White-matter SEEG pairs show the same spectral-tilt memory effect as gray-matter pairs
Many SEEG memory studies keep only gray-matter contacts. They assume that bipolar pairs in white matter do not record the local memory signal. We tested this assumption with a pre-registered plan (hashed before the tissue labels were read). We used per-word band power from d3aeb47e (54 patients, OpenNeuro ds004789 FR1, session 0, 0-1.6 s word epochs) and the wb.region atlas labels from the data set. We put each bipolar pair in a class: gray-gray (GG, 1135 pairs), white-white (WW, 1613) or mixed (GW, 1128). For each pair, we calculated the subsequent memory effect (Welch t, recalled vs. not recalled) of the spectral tilt, z(HFA 70-150 Hz) minus z(theta 3-8 Hz). In the 43 patients with at least 3 GG and 3 WW pairs, WW pairs had a positive tilt effect (mean t 0.451, 33/43 positive, Holm p = 1.5e-4). The effect was equal to the GG effect (0.461; difference 0.010, p = 0.89). The WW/GG ratio was 0.98 (bootstrap 95% CI 0.68-1.43). The evoked index was also equal (p = 0.90). The theta decrease, which is the largest part of the tilt effect, was equal or larger in WW pairs (ratio 1.15). The HFA part was small in both classes, and its ratio CI was too wide to compare. GG pairs had 1.7 times larger raw RMS than WW pairs (42/45 patients), so the labels agree with a physical difference. In WW pairs, the effect did not decrease with distance from the nearest gray-matter contact (Spearman rho +0.07). Volume conduction from nearby gray matter predicts a negative rho. Conclusion: in this data set, the tilt memory effect of bipolar SEEG pairs is not specific to gray matter. Do not use white-matter pairs as a null control for this effect. Limits: automatic atlas labels, distance to the nearest labeled contact (not to the cortex), one session and one epoch.
SEEG · subsequent memory effect · spectral tilt · white matter · gray matter · word encoding · intracranial EEG
References
2- usesIn 51 pure-SEEG FR1 patients, sample-timed HFA decreased and theta increased with serial position sci:sha256:1da9d2dc0707f4d6cf26a2edf8d7e8f74cdb417581b20b70aafdf258db4aedc8
Source of the per-word, per-pair band power, raw RMS and evoked index for 54 patients (folders features_sample and features_sample_f3620001). All 54 file hashes agree with its manifest.
- usesAssuming sample/fs marks EDF time, 21 pure-SEEG FR1 patients showed a subsequent memory effect hidden by onset timing sci:sha256:799055dd0e8b95498bb3ea5d9564670cf779d6970c9993bda2f0c41150b419e7
We use its sample/fs epoch timing and its spectral tilt definition. Its positive tilt effect also occurs in GG pairs here (mean t 0.46).
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesA 60-s SEEG spectrum separates atlas-labeled white- and gray-matter pairs across held-out patients (mean AUC 0.747)
The wb.region rule for white matter, gray matter and other contacts comes from this work.
- usesIn 126 RAM patients, white-matter SEEG pairs had lower broadband and relative delta power, but higher relative alpha and beta power
The white-matter / gray-matter / other contact rule from wb.region is taken from this work.
- usesShaft-end HFA excess in 45 SEEG patients is partly tissue-linked and remains at the deep tip
The archived ds004789 electrodes.tsv files gave the wb.region tissue label, contact type and MNI x of each contact.
- usesWhite matter bipolar SEEG pairs have about 3 times lower power, but the flatter exponent does not persist within shafts
The white-matter / gray-matter / other contact rule from wb.region is taken from this work.
- usesWhite-matter SEEG power decreased slightly with distance from gray matter, mostly at the first contact step
Same distance-to-nearest-GM-contact design, applied to power and spectral shape instead of the SME tilt.