Big data resources for EEGs: Enabling deep learning research

The Temple University Hospital (TUH) electroencephalography (EEG) Corpus is the world's largest open source EEG corpus of its kind [1]. This corpus consists of over 25,000 EEG studies and over 14,000 patients, and includes a neurologist's interpretation of the test, a brief medical history of the patient, and demographic information about the patients such as gender and age. This database represents the efforts of the Department of Neurology and the Neural Engineering Data Consortium to support the use of EEG data in machine learning research. The data was collected in normal clinical settings and hence includes many non-epileptic features such as muscle and movement artifacts, and a variety of channel configurations that cannot be found in currently available, more sanitized datasets. This is the first dataset of its kind to contain a sufficient amount of EEG data to support the application of state of the art deep learning algorithms. The most recent release of this corpus is vl.0.0 which includes 13,550 patients, 23,218 EEG sessions with reports and 61,634 EEG files.