High-order brain interactions in ketamine during rest and task: a double-blinded cross-over design using portable EEG on male participants

Abstract

Ketamine is a dissociative anesthetic that induces a shift in global consciousness states and related brain dynamics. Portable low-density EEG systems could be used to monitor these effects. However, previous evidence is almost null and lacks adequate methods to address global dynamics with a small number of electrodes. This study delves into brain high-order interactions (HOI) to explore the effects of ketamine using portable EEG. In a double-blinded cross-over design, 30 male adults (mean age = 25.57, SD = 3.74) were administered racemic ketamine and compared against saline infusion as a control. Both task-driven (auditory oddball paradigm) and resting-state EEG were recorded. HOI were computed using advanced multivariate information theory tools, allowing us to quantify nonlinear statistical dependencies between all possible electrode combinations. Ketamine induced an increase in redundancy in brain dynamics (copies of the same information that can be retrieved from 3 or more electrodes), most significantly in the alpha frequency band. Redundancy was more evident during resting state, associated with a shift in conscious states towards more dissociative tendencies. Furthermore, in the task-driven context (auditory oddball), the impact of ketamine on redundancy was more significant for predictable (standard stimuli) compared to deviant ones. Finally, associations were observed between ketamine's HOI and experiences of derealization. Ketamine appears to increase redundancy and HOI across psychometric measures, suggesting these effects are correlated with alterations in consciousness towards dissociation. In comparisons with event-related potential (ERP) or standard functional connectivity metrics, HOI represent an innovative method to combine all signal spatial interactions obtained from low-density dry EEG in drug interventions, as it is the only approach that exploits all possible combinations between electrodes. This research emphasizes the potential of complexity measures coupled with portable EEG devices in monitoring shifts in consciousness, especially when paired with low-density configurations, paving the way for better understanding and monitoring of pharmacological-induced changes.

Fig. 1. Overview of experimental design and data analysis.

A Subjects participated in a double-blind crossover design using portable EEG, capturing both resting states and task-based recordings (namely, a gamified oddball paradigm inducing a typical mismatch negativity). In randomized sessions, participants received both ketamine and saline infusions. B Analysis of high-order interactions (HOI) entailed measurements of total correlation (TC) and dual total correlation (DTC), O-information and S-information (see section “Methods”). C Feature selection across the two designs (rest and task) was employed to pinpoint the primary differences between the ketamine and saline conditions.

Fig. 2. Ketamine increases redundancy in the resting state alpha band.

A Each panel shows the effect sizes of each n-plet (i.e. each possible combination of electrodes from 2 to 16) sorted in decreasing order, with a red horizontal line showing the 0. The red dot denotes the maximum effect size. B The n-plet with the largest effect under saline and ketamine for each subject (gray) and for the average (black). Only O and S-info yielded FDR-corrected Wilcoxon sign rank p-values < 0.001. C EEG layout with the electrodes composing the n-plet in black.

Fig. 3. Redundancy increases in the standard tone for mismatch negativity task.

A Effect sizes of each n-plet for the oddball paradigm, sorted in decreasing order, with a red horizontal line showing the 0. The red dot denotes the maximum effect size B The n-plet with the largest effect under saline and ketamine for each subject (gray) and for the average (black). C The electrodes involved in the best feature. Wilcoxon sign rank FDR-corrected p-value < 0.005.

Fig. 4. Association between subjective scores and HOI.

A Average R2 only for the significant (FDR, p < 0.005) associations between the change in each dimension of subjective score (see section “Methods”) and change in HOI. B Same as A, but showing the number of significant HOI in log10 scale to improve visualization.