ABSTRACT
An electroencephalogram (EEG) records the electrical activity of the brain. It is obtained by placing multiple electrodes over the scalp and recording the summated potential difference under the electrode. This potential difference is due to the underlying brain electrical activity. EEG has become an important diagnostic tool in routine clinical practice. Domain experts examine EEG for various named and unnamed patterns while interpreting the EEG. This is time consuming considering that EEG records are typically multichannel and change rapidly over time and with mental status. Currently, EEG evaluation is the standard of care for epilepsy management, and its role in other conditions is evolving. In recent times, with improvements in computing power and other technology, much work has been done on EEG using machine learning. All this work has shown the ability of EEG data to contribute important diagnostic and prognostic information in many neurological and nonneurological disorders.
In this chapter, we begin with a brief introduction to EEG, its generators, and recording techniques. Then we discuss the status of machine learning in EEG–the processing pipelines, applications in a few common neurological disorders, and its impact on clinical practice and further research.
