Ketamine has distinct electrophysiological and behavioral effects in depressed and healthy subjects

Major depressive disorder (MDD) is commonly resistant to first-line treatments and its neurobiology, as well as the mechanisms of antidepressant drugs, remain incompletely understood. Prior clinical and preclinical work indicates that the NMDA receptor antagonist ketamine produces rapid antidepressant effects in treatment-resistant depression and that enhanced AMPA receptor throughput and synaptic potentiation are important to its action. Acute ketamine administration—and its active metabolite (2R,6R)-hydroxynorketamine—has been associated with increases in gamma-frequency oscillations, a signal thought to reflect the balance of excitation and inhibition in neuronal circuits because gamma depends on both AMPA-mediated excitation and GABAA-mediated inhibition. Disruptions of this inhibition/excitation homeostasis have been implicated in depressive phenotypes in animals and humans, motivating the use of gamma power as a surrogate marker of synaptic homeostasis in clinical investigations of ketamine. This report, drawn from the Ketamine Mechanism of Action (Ket-MOA) study, set out to examine ketamine’s effects on mood and resting-state gamma power in unmedicated, treatment-resistant inpatients with MDD as well as in psychiatrically healthy controls. Nugent and colleagues hypothesised that ketamine would produce sustained increases in gamma power that would correspond to antidepressant improvement. The study therefore measured clinical outcomes and magnetoencephalography (MEG) gamma power at baseline, approximately six to nine hours after infusion, and 11–13 days after infusion to probe both clinical and electrophysiological effects beyond the acute infusion period.

The investigators used a double-blind, placebo-controlled, randomized crossover design in which each participant received ketamine and placebo infusions two weeks apart; infusion order was randomised. Clinical ratings were obtained before and repeatedly after each infusion (at -60 minutes baseline and at multiple time points up to Day 11), and resting-state MEG recordings were acquired at three sessions: baseline (two to four days before the first infusion), the day of infusion approximately six to nine hours post-infusion, and 11–13 days post-infusion. The primary clinical outcome was the Montgomery–Åsberg Depression Rating Scale (MADRS). The extracted Methods text does not clearly report the ketamine dose used in the infusions. Eligibility criteria for the MDD group required DSM-IV recurrent MDD without psychotic features confirmed by structured interview, a MADRS score ≥20 at screening and before each infusion, failure to respond to at least one adequate antidepressant during the current episode, and being free of psychotropic medications for specified washout intervals. Healthy controls had no Axis I disorder and no family history of Axis I disorders in first-degree relatives, and were free of medications affecting neuronal function. The investigators calculated that a patient sample size of 34 would provide 80% power to detect an antidepressant effect of ketamine of d=0.5 at p<0.05, two-tailed. Clinical outcome analyses used linear mixed models (restricted maximum likelihood) with time and drug as within-subject factors, including the phase-specific baseline as a covariate; participants were included if at least one pre- and one postinfusion measure were available for a phase. Post-hoc contrasts used Bonferroni adjustment for primary outcomes. A broad set of secondary and exploratory scales were administered, including HAM-D17, SHAPS, TEPS, HAM-A, BDI, and dissociation measures (CADSS). MEG data were recorded on a 275-channel CTF system at 1200 Hz. Preprocessing included a 2 Hz high-pass filter, visual inspection and exclusion of artifact-contaminated segments, and selection of up to ten 15-second artifact-free epochs (datasets with fewer than five such segments were discarded). Source localisation used synthetic aperture magnetometry (SAM) beamforming on a 5 mm grid with a multisphere head model co-registered to individual T1 MRI scans. Gamma-band root-mean-square (RMS) power images were normalised relative to the projected noise floor and then divided by the square root of the summed squared images across canonical bands (2–100 Hz). Group-level imaging analyses employed linear mixed models implemented in AFNI’s 3dLME, modelling session, diagnosis, and pre/post-task factors, with age and gender included initially. Because regions of interest (ROIs) were functionally defined from contrasts, the authors note that unbiased effect-size estimates from ROI values are not interpretable. Secondary analyses examined relationships between regional gamma power and clinical response by adding absolute MADRS change as a covariate; exploratory moderation analyses tested whether baseline gamma moderated the association between post-ketamine gamma and MADRS response, with both absolute and percent-change response metrics examined. The exploratory MEG moderation analyses were based on a reduced MDD subsample for whom both baseline and post-ketamine data were available (n=17).

The analysed sample comprised 35 unmedicated, treatment-resistant MDD subjects and 26 healthy control subjects; one healthy control had a baseline MEG but was not randomised. Clinical outcomes in the MDD group showed a robust antidepressant effect: MADRS scores were significantly lower following ketamine versus placebo (F 1,77 = 84.5, p<0.001). Dissociative symptoms measured by CADSS increased acutely with ketamine (significant main effects of drug, time, and drug-by-time interaction), peaking at 40 minutes post-infusion. Ketamine produced improvements across multiple symptom domains in MDD, including measures of anhedonia, anxiety, PTSD symptoms, suicidality, and quality of life, as reported in supplementary materials. Contrary to expectations, healthy control participants displayed a transient worsening of depressive symptoms after ketamine. There were significant main effects of drug and time and a drug-by-time interaction for MADRS in controls (F 1,328 = 61.87, p<0.001). The increase in MADRS was significant at 40, 80, and 120 minutes post-infusion and at Day 1; by Day 2 only one control still scored above the threshold. Seventeen of 24 healthy controls receiving ketamine (71%) showed an increase of at least five MADRS points at any time point versus 1 of 23 receiving placebo (4%). Symptom domains most affected in controls were inner tension, lassitude, inability to feel (anhedonia), psychic anxiety, and somatic anxiety. CADSS scores also increased transiently in controls, peaking at 40 minutes, and CADSS at 40 minutes was only weakly correlated with MADRS (R=0.358, p=0.086), implying that dissociation did not fully account for the transient mood worsening. MEG analyses revealed increased resting-state gamma power following ketamine relative to placebo in both MDD and healthy groups approximately six to nine hours post-infusion. These increases were observed in widespread cortical and subcortical regions implicated in the central executive, salience, and default mode networks. Primary analyses found no simple association between regional gamma power and magnitude of antidepressant response in either group when absolute MADRS change was modelled. In exploratory moderation analyses restricted to ROIs showing post-ketamine gamma increases, baseline gamma power significantly moderated the relationship between post-ketamine gamma and MADRS response in multiple regions (8 of 11 ROIs). The interaction was strongest in the thalamus (F 1,12 = 47.46, p<0.001) and the right insula (F 1,13 = 22.24, p<0.001): in MDD subjects with lower baseline gamma, larger post-ketamine increases in gamma were associated with better antidepressant response, whereas in those with higher baseline gamma, larger increases were associated with worse response. When response was defined as percent change at later time points, the moderating effect attenuated to trend levels in some regions. The authors emphasise that the exploratory MEG moderation findings derive from a small subsample (n=17) and that ROIs were functionally defined so effect sizes cannot be straightforwardly interpreted.

Nugent and colleagues interpret their results as confirming a rapid and relatively sustained antidepressant effect of a single ketamine infusion in treatment-resistant MDD, accompanied by increased resting gamma power several hours post-infusion. They also highlight the unexpected but reproducible finding that ketamine transiently increased depressive symptoms in psychiatrically healthy controls, primarily in domains of anxiety, emotional blunting and anhedonia; these mood changes were only weakly related to dissociative side effects, suggesting a separable mood-lowering effect in healthy volunteers. The authors note that gamma power remained elevated six to nine hours after ketamine in both groups, indicating that ketamine’s influence on synaptic plasticity persists beyond the acute infusion. However, the absence of a simple correlation between post-ketamine gamma power and antidepressant response led them to probe heterogeneity within the MDD sample. Exploratory analyses suggested that baseline gamma moderates the relationship between increases in gamma and clinical improvement: increases in gamma appear beneficial when baseline gamma is low but potentially deleterious when baseline gamma is already high. This pattern is consistent with a homeostatic model in which both deficits and excesses of inhibition/excitation balance can be maladaptive, and where restoring an optimal balance may be necessary for antidepressant effect. Several limitations are acknowledged. The moderation findings are post-hoc and based on a small MEG subsample (17 MDD subjects), preventing use of the results as a validated biomarker; functionally defined ROIs preclude unbiased effect-size estimation. The authors also note that there is no clear single ‘‘ideal’’ gamma value applicable across subjects given high inter-subject variability, and that percent-change outcomes at later time points reduced the strength of some associations. Methodological constraints—such as MEG sensitivity differences for deep versus surface sources—are discussed, alongside the need for longitudinal studies with multiple post-infusion time points to characterise the temporal dynamics of gamma, connectivity, and glutamatergic markers. Finally, the investigators emphasise two pragmatic implications derived from their data: first, findings obtained in healthy volunteers may not generalise to clinical populations and thus should be interpreted with caution when used to infer antidepressant mechanisms; second, gamma power may be a candidate marker of synaptic homeostasis that, with further validation, could help stratify biological subtypes of MDD and guide personalised interventions. They recommend hypothesis-driven follow-up studies to test whether baseline electrophysiological measures can prospectively predict who will benefit from interventions that modulate inhibition/excitation balance.