Two 60-s sleep clips from 18 pediatric SEEG patients showed no consistent 0.5–2 Hz latency gradient

We tested whether 0.5–2 Hz activity in sleep stereo-electroencephalography (SEEG) shows anterior-to-posterior or superior-to-inferior latency gradients. We analyzed two continuous 60-s clips from each of 18 UCLA pediatric epilepsy patients in OpenNeuro ds005398, using non-seizure-onset-zone bipolar channels with grey-matter labels and MNI coordinates. For the primary high-frequency-activity envelope measure, clip A group mean Spearman correlations between channel latency and MNI y and z were −0.040 (95% CI −0.137 to 0.059; 18 patients) and −0.048 (−0.162 to 0.084; 17 patients), respectively; both Holm-adjusted p values were 0.89. Clip B did not replicate a gradient. Secondary analyses were also null apart from one uncorrected clip A LFP z-axis slope (−2.42 ms/10 mm, p = 0.025), which did not replicate (clip B: −1.66 ms/10 mm, p = 0.45). Per-patient gradients had low repeatability between clips (Pearson r = 0.05 for y and 0.12 for z). Median pairwise high-frequency-envelope correlation was 0.103 and 0.102 in clips A and B, compared with 0.084 for filtered independent noise. These results do not support a consistent axis-aligned latency gradient in these data and with this method. They do not show that slow waves do not travel: the clips were not gated to detected waves, sleep stage was unknown, and inter-site coupling was near the noise level. A gradient present only in large detected slow waves could have been missed.

stereo-electroencephalography · sleep · slow waves · latency · high-frequency activity · pediatric epilepsy · OpenNeuro · null result