Ketamine induces EEG oscillations that may aid anesthetic state but not dissociation monitoring

Ketamine anesthesia produces structured changes in brain electrical activity that track anesthetic state but not dissociation. In an open-label study of 15 participants given 2 mg/kg ketamine, unresponsiveness was linked to increased frontal brainwave power across several frequency bands, while a responsive but dissociated state showed decreased power across broad ranges. The benzodiazepine midazolam reduced dissociation scores and altered power in specific bands. A mixed-effects model found that only midazolam, not any electroencephalogram feature, predicted dissociation scores. The authors conclude that electroencephalogram signatures may enable principled monitoring of anesthetic state but not dissociation.

Study at a glance

Abstract

Objective: Ketamine is an anesthetic drug associated with dissociation. Decreased electroencephalogram alpha (8 to 13 Hz) and low-beta (13 to 20 Hz) oscillation power have been associated with ketamine-induced dissociation. We aimed to characterize surface electroencephalogram signatures that may serve as biomarkers for dissociation. Methods: We analyzed data from a single-site, open-label, high-density surface electroencephalogram study of ketamine anesthesia (2mg/kg, n = 15). We assessed dissociation longitudinally using the Clinician Administered Dissociation States Scale (CADSS) and administered midazolam to attenuate dissociation and enable causal inference. We analyzed spectral power and global coherence with multitaper spectral methods. Mixed effects models were used to assess whether electroencephalogram power and global coherence signatures of ketamine could be developed into dissociation-specific biomarkers. Results: Compared to baseline, ketamine unresponsiveness was associated with increased frontal power between 0.5 to 9.3 Hz, 12.2 to 16.6 Hz, and 24.4 to 50 Hz. As subjects transitioned into a responsive but dissociated state (mean CADSS ± SD, 22.1 ± 17), there was a decrease in power between 0.5 to 10.3 Hz and 11.7 to 50 Hz. Midazolam reduced dissociation scores (14.3 ± 11.6) and decreased power between 4.4 to 11.7 Hz and an increase in power between 14.2 to 50 Hz. Our mixed-effects model demonstrated a quadratic relationship between time and CADSS scores. When models (frontal power, occipital power, global coherence) were reanalyzed with midazolam and electroencephalogram features as covariates, only midazolam was retained. Conclusions: Ketamine is associated with structured electroencephalogram power and global coherence signatures that may enable principled anesthetic state but not dissociation monitoring. Significance: A neurophysiological biomarker for dissociation may lead to a better understanding of neuropsychiatric disorders.