Abstracts

Rationale: Intracranial recording from microwires, implanted with clinically required depth electrodes in epilepsy patients, permits observation of single neuron activity in the conscious human brain. Although these microwire electrodes have typically been placed for research purposes, we wondered how the characteristics of the single-unit activity recorded from these microwires might be correlated with clinical characteristics of epilepsy and if the recordings could be used as an additional information for clinical evaluation. Methods: Toward this end, we examined how the yield of isolated cell activity varies as a function of brain area, the number of days after implantation surgery, and whether the area was determined to be a likely origin of the seizures and subsequently resected. For each pair of implanted microwires, the extracellular voltage was filtered (300-3000 Hz) and 1.15 ms epochs surrounding voltage extrema (> 2.8 s.d. for the channel, 1.15 ms long) were isolated for further analysis. Events for each channel of recording were automatically grouped into clusters of similar waveform shape (G. Celeux, G. Govaert, Computational Statistics and Data Analysis 14, 315, 1992). Each cluster was graded whether it represented isolated single neuron activity on three (0-4) scales: quality of waveform shape, lack of inter-event intervals < 3 ms, lack of power-line frequency harmonics in the event time power spectrum. Using the sum of all three scales, each cluster was classified as either noise (0-3), potential single neuron activity (4-5), or single neuron (SPIKE) activity (?6). Results: We examined recordings from 1205 microwires implanted in 14 different cortical areas in 8 patients resulting in 5895 separate clusters of similar waveform shape. 508 clusters of SPIKE activity were identified (average 0.42 SPIKEs/ch). Both the brain area and its subsequent resection had a significant effect (p=0.0013) on SPIKE yield (SY). For the four brain areas with the largest number of recordings (hippocampus-H, amygdala-A, entorhinal cortex-EC, and orbitofrontal cortex-OF) considered separately, subsequent resection had a significant effect (p<0.05) on SY in 3 areas, increasing yield in H, but decreasing yield in EC and OF. The number of days after implantation surgery had a significant (p=1e-5) effect on SY, trending upward toward 0.64 SPIKEs/ch at 7 days post-implantation.