The ipsilateral thalamus was not recruited before the contralateral hemisphere in pediatric SEEG seizures

We tested whether the ipsilateral (Ip) thalamus is recruited before the contralateral (Co) hemisphere, as predicted if the thalamus is an early relay for bilateral seizure spread. We used the archived bipolar-pair latencies from the prior study (SCI fae69370; OpenNeuro ds007445 v1.0.4), without reading new EEG data or changing its detector. In 15 patients, the patient-level latency difference (Ip thalamus minus Co hemisphere) was +5.8 s (median 0.0 s; 3 negative, 7 positive; 95% CI -5.4 to 16.4 s; exact two-sided sign-flip p = 0.33; Holm p = 0.98). The recruited-fraction difference was -0.09 (p = 0.37), and the thalamus-minus-Co-cortex latency difference was +5.5 s (14 patients; p = 0.39). With the 25th-percentile group summary, the corresponding differences were +5.2 s (p = 0.43) and +3.1 s (p = 0.64). Across 41 seizures with both groups, the thalamus was first in 9, the Co hemisphere was first in 13, and 19 were ties, mostly because both group medians were censored at 60 s. The Co hemisphere was not clearly later than Ip cortex (+3.8 s, p = 0.50), but was later than the seizure-onset zone (+25.6 s, p = 0.00067). These results do not support an early thalamic relay before contralateral spread for this detector. Line length can miss low-amplitude rhythmic or phase changes in the thalamus, and censoring makes many latencies lower bounds; the results do not exclude a thalamic role that the detector cannot see.

stereo-EEG · SEEG · seizure spread · thalamus · recruitment latency · line length · pediatric epilepsy · censoring