Permutation entropy analysis of EEG during a Go/No-Go task to differentiate patients with suicide attempt from those with suicidal ideation in major depressive disorder
摘要
Objective neurophysiological markers are needed to complement clinical assessment in distinguishing patients with suicide attempt (SA) from those with suicidal ideation (SI) within major depressive disorder (MDD). Task-evoked electroencephalography (EEG) complexity was quantified during an auditory Go/No-Go task in 126 right-handed, psychotropic-medication-free participants with MDD (SA: n = 63, 23 males and 40 females, mean age 41.43 ± 14.93 years; SI: n = 63, 25 males and 38 females, mean age 39.13 ± 14.52 years). Permutation entropy (PE) was computed in sliding stimulus-locked windows at midline electrodes (Fz, FCz, and Cz). Normalized PE (H, a 0–1 ordinal-pattern complexity index) was estimated for Go and No-Go trials across embedding delays (τ = 1–60). Primary analyses focused on No-Go H; additionally, a Go/No-Go asymmetry index (A)—calculated as the normalized difference between Go and No-Go H—was evaluated. Group differences were assessed using cluster-based permutation tests, with Bonferroni correction across complexity metrics. Task accuracy and reaction time were comparable between groups. No-Go H was reduced in the SA group at Fz during 100–190 ms post-stimulus across τ = 26–40 (Bonferroni-adjusted cluster p = 0.0497). Based on the Nyquist-derived τ-to-frequency mapping, this range corresponds to frequencies in the low-beta band (12.5–19.2 Hz). However, clusters for A at Fz (100–190 ms, τ = 27–38) and Cz (50–140 ms, τ = 14–25) did not remain significant after correction. The findings indicate that an early frontocentral reduction in EEG complexity during response inhibition may serve as a candidate task-evoked neurophysiological feature that differentiates patients with SA from those with SI within MDD.