Abstracts

Rationale: Intracranial single-pulse electrical stimulation (SPES) may evoke high-frequency oscillations (HFO) and delayed responses (DR), considered as biomarkers of the epileptogenic networks. A number of brain structures are known to exhibit such responses even when non-pathologic. In some patients, these responses are not present at all. These are limiting factors for the use of such biomarkers. We perform a retrospective analysis in 11 patients of the responses in various brain structures that have been explored with depth electrodes and calculate the probability for each structure to exhibit normal or pathological responses. Methods: We recorded responses to SPES in 11 subjects undergoing presurgical evaluation for focal epilepsy using stereoelectroencephalography (SEEG). We designed a SPES protocol consisting in biphasic pulses of 3 ms duration and variable intensity in the range 0.25 to 5 mA, to explore the commonly used range of stimulus intensities that could evoke HFOs and DRs. We stimulated in each patient between 13 and 30 contact pairs and recorded responses on the other 62 contacts. HFOs and DRs were identified on each contact, and their frequency of occurrence in various brain structures was analyzed in relationship with three parameters: stimulus location, response location, and epileptogenicity of each structure. The epileptogenicity of each structure was further divided in seizure onset zone (SOZ), epileptogenic zone (EZ) and irritative zone (IZ), based on the standard analysis of the spontaneous inter-ictal activity and ictal propagation. 3D probability maps by stimulus location (outbound maps) and response location (inbound maps) were created for physiological and pathological occurrence of the biomarkers. Results: The DR and HFO biomarkers were evoked by SPES in 7 (DR) and 9 (HFO) of the 11 patients. The probabilities of DRs and HFOs occurrence were calculated for 8 cortical areas and structures that we sampled by recording or stimulation. Certain structures, like amygdala, exhibited inbound DRs in 75% of the cases where it was pathological (EZ) and outbound DRs in 40% of the cases. In the same structure, inbound HFOs were present in 46.7% of the cases where part of EZ, while the outbound HFOs were present in 26.7% of the stimulations. Accordingly, the physiological HFOs accounted for 53.3% to 73.3% of the responses, depending on the stimulus/recording location classification factor. Conclusions: We provide an estimation of the stimulation-evoked pathological versus physiological HFO and DR frequency, that can help in taking decisions regarding the delineation of the resection area.