The spike onset zone: The region where epileptic spikes start and from where they propagate

Abstract

Objective: To determine whether the maximum hemodynamic response to scalp interictal epileptic discharges (IEDs) corresponds to the region where IEDs originate and from where they propagate.

Methods: We studied 19 patients who underwent first an EEG-fMRI showing responses in the gray matter, and then intracranial EEG (iEEG). We coregistered the hemodynamic responses to the iEEG electrode contacts and analyzed IEDs in the iEEG channel adjacent to a maximum response (labeled the main channel), in relation to IEDs in other channels during a widespread intracranial IED event. IEDs in the main channel were aligned at their peak, and IEDs in each channel were averaged time-locked to these instants. The beginning and peak of IEDs in the averaged trace were identified, blinded to the identity of the main channel. The latency of IEDs was computed between the earliest and all other channels.

Results: The median latency of IEDs in the main channel was significantly smaller than in other channels for either the peak (15.5 vs 67.5 milliseconds, p = 0.00037) or the beginning (46.5 vs 118.4 milliseconds, p = 0.000048). The latency of IED was significantly correlated to the distance from the maximum hemodynamic response (p < 0.0001 for either the peak or the beginning).

Conclusion: IED adjacent to a maximum hemodynamic response, which often corresponds to the seizure onset zone, is more likely to precede IEDs in remote locations during a widespread intracranial discharge. Thus, EEG-fMRI is a unique noninvasive method to reveal the origin of IEDs, which we propose to label the spike onset zone.

Figure 1. Methods to yield the averaged traces of intracranial EEG

(A) Approximately 100 IEDs were marked on the main channel and aligned at their first clear peak (blue arrows). Then, the intracranial EEG at other channels were averaged time-locked to these instants (blue dashed lines), resulting in averaged traces in which time 0 was aligned at the peak of the main channel (green dashed line). The first clear peak of the IED on each of the averaged traces was marked during the second marking (red dashed line with arrowhead), which was performed blinded to the identity of the main channel. Latencies were calculated as the time difference of peaks between the first channel (the channel showing the earliest peak; in this example, the first channel is the main channel) and each of the remaining channels (the differences between the green and red dashed lines as indicated by pairs of the black arrows). (B) Example of an IED on an averaged trace of an intracranial EEG channel. The peak was visually marked and the beginning was automatically detected. IED = interictal epileptic discharge.

Figure 2. Example of IED recorded closest to the maximum hemodynamic response being the earliest (patient 13)

(A) Scalp EEG, bipolar montage, spikes with equipotentiality at F8-T4 and F10-T10. (B) White circles indicate the contacts visible in these slices. Contacts corresponding to channels closest to fMRI maximum (main channel) were marked in red (2 adjacent contacts make 1 channel on intracranial EEG). Yellow-red scale corresponds to t values of positive hemodynamic responses. (C) IEDs of all intracranial EEG channels. Each line corresponds to an averaged trace of an intracranial EEG channel. Red = main channel. Dashed black line = peak of IED in the main channel. Dotted black line = peak of IED in the first channel. The scalp IEDs (see panel A) resulted in a maximum hemodynamic response in the right lateral orbitofrontal gyrus (t value = 14.1). The patient underwent stereo-EEG with a suspicion of right frontal lobe epilepsy. Nine electrodes were inserted in the right hemisphere: 6 in the frontal lobe, 3 in the temporal lobe. In the frontal lobe, ROF aimed the mesial orbitofrontal, RCa the anterior cingulate, RCm the midcingulate, RSMAa the anterior supplementary motor area, and RSMAp the posterior supplementary motor area, RIm the insula through the frontal lobe; in the temporal lobe, RA the amygdala, RH the hippocampus, RHp the posterior hippocampus. The intracranial study revealed very active IEDs in the lateral orbitofrontal region. The peak of IED in the main channel was the earliest (among 25 channels showing IEDs in the averaged traces). IED = interictal epileptic discharge.

Figure 3. Example of IED recorded closest to the maximum hemodynamic response being the earliest (patient 10)

(A) Scalp EEG, average montage, Fp2, F4, F8, T4, T6, F10, T10, P10 spike and waves. (B) White circles indicate the electrode contacts visible in these slices. Contacts corresponding to channels closest to fMRI maximum (main channel) were marked in red (2 adjacent contacts make 1 channel on intracranial EEG). Green-blue scale corresponds to t values of negative hemodynamic responses. (C) IEDs of all intracranial EEG channels. Each line corresponds to an averaged trace of an intracranial EEG channel. Red = main channel. Dashed black line = peak of IED in the main channel. Dotted black line = peak of IED in the first channel. The scalp IEDs (see panel A) resulted in a maximum hemodynamic response in the right angular gyrus (t value = −24.2). The patient underwent stereo-EEG with a suspicion of right posterior quadrant epilepsy. Eleven electrodes were inserted in the right hemisphere: 2 in the frontal lobe, 9 in the temporal lobe and posterior quadrant structures. In the frontal lobe, ROF aimed the mesial orbitofrontal, RCa the anterior cingulate; in the temporal lobe and posterior quadrant, RTp aimed the temporal pole, RA the amygdala, RH the hippocampus, RpT1 the inferior insula, RFug the fusiform gyrus, RSmg the supramarginal gyrus, RAg the angular gyrus, RLi the lingual gyrus, and RCu the cuneus. The intracranial study revealed very active IEDs in the right posterior quadrant structures. The peak of IED in the main channel was the earliest (among 56 channels showing IEDs in the averaged traces). IED = interictal epileptic discharge.