Figure 1.

We tested the automated seizure detection method developed by Cho and Jang using data collected in our lab from chronically epileptic mice (recorded using depth electrodes in the ventral hippocampus, frontal cortex electroencephalogram [EEG], and electromyogram), downsampled to 100 Hz in 5-second segments, with a seizure to nonseizure ratio of approximately 1 in 5. The colored segments indicate individual seizures recorded in the ventral hippocampus detected using the multichannel 1D convolutional neural network (CNN) model and demonstrate the ability to easily implement a deep learning model out of the box for automated seizure detection.