In 61 SEEG subjects, 2–10 Hz hippocampal PSI showed no consistent posterior-to-anterior lead during word encoding

Question: Zhang and Jacobs (2015) reported that human hippocampal theta (2-10 Hz) moves as a traveling wave from posterior to anterior. We tested this claim in a large public SEEG data set. Data: OpenNeuro ds004789 (RAM FR1, session 0, bipolar EDF). We used 200 s of word-encoding signal for each subject (20 s from the first word of each of the first 10 lists). We selected 62 subjects before analysis with MNI coordinates only. Method: we computed the complex cross-spectrum for each pair of hippocampal bipolar channels in the same hemisphere without a shared contact. We oriented each pair from posterior to anterior (MNI y). The primary measure was the phase slope index (PSI) in 2-10 Hz. PSI does not change when a bipolar polarity flips, and it ignores zero-lag volume conduction. The unit of inference was the subject. We used a two-sided sign-flip test and Holm correction. The plan was written before the analysis. Results: 61 subjects and 927 pairs had an anterior-posterior separation of 5 mm or more. The subject mean PSI was 0.0016 (95% CI -0.013 to 0.016; 52% of subjects positive; p = 0.83, Holm p = 0.93). The within-subject correlation between separation and delay was not significant (n = 53, mean Fisher z = -0.05, p = 0.46). The pairs were coupled (mean coherence 0.28, against 0.07 for a shifted control). The group delay was -2.4 ms (95% CI -8.3 to 3.5 ms) at a median separation of 13 mm. A polarity-invariant phase lead was -0.7 deg (CI -4.6 to 3.3 deg). Limits: bipolar signals, MNI y is an approximation of the long axis, no oscillation detection, and subject means can cancel waves that change direction. The package contains code, cross-spectra, byte-range provenance with SHA-256, and results.

hippocampus · theta oscillations · SEEG · phase slope index · traveling waves · free recall · word encoding