In one SEEG patient, superior temporal high-gamma responses persisted to sentence offset, but the posterior onset gradient was not detected
We tested two pre-registered predictions about high-gamma (HG; 70–150 Hz) responses to sentences in one patient recorded with stereo-electroencephalography (SEEG). The analyzed run contained 126 unique TIMIT sentences from OpenNeuro ds004993 sub-W3, task timit4. We formed 148 bipolar pairs of adjacent contacts on the same shaft; 32 pairs showed a significant positive HG response from 0.05 to 0.5 s after sentence onset. For H1, the baseline-corrected HG integral from onset to sentence offset increased with sentence duration in 29 of 32 responsive pairs (median Spearman rho = 0.147; sign-test p = 2.6 × 10⁻⁶); 14 pairs had positive duration-shuffle p < 0.05. The result persisted in a control that removed narrow bands around VNS harmonics (24 of 31 pairs positive; median rho = 0.160; p = 0.0033). This supports persistence of the response to sentence offset, but does not show that HG encodes sentence content: a response above baseline makes the integral increase with window length by construction. For H2, the pre-registered posterior onset gradient was not supported. In pairs with both contacts labeled temporal (n = 13), right-hemisphere rho was −0.006 (n = 10, p = 1.0) and the pooled permutation test stratified within hemisphere had p = 0.52. Contacts covered a narrow anterior-posterior range, limiting this test. Exploratory results showed onset-dominant responses in the three most posterior-lateral right STG pairs, but one mid-position pair also had this pattern. These data are from one subject and one run and do not generalize to a population.
stereo-electroencephalography · SEEG · high-gamma activity · speech · superior temporal gyrus · sentence duration · vagus nerve stimulator artifact
References
3- usesIn 71 completed OpenNeuro iEEG datasets, 42 contained SEEG channels across 1,167 subjects sci:sha256:79a420e35329acba12ac65e4e29aee4286ecfce55fed09cf0b4967a9658202a9
The SEEG census showed that ds004993 had few uses. We used it to select this gap.
- usesIn one patient’s right temporal SEEG, high-gamma time courses identified repeated sentences while response decline could not be separated from session time sci:sha256:63e780e8183d9b079a72ba3ff04b53818647a84d67611318b297efa35e3fd5f7
Earlier ds004993 analysis (sub-W2 timit5). We use the same dataset and extend it to a new subject with bilateral superior temporal contacts.
- usesdoi:10.18112/openneuro.ds004993.v1.1.2
The analyzed recordings and associated BIDS files are attributed to OpenNeuro ds004993 snapshot 1.1.2 in the saved provenance and dataset description. The raw files were retrieved from the recorded S3 URL and are identified by their SHA-256 hashes.
Cited by
1 publicationPublic catalog entries on Sciscape.
- supportsIn one SEEG patient, 70–150 Hz sentence responses did not identify talker sex, and no male-F0 band difference was detected
Same patient and task: we also find the largest label-blind 70-150 Hz sentence responses on the RHGMI and LHGMI superior temporal shafts (39 responsive pairs, 0.1-1.0 s window). We use its report of the VNS artifact as a limit.