Abstracts

Rationale: Direct electrical stimulation (DES) to identify the seizure onset zone (SOZ) has potential as a tool to improve the success rate of surgery for patients with drug-resistant epilepsy. DES also offers a potentially powerful research tool to understand the neurophysiology of the SOZ and how it interacts with surrounding brain regions. We performed low-frequency DES and measured the evoked responses. We hypothesized that the evoked responses would have higher amplitudes when SOZ was stimulated or when recording within SOZ, as compared to outside the SOZ (non-SOZ), in accord with the concept that the SOZ is pathologically hyperexcitable.

Methods: We analyzed data from low frequency DES (bipolar/biphasic stimulation, 3 mA current, 300-500 µs pulse width, 1 Hz frequency) performed in 26 adult (16 female, age 21-53) and 42 pediatric (20 female, age 1.8-23) patients who underwent intracranial EEG monitoring for epilepsy surgery planning at the Hospital of the University of Pennsylvania or at the Children’s Hospital of Philadelphia. After an initial artifact rejection step, voltage time series data from the ~30 trials for each stimulation (stim) site were averaged for each gray matter response site. We then calculated the root mean square (RMS) of the resultant evoked responses during the +25 - > +250 ms time window after stim, scaled by the pre-stim baseline (-100 - > -15 ms). We utilized mixed-effects regression models to analyze the effect that recording or stim within the SOZ would have on the RMS value of the evoked responses; this approach accounts for the non-independence of within-subjects measures. We also controlled for the confound that response amplitudes recorded close to the stim site tend to be larger (distance : RMS correlation r = -0.23, p< 0.0001) by using a rational polynomial model of RMS as a function of stim-recording site distance.