Mean delay theta coupling was higher for the same trial than for other trials within session blocks in 12 patients from a release with SEEG labels

We tested delay theta coupling in OpenNeuro ds006136 version 1.0.1. Mean squared imaginary coherency at 4 to 8 Hz was 0.078966 for signals from the same trial. The mean was 0.073915 for signals from other trials within five session blocks. The patient mean excess was 0.005052. The 95 percent patient bootstrap interval was 0.002003 to 0.008637. The two-sided patient sign flip probability was 0.003418, assuming independent patients and effect distributions symmetric about zero under the null. Ten of 12 patients had a positive excess. The test used 421 channel pairs and 2,030 pair blocks. Six fixed descriptive checks had positive group means. The mean excess was smaller with three segments without overlap. The control is the exact mean of all other trial matches within a block. All 14 public EDF files and 103 metadata files matched fixed hashes. Rebuilt Fourier inputs matched exactly. The package preserves code, input arrays, trial scores, inference samples, validation, and reproduction instructions. The release labels its signals as SEEG, but depth contact status cannot be verified for every signal. The result applies to processed accepted clinical signals. It does not establish direct neural communication or a memory-specific process.

theta coupling · imaginary coherency · trial control · stereo-electroencephalography · working memory · OpenNeuro