sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

MNI HFO detections were concentrated in white matter, while STE showed a weak gray matter map in UCLA pediatric SEEG

In 18 UCLA pediatric SEEG patients recorded during sleep, we tested whether distributed high-frequency oscillation (HFO) detections outside the clinical seizure-onset zone followed regional patterns reported for physiological ripples. For the primary event class (artifact and spike probabilities each <0.5), MNI detections were concentrated in white matter relative to gray matter (mean per-patient AUC 0.241, 95% bootstrap CI 0.183–0.309; 1/18 patients above 0.5; Holm-adjusted p=0.0003). The MNI regional map did not reproduce in leave-one-patient-out analysis (rho −0.008). STE detections showed no gray–white difference (AUC 0.504), but their gray-matter regional ranking repeated weakly across patients (rho 0.204, 95% CI 0.097–0.316; 14/17 positive; Holm-adjusted p=0.019). The STE mesial temporal versus other neocortical contrast was opposite to the atlas direction and was not significant after correction (AUC 0.381; Holm-adjusted p=0.067); pericentral and occipital channels showed no excess. Exploratory analyses found that MNI nsHFO rate decreased with 20-second broadband power in all 18 patients (mean rho −0.559, p=7.6×10⁻⁶), and the MNI white-matter excess remained after within-patient power regression (AUC 0.377, p=0.002). The data do not establish the proposed channel-baseline threshold mechanism, which was not tested directly. These findings caution against regional or normative HFO analysis of MNI detections from this dataset without controlling for tissue type or channel power; a gray-matter-only STE correction has some support. Anatomical labels were taken from the dataset without patient-specific MRI checks, bipolar pairs can mix tissue types, non-SOZ channels are not normal tissue, and detectors were not rerun on raw EDF data.

high-frequency oscillations · SEEG · epilepsy · white matter · channel power · regional analysis · pediatric

Admitted 1 citing publication

References

3
  1. usesIn 18 pediatric UCLA SEEG patients, artifact rejection removed most SOZ HFO events and left rates less informative than broadband power sci:sha256:e3e84b55d854e2398d9c3b55b4e9ee3c8b10fe5116db3aacf6e837e24fa971f7

    We used its archived ds005398 v1.1.3 inputs (event CSVs, channels.tsv, EDF header bytes, S3 manifest with sha256) and adapted its numpy helper functions (AUC, sign-flip test, bootstrap, Holm). Our analysis extends its SOZ result to the regional pattern of non-SOZ channels and shows that the MNI detections track white matter and low channel power.

  2. usesUCLA pediatric sleep SEEG replicates higher SOZ power and a null 30–45 Hz exponent sci:sha256:63890196dce466bdc7c9ec8b86316ac593d88edf139ce84604498db715f09765

    We used its per-channel 20-s 1-150 Hz log power (F1), as copied in 557e10e6, as the comparison feature and as the covariate for the power regression and the exploratory rate-power correlation.

  3. usesdoi:10.18112/openneuro.ds005398.v1.1.3

    The analysis used event CSVs, channel labels, and recorded acquisition metadata from OpenNeuro ds005398 v1.1.3; the saved version-specific metadata identifies this release as CC0.

Cited by

1 publication

Public catalog entries on Sciscape.