Abstracts

Rationale:
Much recent debate has focused on cell-type-specific firing patterns at seizure onset. Action potential waveform alterations during seizures, however, can impede the ability to track neuronal firing through ictal recruitment. We therefore employed a novel template-matching method to identify single neurons in microelectrode recordings in patients with intractable focal epilepsy, in order to assess firing patterns through local recruitment to the seizure.
Method:
Twenty-seven patients with focal onset seizures undergoing epilepsy surgery evaluation were simultaneously implanted with either Utah microelectrode arrays (UMA; n = 6) or Behnke-Fried microwires (BF; n = 21). Single neurons (units) were defined in the peri-ictal period using standard spike sorting techniques and were used to build templates in principal component space to define units during the seizure. Probabilities of each unit being a fast-spiking interneuron were calculated via waveshape (trough to peak delay) and its autocorrelation – a cell-intrinsic firing pattern. Results1,239 single neurons were tracked through 41 seizures, with 14% being defined as > 80% confidence of being interneurons. We identified two distinct firing patterns at seizure onset. Type 1 was characterized by tonic firing and waveform alterations (p < 0.05, Mann-Whitney U test; UMA: 6 seizures, 3 patients, BF: 18 seizures, 12 patients). Type 2 showed burst firing with no change in waveform (UMA: 4 seizures, 3 patients, BF: 13 seizures, 9 patients). Type 2 recordings showed stable firing across the population. In neocortical (UMA) Type 1 recordings, interneuronal firing became out-of-phase to the pyramidal cells in the seconds prior to tonic firing and waveform changes (p < 0.05, Kuiper’s test). The interneuronal firing rate peaked earlier than the pyramidal cell population, prior to decreasing again, but remained above baseline throughout the seizure, with 43 of 70 interneurons showing firing rates > 3 SD over preictal levels, and zero significantly decreasing. Conversely, in Type 1 recordings in mesial temporal structures (BF), 10 of 11 interneurons significantly reduced firing rate at seizure onset prior to waveform alterations, and none increased.
Conclusion:
We conclude that action potential waveform alterations, out-of-phase interneuronal firing, and the transient increase followed by decrease in inhibitory firing found in Type 1 recordings are key events in the ictal transition in neocortical recordings. These observations are consistent with our predictions of single unit dynamics in recruited neocortical tissue, and have implications for single unit studies of this time period. Meanwhile, dichotomous interneuronal firing patterns in mesial temporal vs. neocortical recordings suggest that seizure propagation in these two regions may occur through different mechanisms. Finally, we conclude that fast-spiking interneuronal firing does not drive seizure activity, with firing occurring out-of-phase to the dominant ictal rhythm in neocortical recordings, and cessation of interneuronal firing in many mesial temporal recordings.
Funding:
:NIH R01 NS084142 CRCNS R01 NS095368