Abstracts

Rationale: Tripolar concentric ring electrode (TCRE) electroencephalography (tEEG) was first introduced by Besio [1]. The novelty of the TCRE and instrumentation is that two bipolar differential signals from three closely spaced electrode elements are recorded. Then the tripolar Laplacian derivation first described in [1] as a weighted sum {16*(M-D)-(O-D)} where O, M, and D are the potentials on the outer ring, middle ring, and central disc of the TCRE, respectively, is performed. We have shown that compared with conventional EEG signals, tEEG has nearly 4-fold (374%) the signal to noise ratio and less than one-tenth (8.27%) the mutual information [1, 2].

The goal of this work was to demonstrate that TCREs provide a unique opportunity to record high-frequency oscillations (HFOs) from scalp and develop a procedure to detect them that may be automated. We expect these techniques to improve diagnosis of epilepsy. Methods: The recording protocol was approved by the IRB committees and did not interfere with the clinical EEG recording and evaluation. The tEEG recordings were performed concurrently with the clinical EEG. The TCREs were placed just behind the disc electrodes in locations close to the 10-10 sites.

After acquiring the t/EEG we used a modified version of the time course algorithm reported by Gardner et al. for detection of HFOs [3]. The algorithm performs a continuous short-time Fourier transform to calculate the power within a particular frequency band over consecutive half-overlapping one second epochs using a Hamming tapering window. As a result, the time course of power modulations was obtained. We found that HFOs were visually evident in the tEEG but were not present in the EEG.

However, with a threshold set at the mean plus one standard deviation the time course was only able to automatically detect the HFOs in one of our five patient's data. Various thresholds were tested with no improvement. Therefore, instead of calculating the power within a particular frequency band over time, we calculated the power spectrum within a short period of time using the same mean plus one standard deviation threshold. Results: For each of the five patients we were able to visually observe where the peaks in the high gamma-band burst HFOs were with both methods. The automated threshold correctly detected the high gamma-band burst HFOs in four of the five patient's data using the power spectrum. Conclusions: The power spectrum method more consistently lead to automatic HFO detection than the time course on our tEEG data for these five patients.

[1] WG. Besio et al."Tri-polar Concentric Ring Electrode Development for Laplacian" IEEETBME. 2006.
[2] K. Koka, WG. Besio, "Improvement of Spatial Selectivity and Decrease of Mutual Information of TCRE" Neurosci Meth, 2007.
[3] AB Gardner et al. "Human and automated detection of HFOs" Clin. Neurophys, 2007.

Funding - Fogarty International Center of the NIH R21TW009384 and the NSF OISE 10494. Dr. Fisher is supported by the Anderson Research Fund for Epilepsy and the Maslah Saul MD Chair.