Interictal High Frequency Oscillations Detected with Simultaneous Magnetoencephalography and Electroencephalography as Biomarker of Pediatric Epilepsy

The overall goal of this protocol is to provide a standardized methodology for the reliable non-invasive recording detection and localization of interictal high-frequency oscillations from pediatric patients with medically refractory epilepsy using simultaneously recorded scalp electroencephalography and magnetoencephalography. This method can help answer key questions in pediatric epilepsy, such as which brain area is epileptogenic and should be resected during surgery. The main advantage of this technique is that it allows the non-invasive localization of the area that generates high-frequency oscillations in the brain of children with epilepsy.

Different studies have shown that high-frequency oscillations can be detected noninvasively using scalp EEG and MEG, but only a few studies localize their generators at the source level. We present the localization of the underlying generators of high-frequency oscillations obtained by solving the inverse problem and compare it with the determined by the epileptologist. Begin by placing the electroencephalography or EEG cap on the patient’s head according to the international 10-20 system.

Cleanse the skin where each electrode is located, move hair out of the way, and then apply gel for each electrode. Next, place the ground and reference electrodes on the patient’s head. Connect additional electrodes for measuring horizontal and vertical electrocochleography, EOG, electrocardiography, ECG, electromyography, EMG, and additional EEG electrodes at locations covering the temporal regions.

Then, use an EEG-o-meter to measure the impedance for each electrode. If the impedance is above 10 kiloohms, cleanse the skin again and ensure that it decreases below 10 kiloohms. Place four HPI coils on the EEG cap.

Use the digitizer to obtain the locations of fiducial landmarks, including left/right pre-auricular points and the nasion as well as the location of the HPI coils and the EEG electrodes. Digitize additional head points to obtain precise head shape. Then, transfer the patient into the magnetic shielded room, or MSR, where the magnetoencephalography, MEG, system is located.

Lay the patient down on the bed. Put their head into the MEG helmet and apply appropriate pads under the patient’s head for comfort. Adjust the patient’s head position in the scanner, ensuring that it is located as deep as possible in the helmet.

Finally, close the door to the MSR and communicate with the patient via an intercom system. Start the recordings by clicking the go button in the MEG acquisition software. Check online for all the recorded signals and fix bad MEG channels by using a sensor tuner.

Next, measure the patient’s head position by clicking the measure button in the MEG acquisition software. If the patient’s head is not well covered by the sensory array, ask the patient to move his or her head deeper into the helmet. Then, in the MEG acquisition software, click the record button to record MEG, EEG, and obtain peripheral recordings for 60 minutes.

After the recording is complete, open the MSR and take out the patient from the MSR room. Gently remove all the tapes, electrodes, HPI coils, and the EEG cap. Finally, after escorting the patient out of the testing space, record the magnetic signals of the empty MSR for two minutes without the patient present.

Begin by opening the data in the analysis software. Display EEG and MEG data with two vertically aligned windows of 10 seconds per page. Go to the filter tab and set the high-pass filter to one hertz, the low-pass filter to 70 hertz, and the notch filter at 50 or 60 hertz.

Inspect the data and identify portions with interictal epileptic discharges or IEDs. Mark the peak of each IED occurring in both the EEG and MEG data. Run the algorithm for the automatic detection of HFOs on the portions of EEG data with IEDs and import the detected HFOs into the software for data visualization.

To review the detected HFO events, display EEG, MEG, and peripheral recordings with vertically aligned windows of two seconds per page. Go to the filter tab and set the low-pass filter to 250 hertz and the high-pass filter to 80 hertz. To ensure that the detected HFOs are not due to artifacts, verify that there is not concurrent activity in the peripheral recordings.

Furthermore, consider only the HFO events that occur in both EEG and MEG signals and disregard HFOs that do not overlap with the marked IEDs. Next, segment the patient’s MRI and obtain the cortical surface using brain imaging analysis software and estimate the forward model with the boundary element method. For each HFO event, perform the source localization on both EEG and MEG data using the wavelet maximum entropy on the mean or WMEM method.

Average the localization results along the event duration to obtain a source localization map. Finally, visualize the HFO zone over the cortical surface, applying a threshold of 60%of the maximum activation amplitude using both EEG and MEG data. In these patient examples, HFOs were identified in the ripple frequency band at 80 to 150 hertz occurring in both EEG and MEG and overlying IEDs.

The time frequency plane of two representative EEG and MEG channels shows the typical isolated peak at the time of the HFO. Here, HFO localization results are shown from both scalp EEG and MEG for a 15-year-old girl with encephalomalacia of the right middle cerebral artery region. Both techniques localize the HFOs at the vicinity of the lesion in a location close to the right temporoparietal junction, although they were slightly different.

Here, the results for an 11-year-old boy with a left parietal superior temporal encephalomalacia are shown. The figure shows the spatial concordance between the HFO zone localized noninvasively with scalp EEG and MEG, the IEDs localized by MEG, and the HFO zone localized invasively with iEEG. The location of the intracranial electrodes with the highest number of HFOs, LA51, LA52 and LA53, was concordant with the HFO activity localized noninvasively and both of them overlapped with the seizure onset zone.

This is the first study to report the localization of interictal high-frequency oscillations with simultaneous EEG and MEG recordings that also investigates the concordance of the localization results with those from intracranial recordings. The noninvasive recording, detection, and localization of high-frequency oscillations is challenging because HFOs are very weak signals generated by small brain regions on the order of cubic millimeters and hindered by noise and background brain activity. So far, few studies managed to show that high-frequency oscillations can be detected noninvasively by using electroencephalography and magnetoencephalography and localized this activity by solving the inverse problem.

Mastering, improving, valuating the proposed protocol will provide physicians reliable, noninvasively recordable biomarker for the identification of the epileptogenic zone. The development of biomarker has the potential to reduce the requirement for long-term monitoring and invasive intracranial recordings and significantly improve the pre-surgical evaluation procedure of pediatric patients with epilepsy.