical signals are incredibly small in comparison to electrical noise from movements, with eye blinks in particular 2-3 orders of magnitude larger than brain activity of interest. Approximately 10% of our current study's data is contaminated by artifacts. This data is taken from 2 procedures, one where the phone serves as a distractor while the participant completes a task, and another involving passive phone viewing with no task. One approach to reclamation of otherwise unusable data is a preprocessing technique called Independent Component Analysis (ICA). ICA allows for blind separation of signals into separate components, like separating a combined music track into individual tracks for each instrument. Using this, we are able to separate brain and artifact contributions and reconstitute a signal factoring out major noise components. So far I have used ICA for the passive viewing portion of our EEG data, increasing the percent usable data from 90% to 98%. Datasets from participants with less than 90% usable EEG data disproportionately benefited from ICA, as those had more data to reclaim.
