Abstracts

The variety of interictal spike topologies may provide insight into the number and size of putative epileptogenic zones. We wished to examine whether high spatial density EEG recording could provide more information on spike waveforms and topology leading to better characterization of the interictal activity.
Six epilepsy patients were recruited for this study, selected for frequent interictal activity and a diagnosis of medically refractory extratemporal epilepsy. We recorded 30-90 minutes of 256 channel EEG at 500 samples/second using a comfortable electrode net that allowed the subject to lie on a bed in our sleep lab. After data acquisition, obvious bad channels were removed and all remaining channels were high-pass filtered at 1.0 Hz. Automatic spike detection was done off-line using Insight software [Persyst Development Corp, AZ], first with the full 256 channel data and then with subsets of 128, 64, 32, and 19 channels. The identical default detection parameters were used throughout. After spike detection, automatic clustering was performed, followed by manual deletion of events that were not spikes.
As the electrode density increased, the number of detected spikes also increased approximately linearly. The number of clusters also increased, though not quite so rapidly as the number of detected spikes.
Higher spatial sampling of the EEG can be used to objectively identify and further sub-classify interictal spike waveforms and topology.
[Supported by: Electrical Geodesics, Inc. loaned a 256 channel EEG system to Dr. Holmes for data acquisition.]