Abstracts

Rationale: LiTT outcomes in mTLE are variable. This study evaluates whether SEEG-based time-frequency patterns at ictal onset can predict which patients will benefit most from LiTT.

Methods: We retrospectively reviewed patients with refractory mesial temporal lobe epilepsy
(mTLE) who underwent SEEG-confirmed seizure onset in the hippocampus and amygdala,
followed by LiTT targeting these regions. At least two typical seizures were analyzed per patient.
TF analysis was performed using the Short-Time Fast Fourier Transform (STFFT) in CURRY 9
(Compumedics NeuroScan). A band-pass filter (80–250 Hz) was used for high-frequency
components, and a second-order Butterworth filter was applied for 0–80 Hz frequencies. STFFT
settings included a resolution of 128 ms, scaling factor of 4.0, and frequency range of 0–250 Hz.
Neuroimaging data were also reviewed.

Results: Eight patients met inclusion criteria. Two patients achieved Engel Class I-A seizure
outcomes after LiTT. In these cases, TF analysis of ictal onset revealed sustained high-frequency
activity (125–150 Hz) with simultaneous suppression of lower frequencies in two or fewer
hippocampal and/or amygdala SEEG contacts, with no activity in extra-mesial temporal (MT)
electrodes. Among six other patients (Engel Class II–IV), TF analysis revealed either (1)
multifrequency activity below 80 Hz within MT electrodes, with or without preictal spikes (n=5),
or (2) sustained high-frequency activity in both MT and extra-temporal electrodes (n=1). Two
patients with Class IV outcomes later underwent temporal lobectomy and subsequently attained
Class I-A outcomes. Brain MRI was normal except in two cases: Patient 7 had a hyperintense
signal in the right medial temporal region, and Patient 10 had left mesial temporal sclerosis
confirmed pathologically. PET imaging revealed temporal lobe hypometabolism in all patients.

Conclusions: In our cohort, all patients exhibited seizure onset within the hippocampus and
amygdala; however, only two achieved seizure freedom following LiTT targeting these regions.
This variability underscores the need for more precise predictive tools. SEEG-based ictal onset
fingerprinting—characterized by sustained high-frequency ripple activity with simultaneous low-
frequency suppression in two or fewer SEEG electrode contacts—may serve as a useful
biomarker for predicting favorable LiTT outcomes in mTLE. Conversely, patients lacking this
pattern often experience suboptimal results with LiTT. In such cases, temporal lobectomy has

demonstrated superior efficacy, achieving seizure freedom. These findings highlight the
importance of incorporating SEEG-based biomarkers into preoperative evaluations to guide
surgical decision-making and optimize treatment strategies for individuals with mTLE.

Funding: N/A