Sustained pharmacodynamic effects of S‐ketamine on cortical excitability and resting‐state brain activity: A randomized, placebo‐controlled trial

Abstract

Aims

This study aimed to investigate a set of pharmacodynamic biomarkers reflecting acute, delayed and sustained central nervous system effects of S‐ketamine, used as a tool compound for rapid‐acting antidepressant activity, with the goal of informing biomarker strategies for delayed antidepressant effects.

Methods

In this randomized, double‐blind, double‐dummy, placebo‐controlled, 4‐way crossover study in 16 healthy participants, we administered S‐ketamine in a therapeutic intravenous dose (IV), a low and high oral dose, versus placebo. Measurements were conducted from baseline up to 7 days post‐dose using Transcranial Magnetic Stimulation combined with electromyography (TMS‐EMG) and encephalography (TMS‐EEG), pharmaco‐electroencephalography (pEEG), alongside drug and metabolite plasma concentrations. Outcomes were analysed using mixed‐effects ANCOVA and cluster‐based permutation testing. Post‐hoc concentration–effect relationships were explored.

Results

IV S‐ketamine induced an acute reduction in motor‐evoked potential (MEP) amplitude and sustained attenuation of long‐interval intracortical inhibition (LICI₅₀), the latter showing a linear concentration–effect relationship with the parent compound. Acute TEP modulation was observed across all treatments, whereas delayed effects occurred only after IV and high‐dose oral. pEEG showed acute reductions in alpha, beta and delta power (eyes closed) and sustained increases in delta power (eyes open) following IV and high‐dose oral S‐ketamine; with delta power exhibiting a less analyte‐specific linear concentration–effect relationship.

Conclusion

We provide evidence suggestive of delayed pharmacodynamic effects of S‐ketamine in healthy participants sustained up to 7 days post‐dose, using TMS and pEEG derived measures, which are markedly distinct from its acute effects and may be relevant to understand its antidepressant efficacy.

What is already known about this subject?

• S‐ketamine has emerged as a rapid‐acting antidepressant, yet the underlying mechanisms mediating S‐ketamine's antidepressant effects in humans remain unclear.

• Efforts to develop alternative rapid‐acting antidepressant drugs are currently limited by a lack of pharmacological biomarkers.

What this study adds

• This study is the first to demonstrate the delayed pharmacodynamic effects of S‐ketamine beyond 24 h post‐dose on cortical excitability and spontaneous brain activity in healthy volunteers.

• The delayed and sustained effects are distinct from the acute effects in terms of their directionality, showing an increase in cortical excitability and spontaneous brain activity, rather than the decrease observed during the acute phase.

• TMS and pEEG are potential biomarker candidates for S‐ketamine's sustained antidepressant efficacy and offer mechanistic insights that may guide future development of rapid‐acting antidepressants.

1. INTRODUCTION

Major depressive disorder (MDD) affects approximately 280 million people worldwide and remains a leading cause of disability. ¹ Conventional monoaminergic antidepressants have limitations in terms of delayed onset, limited efficacy and poor tolerability, that underscore the urgent need for rapidly acting pharmacological treatments with novel mechanisms to effectively address treatment‐resistant depression (TRD). ²

S‐ketamine (SKET) is a non‐competitive, non‐selective, glutamate N‐methyl‐D‐aspartate receptor (NMDAR) antagonist, originally approved as an anaesthetic, but recently FDA‐approved as monotherapy for TRD in an intranasal formulation (Spravato®). SKET produces a rapid onset of antidepressant effects; however, its broader clinical use in psychiatry is limited by transient sedation and psychomimetic adverse effects. ³ Accordingly, the development of novel compounds that preserve the antidepressant efficacy associated with SKET while minimizing these untoward effects may offer safer and more scalable rapid‐acting antidepressant strategies.

However, the mechanisms underlying SKET's antidepressant effects remain incompletely understood, and efforts to develop alternative rapid‐acting antidepressant drugs (RAADs) are constrained by a lack of pharmacodynamic (PD) biomarkers. A biomarker platform capable of quantifying both therapeutic and central nervous system (CNS) mediated side effects in healthy volunteers (HVs) would enable efficient early‐phase screening of candidate compounds and dose selection. Importantly, optimizing such a platform requires a deeper understanding of SKET's underlying pharmacological effects and their temporal evolution.

Clinically, SKET's antidepressant effects typically emerge by 24 h (h) post‐dose, ⁴ , ⁵ following its rapid systemic clearance, and persist for several weeks. ³ These delayed effects are distinct from the acute concentration‐dependent psychomimetic and CNS‐depressant effects observed shortly after dosing. ⁶ Results from a concurrently conducted study, ⁷ comparing IV with orally administered SKET (PO‐SKET) using a validated CNS test battery (NeuroCart), further support the hypothesis that SKET's acute effects up to 6 h post‐dose are primarily mediated by the parent compound through (direct) NMDAR antagonism. ⁸ PO‐SKET was associated with a significantly altered PD profile, consistent with differences in systemic exposure and metabolite‐to‐parent ratios (MPRs) resulting from first‐pass metabolism. ⁹ Importantly, these differences were primarily driven by lower parent‐compound exposure rather than increased exposure to the active metabolite S‐norketamine (SNOR). ⁷ The same study further demonstrated the sensitivity of biomarkers such as saccadic peak velocity, adaptive tracking, body sway and subjective drug effects for SKET's acute NMDAR‐mediated adverse effects.

In contrast, SKET's sustained antidepressant efficacy may depend on glutamatergic modulation through mechanisms extending beyond initial NMDAR target engagement. These downstream processes potentially involve synaptic plasticity and network‐level reorganization, ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ for which PD biomarkers are currently lacking. In this context, Transcranial Magnetic Stimulation (TMS) and resting‐state pharmaco‐electroencephalography (pEEG) have emerged as potential methodologies for probing such longer‐lasting neurophysiological changes. Previous studies have demonstrated decreased excitability ¹⁶ , ¹⁷ , ¹⁸ and a shift in spectral power to the lower frequencies in patients with TRD, ¹⁹ , ²⁰ suggesting potential relevance of these biomarkers to the disease pathology, and thus possibly to antidepressant efficacy. Although it remains unclear whether these measures can serve as translational biomarkers of delayed pharmacodynamic responses and subsequent clinical benefit, TMS and pEEG biomarkers were selected based on the a priori hypothesis that delayed antidepressant response to SKET is associated with sustained modulation of cortical excitability and oscillatory dynamics.

To date, TMS and pEEG have been applied predominantly to characterize the acute effects of ketamine, ²¹ , ²² , ²³ but not its delayed effects. Nevertheless, TMS is of particular interest because of its demonstrated sensitivity to pharmacological modulation of glutamatergic neurotransmission though through different mechanisms. For example, TMS has demonstrated sensitivity to other NMDAR antagonists, such as dextromethorphan and memantine, in healthy participants. ²⁴ In addition, TMS has been shown to detect PD effects of the AMPA receptor positive allosteric modulator osavampator (TAK‐653/NBI‐1065845), a candidate RAAD, which also produced a psychostimulant‐like profile on the NeuroCart. ²⁵ , ²⁶ Together, these findings support the potential of TMS‐ and pEEG‐derived measures as mechanistic biomarkers to investigate delayed glutamatergic effects relevant to antidepressant response.

Building on prior work by van Mechelen et al. characterizing the acute PD effects of SKET using the NeuroCart, ⁷ the present study aims to evaluate TMS and pEEG as candidate biomarkers of corticospinal excitability, cortical excitability and resting‐state brain, respectively, for not only acute (≤6 h), but also delayed (6–24 h) and sustained (24 h–7 days) CNS effects of SKET in healthy participants. Furthermore, by leveraging the pharmacokinetic divergence between IV‐ and PO‐SKET, we explore whether differences in parent‐compound and metabolite exposure relate to differences in the timing and nature of SKET's delayed PD effects.

2. METHODS

This study was approved by the Medical Ethics Committee, Foundation Beoordeling Ethiek Biomedisch Onderzoek (BEBO; ID: NL73916.056.20). The study was conducted in accordance with the principles of the Declaration of Helsinki and the International Conference on Harmonization guidelines on Good Clinical Practice. Written informed consent was obtained from all participants before study participation. Centre for Human Drug Research (CHDR), Leiden, The Netherlands sponsored and conducted this study, from 12 Aug 2021 to 10 Mar 2022, registered in the Dutch Trial Registry (NTR, ID: 28424).

2.1. Participants

Male and female HVs, age 18–45 years, were selected following a medical screening (in‐ and exclusion criteria listed in Table S1). Participants with an increased risk of TMS‐related complications based on the TMS safety questionnaire, ²⁷ or significant history or presence of neuropsychiatric illnesses were excluded.

2.2. Study design

The study was a randomized, double‐blind, double‐dummy, placebo‐controlled, four‐way crossover study (Figure S1). Participants were administered (1) single low‐dose solution of 0.20 mg/kg PO‐SKET + 40‐minute placebo infusion, (2) single high‐dose solution of 0.45 mg/kg PO‐SKET + 40‐min placebo infusion, (3) 0.40 mg/kg 40‐min infusion of SKET (IV‐SKET) + placebo PO solution and (4) 40‐min placebo infusion + placebo PO solution, in randomized order. The randomization code was generated digitally by a study‐independent unblinded statistician, participants were enrolled sequentially by the research physician and the medication was prepared by a study‐independent unblinded pharmacist. Replacement participants received the same treatment allocation as the participant being replaced. Participants were admitted to the research centre each study period on Day −1 to Day 2, with a wash‐out period of 14–21 days, and a return visit on Day 7. Details on safety and PK assessments are described elsewhere. ⁷ PK samples were collected to measure SKET, SNOR and S‐hydroxy‐norketamine (SHNK) plasma concentrations. PEEG was performed pre‐dose, and 2.5, 6, 24 h and 7 days post‐dose. TMS‐EMG–EEG was performed pre‐dose and 4.5, 24 h and 7 days post‐dose. PD effects were assessed over different time periods following SKET administration: effects measured at 2.5, 4.5 and 6 h post‐dose were regarded as acute; those at 24 h as delayed; and those at 7 days as sustained.

2.3. Dose selection

A detailed description of the dose selection is also provided elsewhere. ⁷ In summary, it was based on simulated SKET and SNOR concentrations from two population PK‐models, developed with in‐house data from IV‐SKET administrations ²⁸ and PO‐SKET data from the literature. ⁹ Intravenous (IV) and oral (PO) administration routes were selected to maximize pharmacokinetic contrast between the parent compound and its metabolites, thereby facilitating evaluation of their differential contributions to SKET's delayed pharmacodynamic effects. An IV dose that demonstrated antidepressant efficacy in TRD ⁵ and was expected to reach exposures associated with robust PD effects on the Neurocart test battery ⁶ was selected: 0.40 mg/kg 40‐min IV. Two single PO doses were selected: 0.45 mg/kg, to target similar SNOR exposures as a multiple dose study being investigated for the treatment of TRD elsewhere ²⁹ and 0.20 mg/kg, to ensure similar maximum plasma concentrations of SNOR as the selected IV dose. In order to align the expected time to maximum concentration (T_max), PO administrations were performed 15 min after the start of IV administration.

2.4. Procedures

Per IPEG guidelines, ³⁰ pEEG was recorded with 5‐min alternating periods of eyes closed and open every 64 s (s). TMS was applied per current guidelines. ²⁷ The motor hotspot of the dominant abductor digiti minimi (ADM) muscle, assessed by the Edinburgh Handedness Questionnaire, ³¹ was stimulated. The rMT was determined ³² at the start of each measurement, after which single‐pulse (spTMS) and paired‐pulse (ppTMS) protocols were applied. TMS‐EMG and ‐EEG were recorded simultaneously during TMS stimulation.

Data acquisition and processing was performed as described previously, ³³ and a comprehensive explanation is provided in the Supporting Text S1. Endpoints for pEEG included total power per frequency band per electrode and per eye state. TMS‐EMG parameters included: mean single pulse peak‐to‐peak MEP amplitude (mV), paired‐pulse short intra‐cortical inhibition as ISIs of 2 ms (SICI₂) and long intra‐cortical inhibition at ISIs of 50, 100 and 300 ms (LICI₅₀, LICI₁₀₀ and LICI₃₀₀).

2.5. Statistical analysis

The sample size calculation was based on the MEP amplitude from Ruijs et al., ³⁴ with 16 participants providing 80% power to detect mean difference of −350 μV, assuming a SD of 450 μV and a two‐sided significance level of 0.05.

Data for pEEG was log‐transformed and back‐transformed after analysis so the results could be interpreted as percentage change. Each pEEG and TMS‐EMG parameter was analysed with a mixed model analysis of covariance with treatment, time, period and treatment by time as fixed factors and participant, participant by treatment and participant by time as random factors and the (average) baseline measurement as covariate. To evaluate differences in acute, delayed and sustained PD effects, contrasts were calculated per treatment and post‐dose timepoints within the model, e.g., low dose SKET: 0.20 mg/kg PO versus placebo ≤ 6 h for acute effects; 0.20 mg/kg PO versus placebo at 24 h for delayed effects; 0.20 mg/kg PO versus placebo at 7 days for sustained effects. pEEG assessments for 2.5 and 6 h post‐dose were combined. We report the estimated difference (ED) of the least square mean (LSM), the 95% confidence interval (95% CI) and the p value. Graphs of the change from baseline (CFB) of the LSM estimates over time by treatment are presented with 95% CI as error bars. No correction for multiple testing is applied.

TMS evoked potential (TEP) responses were analysed across all EEG electrodes, for the full sample (0–300 ms) and at specific time periods of interest (TOIs) around the characteristic TEP components ³⁵ (N15: 0–20 ms; P30: 20–40 ms; N45: 40–55 ms; P60: 55–80 ms; N100: 80–130 ms; P180: 130–230 ms). To compare TEPs measured under placebo and active treatment conditions, nonparametric cluster‐based permutation analysis was used. ³⁶ Full methodological details, including model specifications and statistical parameters, are provided in the Supporting Text S1.

Post‐hoc exploratory PK‐PD analysis was performed on parameters demonstrating significant effects in the mixed effects model: MEP amplitude, LICI₅₀ and pEEG delta power (eyes open and closed). For the PK‐PD analysis of the acute phase (up to 6.5 h post‐dose), PD observations were matched with PK data from the closest PK sampling time. For the PK‐PD analysis of PD endpoints collected at 24 h and 7 days post‐dose, the PD observations were matched with the participant's area under the concentration–time curve from time zero to the last quantifiable concentration (AUC_last) on Day 1 of that treatment period. Parameters were modelled with an intercept only, a linear concentration–effect relationship, and a nonlinear (E_max) concentration–effect relationship in a mixed‐effects model. A full description of the PK‐PD analysis is provided in Supporting Text S2.

2.6. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22. ³⁷

3. RESULTS

3.1. Demographics, PK and safety

Sixteen participants (eight males and eight females) completed the study. The median (min–max) age and Body Mass Index (BMI) were 25 (21–35) years and 21.9 (19.4–27.2) kg/m², respectively. Safety and PK results are reported elsewhere. ⁷ Plasma concentration over time curves for SKET, SNOR and SHNK are summarized in Figure 1. Treatment emergent adverse events and PK parameters are provided in Table S2 and S3, respectively.

FIGURE 1.

Plasma‐concentration (mean ± SD) over time curve. Average S‐ketamine (SKET; in red), S‐norketamine (SNOR; in green) and S‐hydroxy‐norketamine (SHNK; in blue) plasma time‐concentration profiles up to 24 h post‐dose for IV‐SKET (0.40 mg/kg S‐ketamine 40‐min IV; top panel), HD‐SKET (0.45 mg/kg S‐ketamine PO; bottom panel) and LD‐SKET (0.20 mg/kg S‐ketamine PO; middle panel). Abbreviations: h = hours; HD‐SKET = high‐dose oral S‐ketamine (0.45 mg/kg S‐ketamine PO); IV = intravenous; IV‐SKET = 0.40 mg/kg S‐ketamine 40‐minute intravenous infusion; kg = kilogram; LD‐SKET = low‐dose oral S‐ketamine (0.20 mg/kg S‐ketamine PO); mg = milligram; mL = millilitres; ng = nanograms; PO = per os; SD = standard deviation.

3.2. TMS‐EMG

Acute and sustained effects were observed for 0.40 mg/kg IV‐SKET on TMS‐EMG, with a statistically significantly decreased MEP amplitude at 4.5 h post‐dose (ED: −430.17 μV, 95% CI: −806.74; −53.61, p = 0.03) and attenuation of LICI50 7 days post‐dose (ED: 57.83%, 95% CI: 14.24; 101.43, p = 0.01) compared with placebo, respectively (Figure 2). IV‐SKET did not affect any other TMS‐EMG parameters. Neither PO‐SKET dose demonstrated acute, delayed nor sustained effects. For a complete overview of sp‐ and ppTMS‐EMG results, see Table S4.

FIGURE 2.

Summary graphs of single pulse MEP amplitude and LICI50. Change from baseline of the least square means (LSM) with 95% confidence interval (95% CI) as error bars of the (A) motor‐evoked potential (MEP) amplitude (μV) and (B) long intracortical inhibition (LICI₅₀) (%), using single‐ and paired‐pulse transcranial magnetic stimulation, for all treatments. IV‐SKET corresponds to 0.40 mg/kg S‐ketamine 40‐min IV (in blue), HD‐SKET corresponds to 0.45 mg/kg S‐ketamine PO (in pink) and LD‐SKET corresponds to 0.20 mg/kg S‐ketamine PO (in green). Significant results compared to placebo (in grey) are marked with an asterisk (*). Abbreviations: 95% CI = 95% confidence interval; BL = baseline; CFB = change from baseline; HD‐SKET = high‐dose oral S‐ketamine (0.45 mg/kg S‐ketamine PO); IV = intravenous; IV‐SKET = 0.40 mg/kg S‐ketamine 40‐minute intravenous infusion; LD‐SKET = low‐dose oral S‐ketamine (0.20 mg/kg S‐ketamine PO); LICI₅₀ = long intra‐cortical inhibition at interstimulus‐interval of 50 ms; LSM = least square means; MEP = motor‐evoked potential.

3.3. TMS‐EEG

All treatments showed statistically significant acute (4.5 h) effects, both 0.40 mg/kg IV‐SKET and 0.45 mg/kg PO‐SKET showed delayed effects (24 h), and none of the treatments showed sustained effects (7 days). Overall, a pattern of two consistent clusters was observed. First, an acute effect at 4.5 h post‐dose was observed for both 0.20 and 0.45 mg/kg PO‐SKET, with significant clusters demonstrating an increased N45 component (more negative) and decreased P60 component (less positive) located in the contralateral centro‐parietal region, measured using ppTMS‐EEG at long intervals (ISIs 50, 100 and 300 ms) (Figure 3A–D). A comparable cluster was found as delayed effect measured 24 h post‐dose for 0.45 mg/kg PO‐SKET (Figure 3E). Secondly, a delayed effect measured 24 h post‐dose was observed for 0.40 mg/kg IV‐KET and 0.45 mg/kg PO‐SKET, with significant clusters demonstrating an increased P180 component (more positive) located in the contralateral parieto‐occipital region, measured using ppTMS at ISI 2 ms (Figure 4). For a complete overview of the significant TMS‐EEG clusters and corresponding topo‐plots, see Table S5 and Figures S2–S4.

FIGURE 3.

Significant early ppTMS‐EEG clusters (N45 & P60) for PO‐SKET. Overview of the significant clusters found when comparing TEPs of the placebo (in blue) and oral S‐ketamine (in red) conditions, low‐dose oral (0.20 mg/kg S‐ketamine PO) or high‐dose oral (0.45 mg/kg S‐ketamine PO) for ppTMS‐EEG (A) ISI 50 ms; low‐dose; 4.5 h post‐dose, (B) ISI 100 ms; low‐dose; 4.5 h post‐dose, (C) ISI 300 ms; low‐dose; 4.5 h post‐dose, (D) ISI 100 ms; high‐dose; 4.5 h post‐dose, (E) ISI 300 ms; high‐dose; 24 h post‐dose. On the left side, the grand average (mean ± SEM) over all significant electrodes is presented. On the right side, the amplitude difference in topographical distribution (placebo – S‐ketamine) at the time of the cluster is presented. The thick green bar represents the time window of significant differences, the thick grey bar the time window of the entire cluster (including non‐significant parts), the white cross the stimulation site, and the black dots the electrode positions with the significant electrodes as red stars. Abbreviations: ISI = interstimulus interval; ms = millisecond; ppTMS‐EEG = paired pulse transcranial magnetic stimulation – encephalography; PO = per os; SEM = standard error of the mean; TEP = TMS evoked potential.

FIGURE 4.

Significant late ppTMS‐EEG clusters (P180) for IV‐SKET and HD‐SKET. Overview of the significant clusters found at timepoint 24 h post‐dose when comparing TEPs of the placebo (in blue) and S‐ketamine (in red) conditions for paired pulse TMS at ISI 2 ms for (A) high‐dose oral (0.45 mg/kg S‐ketamine PO) or and (B) IV dose (0.40 mg/kg S‐ketamine 40‐min IV). On the left side, the grand average (mean ± SEM) over all significant electrodes is presented. On the right side, the amplitude difference in topographical distribution (placebo – S‐ketamine) at the time of the cluster is presented. The thick green bar represents the time window of significant differences, the white cross the stimulation site, and the black dots the electrode positions with the significant electrodes as red stars. Abbreviations: ISI = interstimulus interval; ms = millisecond; ppTMS‐EEG = paired pulse transcranial magnetic stimulation – encephalography; PO = per os; SEM = standard error of the mean; TEP = TMS evoked potential.

3.4. pEEG

For a complete overview of the pEEG results, see Tables S6–S7. In summary, 0.40 mg/kg IV‐SKET and 0.45 mg/kg PO‐SKET showed significant acute, delayed and sustained effects on pEEG. Most consistent were decreases in delta, alpha and beta power in the eyes‐closed condition up to 6 h post‐dose, and increases in delta power in the eyes‐open condition in the parieto‐occipital regions for both 0.40 mg/kg IV‐SKET and 0.45 mg/kg PO‐SKET up to 24 h or 7 days post‐dose (Figure 5).

FIGURE 5.

Summary graphs of significant pEEG results. Change from baseline (%) of the least square means (LSM) with 95% confidence interval (95% CI) as error bars of pEEG power in the frequency bands delta (δ) 1.5 < 6.0 Hz; alpha (α): 8.5 < 12.5 Hz; and beta (β): 12.5 < 30.0 Hz, for the parieto‐occipital lead PzO1 in the eyes‐closed condition (A, C, E), and the eyes‐open condition (B, D, F) for all treatments at baseline, 2.5 h + 6 h, 24 h and 7 days post‐dose. IV‐SKET corresponds to 0.40 mg/kg S‐ketamine 40‐min IV (in blue), HD‐SKET corresponds to 0.45 mg/kg S‐ketamine PO (in pink) and LD‐SKET corresponds to 0.20 mg/kg S‐ketamine PO (in green). Significant results compared with placebo (in grey) are marked with an asterisk (*). Abbreviations: 95% CI = 95% confidence interval; BL = baseline; CFB = change from baseline, d = days; h = hours; HD‐SKET = high‐dose oral S‐ketamine (0.45 mg/kg S‐ketamine PO); IV = intravenous; IV‐SKET = 0.40 mg/kg S‐ketamine 40‐minute intravenous infusion; LD‐SKET = low‐dose oral S‐ketamine (0.20 mg/kg S‐ketamine PO); LSM = least square means; pEEG = pharmaco‐electroencephalography; PO = per os.

3.5. PK–PD relationship

Significant negative linear concentration–effect relationships were observed for DPzO1C and DPzO2C for both SKET and SNOR up to 6 h post‐dose (acute phase), and a significant positive linear concentration‐effect relationship was observed between LICI₅₀ and SKET at 7 days post‐dose (sustained phase). Results are reported in Table S9 and Figure S5.

For a summary of all the PD results, see Table S8.

4. DISCUSSION

4.1. Summary

In this study, we evaluated TMS‐ and pEEG‐derived measures as candidate biomarkers of acute (≤ 6 h), delayed (6–24 h) and sustained (24 h–7 days) pharmacodynamic effects of SKET in HVs. We observed what appears to be a temporally distinct pattern of change in corticospinal excitability, cortical excitability and resting‐state brain activity following SKET administration. Acute effects were detectable across all modalities, whereas delayed effects were primarily captured by TMS‐EEG and pEEG, and sustained effects by TMS‐EMG and pEEG. Acute reductions in cortical excitability and spectral power within the first 6 h post‐dose were followed by delayed and sustained increases at 24 h and 7 days. Concentration–effect relationships were observed between SKET concentrations for delta power in the acute phase, and for LICI₅₀ in the sustained phase.

4.2. S‐ketamine's effects on corticospinal excitability

Corticospinal excitability was affected only by 0.40 mg/kg IV‐SKET, with reduced MEP amplitudes 4.5 h post‐dose, followed by increased excitability reflected by attenuated LICI₅₀ (reduced inhibition) 7 days post‐dose. Neither PO‐SKET dose affected TMS‐EMG endpoints at any post‐dose time point. Post‐hoc analysis indicated a positive concentration‐effect relationship between SKET exposure and LICI₅₀, suggesting the observed effects are mediated by SKET rather than its metabolites.

Previous TMS‐EMG studies report divergent acute effects of SKET on TMS endpoints (rMT, MEP amplitude and SICI ²¹ , ²² ) and comparisons with the literature are hampered by limited data and methodological heterogeneity, including differences in timing (peak‐exposure vs. post peak‐exposure), dose, study design and compound (e.g., racemic vs. enantiomer‐specific ketamine with differing NMDAR affinities ³⁸ ). Unlike prior studies assessing peak exposure effects during infusion, our measurements capture post peak‐exposure effects (~4 h post‐infusion measured at ~15% of C_max). These findings suggest that the corticospinal effects observed here might reflect a later phase of SKET's PD profile that is distinct from the initial effects reported during infusion. This may reflect a transition from early glutamate‐mediated hyperexcitability to a later compensatory increase in inhibitory tone characterized as post peak‐exposure reduction in corticospinal excitability.

The attenuation of LICI observed at 7 days post‐dose may point toward involvement of longer‐term neuroplastic processes. Preclinical studies have shown that reductions in intracortical inhibitory activity can facilitate synaptic plasticity, including long‐term potentiation. ³⁹ , ⁴⁰ Moreover, a human TMS study using paired associative stimulation, a paradigm thought to induce long‐term potentiation—or long‐term depression—like plasticity, demonstrated attenuation of LICI (less inhibition) following potentiation specific stimulation protocols. ⁴¹ While highly speculative, these observations provide a potential framework for interpreting the delayed reduction in intracortical inhibition observed here, which may reflect induction of downstream NMDAR–dependent plasticity. ³⁹

4.3. S‐ketamine's effects on cortical excitability

TMS‐EEG analysis revealed significant clusters for all treatments at 4.5 h post‐dose, suggesting its high sensitivity to SKET's acute effects. Delayed effects were observed only for 0.40 mg/kg IV‐SKET and 0.45 mg/kg PO‐SKET at 24 h post‐dose, but these effects were not sustained up to 7 days post‐dose. Two consistent cluster patterns emerged, reflecting distinct acute and delayed effects, but without clear differentiation between administration routes. Acute effects involved modulation of early TEP components (N45, P60) (Figure 3), whereas delayed effects were characterized by increased P180 amplitude (Figure 4).

Direct comparison with previous TMS‐EEG studies in HVs is limited by differences in analytical approaches. ⁴² , ⁴³ Nevertheless, early TEP modulation aligns with findings from other NMDAR antagonists, supporting sensitivity to glutamatergic modulationy. ⁴⁴ However, SKET's delayed modulation of the late TEP components cannot be readily attributed to glutamatergic modulation alone. Available evidence suggests that late TEP components are modulated by GABAergic rather than by antiglutamatergic agents. ⁴⁵ Previous studies have reported reductions in P180 amplitude after administration of baclofen, a GABA_B receptor agonist, and levetiracetam, an SV2A ligand. ³⁴ , ⁴⁶ Their effects contrast with the delayed increase in P180 amplitude observed following SKET in the current study, which may suggest enhanced cortical excitability at 24 h post‐dose.

4.4. S‐ketamine's effects on spontaneous brain activity

Significant acute effects, and more limited delayed and sustained effects on spectral power were observed for both 0.40 mg/kg IV‐SKET and 0.45 mg/kg PO‐SKET. Acute decreases in delta, alpha and beta power were observed up to 6 h post‐dose, whereas delayed increases in delta power emerged at 24 h and persisted up to 7 days. Effects observed for both IV‐ and PO‐SKET aligned in terms of directionality, not clearly indicating differential effects attributable to active metabolites. Post‐hoc analysis showed negative linear concentration–effect relationships for delta power for both SKET and SNOR, suggesting both SKET and SNOR contributed to acute delta power changes, with a stronger contribution from SKET (Table S9).

Direct comparison with the existing literature remains limited. Available studies using IV administered racemic ketamine or SKET in HV have primality measured pEEG at peak concentrations, often reporting increased theta and gamma power, ²³ , ⁴⁷ , ⁴⁸ whereas the present study captured post peak‐exposure effects, which may partly explain these discrepancies. After 24 h post‐dose, resting‐state eyes‐open recordings showed increased delta power over parieto‐occipital regions, with unchanged alpha power and no frontal effects. Therefore, this pattern may suggest a region‐specific shift rather than a global change in arousal or vigilance. Notably, both TMS‐EEG and pEEG effects were located in the parieto‐occipital region. While spectral power changes cannot be directly mapped onto specific functional networks, this spatial convergence raises the possibility that posterior cortical networks are preferentially engaged by SKET at delayed time points. One potential interpretative framework is the posterior default mode network, as key nodes such as the posterior cingulate cortex and precuneus are located within these regions and have been shown to be modulated by ketamine in functional imaging studies. ⁴⁷ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ However, given the indirect nature of pEEG measures, this interpretation remains speculative.

4.5. Future perspectives

The acute decreases align with the transient CNS depressant effects reported by van Mechelen et al., ⁷ whereas the delayed and sustained increases may reflect downstream neurophysiological changes extending beyond initial NMDAR antagonism. Of note, these delayed and sustained increases align temporally with the onset of SKET's antidepressant effects at approximately 24 h post‐dose. ³ The possible clinical implication of the present work lies not in demonstrating ketamine‐specific delayed effects per se, but in identifying neurophysiological markers sensitive to delayed and sustained CNS modulation beyond acute drug exposure, considered to be a common and important characteristic of RAAD. In early‐phase CNS drug development, the lack of mechanistically informed biomarkers that bridge acute pharmacology with delayed clinical outcomes remains a major challenge. The present study suggests that TMS‐EMG, TMS‐EEG and pEEG capture distinct but complementary aspects of SKET's pharmacodynamic profile, rather than providing interchangeable measures. Individually, these endpoints offer only a partial view of drug effects; however, when combined, they provide a more comprehensive characterization across temporal and mechanistic domains. TMS‐EEG appeared most sensitive to acute drug effects, whereas pEEG captured both acute and later changes in network‐level brain activity, with these effects appearing less analyte‐specific. TMS‐EMG provided more selective evidence of corticospinal modulation by the parent compound, with MEP amplitude reflecting an acute effect and LICI₅₀ a sustained effect. Together, these findings highlight the importance of multimodal biomarker approaches, in which complementary measures are used to capture distinct phases of drug action and may support more informed endpoint selection in early‐phase studies. While the present study was not designed to optimize dosing for therapeutic efficacy, these biomarkers, especially combined, may support early go/no‐go decisions and inform dose selection including dosing regimen (e.g. weekly or biweekly dosing) by identifying exposure levels and their associated delayed CNS effects, thereby helping to balance efficacy and tolerability in future clinical studies.

Building on a more robust foundation, future studies should then extend these approaches beyond HVs to patient populations, particularly individuals with TRD, to assess the translational relevance of the observed PD effects and whether they are predictive of antidepressant response. While studies in HVs allow controlled characterization of acute, delayed, and sustained neurophysiological responses to SKET in the absence of disease‐related confounds, alterations in baseline cortical excitability, excitation–inhibition balance, and large‐scale network organization in TRD may substantially influence both the magnitude and temporal profile of drug‐induced effects.

4.6. Limitations

Implications of the present work should be interpreted in light of a few limitations. Both TMS and pEEG were not performed at peak SKET concentrations, precluding replication of previously reported acute‐phase findings. Our primary focus was, however, to identify biomarkers of sustained antidepressant effects, emerging after SKET clearance. The cross‐over design, while limiting follow‐up to 7 days to reduce participant burden, was essential to meet study aims. Although relatively low subanesthetic doses were used, these reflect proven therapeutic dose levels, enabling identification of biomarkers related to the intended clinical effect in MDD. Exclusion of the glutamate‐sensitive TMS‐EMG biomarker ICF ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ due to time constraints represents another limitation. Prior studies have suggested that the P180 may reflect residual auditory‐evoked activity. ⁵⁸ , ⁵⁹ Given our cross‐over design and the use of continuous masking noise, ⁶⁰ we believe auditory confounds were sufficiently controlled, and that the observed delayed effects on the late TEP components represent genuine neurophysiological effects. Furthermore, measurements across multiple days introduce day‐to‐day variability, which may have reduced statistical power and increased the risk of Type I errors, particularly at later time points. In addition, no correction for multiple comparisons was applied, further increasing the risk of Type I error; however, this approach was considered appropriate given the hypothesis‐generating and exploratory nature of the study, which aimed to identify candidate neurophysiological markers for future validation in adequately powered confirmatory studies. Despite these limitations, post‐hoc PK–PD analyses demonstrated concentration–effect relationships, including the later time points, suggesting that the observed effects are unlikely to be purely spurious. Nevertheless, these results should be interpreted with caution in light of the aforementioned limitations.

Future work is needed to validate these findings through larger samples, increased statistical power, and replication using similar dosing regimens and time points. Denser post‐dose sampling may further help distinguish true PD effects from between‐day variability and strengthen confidence in TMS‐ and pEEG‐derived biomarkers.

5. CONCLUSION

In this exploratory study, we provide evidence suggestive of delayed PD effects of SKET on cortical excitability and resting‐state brain activity in HV, extending beyond 24 h post‐dose and persisting for up to 7 days post‐dose, when plasma exposure has fully dissipated. The effects emerging after 24 h appeared distinct from the acute effects observed the first few hours after dosing, with a shift in directionality from reduced to increased cortical excitability and resting‐state spectral power. While signal variability at the later timepoints warrants cautious interpretation, the consistency of this temporal pattern across modalities and concentration–effect relationships supports the presence of a delayed neurophysiological response. Notably, the emergence of increased brain activity around 24 h post‐dose coincides with the typical onset of SKET's antidepressant effects, suggesting potential relevance for understanding mechanisms underlying its clinical efficacy; however, the extent to which the neurophysiological changes observed here relate to therapeutic mechanisms or clinical response remains uncertain. Both 0.40 mg/kg IV‐SKET and 0.45 mg/kg PO‐SKET showed broadly similar delayed neurophysiological effects; however, LICI₅₀ outcomes appeared to be dependent on the parent compound concentration, whereas delta power was less analyte‐specific. The combination of TMS and pEEG could serve as biomarkers of sustained antidepressant effects of non‐selective NMDAR antagonists such as SKET, and possibly for novel rapid‐acting glutamatergic antidepressants, although this remains to be confirmed in a TRD population.

AUTHOR CONTRIBUTIONS

C.M.K.E.C., A.A.G., G.E.J. and J.A.A.C.H. wrote the manuscript; L.G.J.M.B., A.G., C.J.H., G.E.J. and J.A.A.C.H. designed the research; J.C.M., L.G.J.M.B. and G.E.J. performed the research; A.A.G., L.A., E.O., R.J.D. and M.J.J. analyzed the data.

CONFLICT OF INTEREST STATEMENT

C.M.K.E.C., A.A.G., J.C.M., L.G.J.M.B., L.A., E.O, R.J.D., M.J.J., G.E.J., and J.A.A.C.H. are or were employees of the Centre for Human Drug Research when the study was conducted and declare no other competing interests; A.G. and C.J.H. are employees at the department of Psychiatry of the Oxford University and declare no other competing interests.

Supporting information

Figure S1. Study Design.

Figure S2. Significant TMS‐EEG clusters for IV‐SKET versus placebo.

Figure S3. Significant TMS‐EEG clusters for HD‐SKET versus placebo.

Figure S4. Significant TMS‐EEG clusters for LD‐SKET versus placebo.

Figure S5. Concentration‐effect relationship graphs.

Table S1. In‐ and Exclusion criteria.

Table S2. TEAEs by SOC and PT.

Table S3. Summary of PK parameters.

Table S4. Summary of TMS‐EMG results: all treatments vs. placebo.

Table S5. Summary of significant TMS‐EEG clusters: all treatments vs. placebo.

Table S6. Summary of pEEG results in the eyes‐closed condition: all treatments vs. placebo.

Table S7. Summary of pEEG results in the eyes‐open condition: all treatments vs. placebo.

Table S8. Overview of statistically significant TMS‐EMG, TMS‐EEG and pEEG effects.

Table S9. Concentration‐effect relationships pharmacodynamic parameters with significant treatment effects.

Text S1. PD‐Data Acquisition and Synthesis.

Text S2. PK‐PD relationship methods.

de Cuba CMKE, de Goede AA, van Mechelen JC, et al. Sustained pharmacodynamic effects of S‐ketamine on cortical excitability and resting‐state brain activity: A randomized, placebo‐controlled trial. Br J Clin Pharmacol. 2026;92(10):3534‐3545. doi: 10.1002/bcp.70649

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.