This brain imaging study (n=11) found that a single intravenous dose of ketamine (70mg/70kg) increased glutamate levels in the anterior cingulate cortex and altered functional connectivity between this region, the dlPFC, and the amygdala in healthy men. Multimodal imaging also suggested that ketamine-related increases in a putative marker of synaptic plasticity were linked to reduced default mode network activity.
We investigated ketamine’s neuroplastic effects in healthy human subjects using integrated Positron Emission Tomography (PET)/Magnetic Resonance Imaging (MRI) measures before and 1–8 days after a single psychedelic dose of ketamine (1 mg/kg, intravenous). Eleven male participants underwent two PET/MRI scans with [ 11 C]-UCBJ (synaptic density/plasticity), 1 H-MRS (glutamate and GABA) and resting-state fMRI (intrinsic brain activity, functional connectivity), before and after ketamine. While group-level analyses showed no significant increases in PET synaptic markers, ketamine administration resulted in significantly elevated glutamate levels within the anterior cingulate cortex (ACC). Functional connectivity analyses revealed reduced coupling between the ACC and the dorsolateral prefrontal cortex (dlPFC) and increased coupling between the ACC and the amygdala in the days following ketamine administration. Our multimodal analysis revealed that participants showing an increase in [ 11 C]-UCBJ volume distribution (VT), a putative index of synaptic plasticity, showed a correlated reduction in intrinsic activity within regions belonging to the default mode network (DMN). By linking molecular, cellular and network-level changes, our results point to the DMN as a central hub where ketamine may reshape brain hierarchies in the long term, providing new directions for understanding its therapeutic mechanisms and developing targeted treatments.
Papers cited by this study that are also in Blossom
Bowdle, A. T., Radant, A. D., Cowley, D. S. et al. · Anesthesiology (1998)
Agnorelli, C., Spriggs, M. J., Godfrey, K. et al. · Preprints (2024)
Ly, C., Greb, A. C., Cameron, L. P. et al. · Cell Reports (2018)
Nardou, R., Sawyer, E., Song, Y. J. et al. · Nature (2023)
Ketamine has dose-dependent psychoactive effects and, at sub-anaesthetic doses, can produce rapid and sustained antidepressant effects that last beyond the period of drug exposure. Earlier animal and human studies suggest that these lasting benefits may involve neuroplastic changes, including altered synaptic structure, glutamatergic signalling, and resting-state brain network activity. However, evidence in living humans has remained incomplete and sometimes inconsistent, especially for direct in vivo measures of synaptic plasticity. Prior PET work using the synaptic marker [11C]-UCBJ, which targets SV2A, had not shown a clear group-level ketamine effect in humans, and the relationship between synaptic measures and functional MRI findings was still poorly understood. This study set out to use a multimodal PET/MRI approach to examine ketamine’s effects across multiple biological levels in healthy men. The authors aimed to test whether a single intravenous psychedelic dose of ketamine would increase [11C]-UCBJ VT, a putative marker of synaptic plasticity, particularly in mood-regulating regions within the default mode network (DMN), over 1-8 days after dosing. They also examined glutamate and GABA in the anterior cingulate cortex (ACC), regional intrinsic brain activity, and functional connectivity, and explored whether changes in synaptic markers were related to changes in whole-brain activity.
The study was an exploratory within-subject imaging investigation in 11 healthy male participants, mean age 32 ± 10 years. Participants were screened to exclude physical or psychiatric illness, substance misuse, recent ketamine or psychedelic exposure, and MRI/PET contraindications. They abstained from alcohol and illicit drugs during the study period, with urine and breath testing at each visit. Ethical approval was obtained in London, and all participants gave informed consent. Each participant received a single intravenous infusion of racemic ketamine at 1 mg/kg over 40 minutes. This was described as a sub-anaesthetic, psychedelic-like dose that is used clinically and had known antidepressant effects. Vital signs were monitored before, during and after infusion. Psychological effects were assessed acutely with the Brief Psychiatric Rating Scale (BPRS), the Modified Observer’s Assessment of Alertness/Sedation scale, and the 6-item Clinician-Administered Dissociative States Scale (CADSS). Mood and wellbeing were measured at baseline, 1 week, and 4 weeks using the Profile of Mood States (POMS) and Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS). Participants underwent two integrated PET/MRI scans. The first scan occurred about 2 weeks before ketamine, and the second occurred after ketamine, either 1-2 days later in five participants or 7-8 days later in six participants; this timing was chosen to sample possible temporal effects across the post-infusion window. PET imaging used [11C]-UCBJ to quantify SV2A-related binding, with [11C]-UCBJ VT as the primary outcome. The authors focused on pre-specified regions including the dlPFC, vmPFC, ACC, PCC, hippocampus and amygdala. MRI included 1H-MRS to measure glutamate and GABA in the ACC, and resting-state fMRI to assess intrinsic activity and functional connectivity. Due to quality control issues, some MRS GABA data were excluded, and the HERMES sequence was not used in the main analysis. PET data were modelled with a 1-tissue compartment approach. Functional connectivity was calculated from pairwise correlations between ACC and the other regions of interest, with the ACC used as a seed. Intrinsic activity was assessed using ALFF, a measure of low-frequency BOLD signal power. Statistical analysis relied on one-sided linear mixed-effects models for within-subject pre-post changes in PET measures and mood scales, two-sided models for glutamate, Spearman correlations for associations between imaging and behavioural measures, voxel-wise mixed-effects analysis for ALFF, and permutation-based testing for functional connectivity. Multiple comparisons were controlled with Benjamini-Hochberg or Bonferroni corrections as appropriate.
Ketamine was followed by statistically significant improvements in mood measures: POMS scores decreased at 7 days and 4 weeks, and WEMWBS showed a trend towards improvement at 7 days. The extracted text reports POMS changes of M: -15 ± 20 at 7 days (β = -14.82, p = 0.031) and M: -14 ± 15 at 4 weeks (β = -13.91, p = 0.011), while WEMWBS increased by M: 6 ± 9 at 7 days (β = 5.64, p = 0.072). For the primary PET outcome, no statistically significant group-level increase in [11C]-UCBJ VT was detected after ketamine in any region of interest when all 11 participants were analysed. There was, however, a non-significant trend towards higher VT across regions, including the amygdala, where the increase approached significance (M: 7.33 ± 13.44%; β = 1.17; p = 0.066; d = 0.495; BF = 0.837). The extracted text is incomplete at the point where the ACC result is reported, but it indicates a similar trend in the ACC. No significant relationships were found between changes in [11C]-UCBJ VT and long-term changes in WEMWBS or POMS in any analysed region. Ketamine produced a statistically significant increase in glutamate levels within the ACC. The extracted text does not give the exact effect size or p-value in the Results subsection, but the Discussion states that this glutamate increase was particularly pronounced in the 7-8 day group. There was no significant correlation between the ketamine-related change in ACC glutamate and the change in [11C]-UCBJ VT in the ACC (n = 9; ρ = 0.01, p = 0.980). In terms of network measures, ketamine altered ACC functional connectivity: the Discussion specifies that coupling between the ACC and dlPFC became more anti-correlated, while coupling between the ACC and amygdala shifted from baseline anticorrelation towards positive coupling. For whole-brain intrinsic activity, ketamine produced a statistically significant reduction in ALFF only in a small occipital cortex cluster. The exploratory multimodal analysis found a negative relationship between changes in [11C]-UCBJ VT and ALFF, but only in regions belonging to the DMN, indicating that participants with larger increases in the synaptic marker tended to show larger reductions in intrinsic activity. The extracted text does not report any significant correlation between the PET and functional measures at baseline in those same DMN regions.
The authors conclude that a single 1 mg/kg dose of ketamine did not produce a clear group-level increase in [11C]-UCBJ VT in DMN regions 1-8 days later in healthy men. They interpret this as inconclusive rather than negative, noting that the data showed a trend towards increased VT and that a larger, better powered sample may be needed to detect any effect. They also report no difference between the early scan group (1-2 days) and the later scan group (7-8 days) for PET metrics. Their broader interpretation is that ketamine may exert complex and heterogeneous effects across molecular, cellular, and network levels rather than a single uniform neuroplastic response. In contrast to the PET findings, the authors emphasise a robust and sustained increase in ACC glutamate after ketamine, particularly in the later post-dose window. They note that this persistence extends beyond the commonly assumed acute glutamatergic surge and may indicate a longer-lasting shift in excitatory/inhibitory balance in the ACC. They place this within the wider idea that ketamine acts through NMDA receptor antagonism, disinhibition of pyramidal neurons, and downstream synaptic remodelling. Although they did not find a correlation between ACC glutamate and ACC [11C]-UCBJ VT, they argue that the ACC remains a key region in ketamine’s mechanism because it also showed altered coupling with the dlPFC and amygdala. The authors interpret the connectivity changes as a reconfiguration of fronto-limbic circuitry that may be relevant to ketamine’s therapeutic actions, although they stress that this healthy-volunteer study cannot test antidepressant efficacy. They suggest that ketamine may reduce maladaptive ACC-PFC synchrony and alter ACC-limbic interactions in ways that resemble previously reported changes in depressed patients. They also discuss the ALFF findings as evidence of reduced intrinsic activity, albeit limited to a small occipital cluster at the whole-brain level. Their exploratory PET-fMRI analysis is presented as especially informative because it linked increases in the synaptic marker with decreases in intrinsic activity in DMN regions, which they interpret as consistent with ketamine perturbing self-referential network processing. The authors compare their findings with earlier work, including prior [11C]-UCBJ PET studies that were also inconclusive, and they suggest several reasons why group-level PET effects may be hard to detect in humans. These include high inter-individual variability, the possibility that repeated dosing is needed, stronger effects in clinical populations than in healthy volunteers, and limitations of the tracer itself. They acknowledge several important limitations: the very small sample, inclusion of only male participants, the fact that none were naive to classic psychedelics, the varying post-dose scan times, and the limited spatial resolution and biological specificity of [11C]-UCBJ PET. They also caution that the tracer may not cleanly index a single form of plasticity and may be influenced by other presynaptic processes. Overall, they present the study as a proof-of-concept multimodal investigation that offers hypotheses for future work rather than definitive evidence of ketamine-induced synaptic plasticity in healthy humans.
A total of 11 healthy male participants (mean age 32 ± 10 years) were screened and included in the study. Informed consent was obtained from all participants. Participants' details are listed in Table. The inclusion criteria were as follows: participants needed to be between 20 and 60 years of age, exhibit no physical or psychiatric medical conditions, and have no history or current incidence of substance abuse. Consumption of ketamine or classic psychedelics in the 6 months preceding the beginning of the study was a ground for exclusion to avoid carry-over neuroplastic effects from previous substance exposure.Of the total sample, none of the participants were naïve to classic psychedelics, while n = 7 participants were ketamine naïve. Additionally, participants were required to abstain from alcohol and illicit substance intake for at least 1 week before and throughout the study period, as assessed via urine drug screen tests and breathalyser tests on each visit. Participants were screened for contraindications to MRI or research PET scanning. Ethical approval was granted by the Brent Research Ethics Committee, London, UK, following the Declaration of Helsinki guidelines (Supplemental Data).
Racemic ketamine was obtained from the St Charles Centre for Health & Wellbeing Pharmacy and administered intravenously by constant infusion over 40 min at a dose of 1 mg/kg. This dosage is sub-anaesthetic and produces a psychedelic-like experience often described as dissociative, and has confirmed antidepressant effects.The amount of ketamine administered was comparable across participants (78 ± 13 mg), and all participants experienced the dissociative effects of the drug (Table). Vital signs (blood pressure and heart rate) were obtained before, during and after ketamine infusion.
Each participant underwent 2 PET/MRI scans, where [ 11 C]-UCBJ was administered intravenously for the quantification of brain SV2A (Figure). Subsequently, 1 H-MRS was acquired to quantify glutamate and GABA concentrations within the ACC. Then, resting-state fMRI data were obtained to assess whole-brain intrinsic brain activity and connectivity. Participants had their first scan (i.e. PET/MRI-1) approximately 2 weeks before the ketamine infusion. Of the 11 participants, a subset of 4 participants underwent their post-ketamine scan (i.e. PET/ MRI-2) 1 day after ketamine infusion. One participant had the scan 2 days after the ketamine infusion due to tracer production failure 1 day post-ketamine administration. Participants who had their second scan either at 1 or 2 days after ketamine are referred to as the Day 1-2 group (n = 5). A subset of 4 participants had their Scan 2 at 7 days after ketamine infusion. Two participants had their Scan 2 at 8 days after ketamine infusion due to tracer production failure at 7 days post-ketamine administration. Those participants are referred to as the Day 7-8 group (n = 6). This design was chosen to sample potential temporal dynamics across the 1-8-day post-infusion interval. The study was not powered to detect between-subgroup differences. Therefore, subgroup comparisons were considered exploratory and are reported in the Supplemental Data.
The psychological safety of participants during the acute effects of ketamine was assessed through the Brief Psychiatric Rating Scale (BPRS),the Modified Observer's Assessment of Alertness/Sedation (MOAA/S) scale,and the 6-item version of the Clinician-Administered Dissociative States Scale (CADSS) (Table).These scales were administered by the study physician immediately after completion of the ketamine infusion. Further, the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS)and The Profile of Mood States (POMS)questionnaires were administered at baseline, 1 week, and 4 weeks after ketamine to measure changes in well-being and mood, respectively.
The PET data were acquired using an integrated General Electric Signa 3 Tesla combined PET/MRI scanner with a 32-channel head coil at the Invicro Imaging Centre (now 'Perceptive'), London, UK. The [ 11 C]-UCBJ tracer was synthesised onsite and administered i.v. as a bolus over 20 s by the study physician. The PET scan acquisition time was 90 min in addition to 30 min of MRI scanning. For the PET modelling, the study employed the 1-tissue compartment model for reversible binding to correlate the parent plasma input function with tissue time-activity curves, producing estimates of the total VT for each predefined ROI. The fMRI data were acquired using an Echo-Planar Imaging sequence sensitive to BOLD contrast with the following settings: TR = 2000 ms, TE = 30 ms, flip angle = 80°, 3 mm × 3 mm in-plane resolution (64 × 64 matrix), slice thickness = 3.6 mm, 36 axial slices, 240 volumes, 10 min. In addition, a T1-weighted IR-SPGR sequence was acquired to provide an anatomical image (TI = 400 ms, TE = minimum, flip angle = 11°, 256 × 256 matrix, 1 mm isotropic voxels, sagittal slices). MRS data included two 1 H-MRS scans: a PRESS-PROBE sequence (TE = 30 ms, 2 × 2 × 2 cm 3 , 96 averages) and a GABAand GSH-edited HERMES sequence (TE = 76 ms, 2.5 × 2.5 × 3 cm 3 , NEX = 8). Voxels were positioned in the ACC based on previous evidence supporting its predominant role in ketamine's mechanism of action.After quality control, datasets from 2 subjects on the PRESS data were excluded, and 7 time points were excluded on the HERMES data. Thus, the data from the HERMES sequence were excluded from the main analysis (Supplemental Table). See Supplemental Data for full details on PET/MRI acquisition and processing.
Pre-defined ROIs were the dorsolateral (dlPFC) and ventromedial (vmPFC) prefrontal cortex, ACC, PCC, hippocampus and amygdala. The selection of the ROIs was based on previously published work on [ 11 C]-UCBJ VT changes induced by ketamine within the DMN.This choice was motivated by several reasons. First, VT was our pre-specified primary outcome measure. Also, there was a trend in the data suggesting a difference in the control region VT between pre-and post-ketamine visits, which could potentially influence DVR-1 estimates. Further, fp did not differ between sessions, indicating stable peripheral free fraction. Therefore, fp-correction was not expected to reduce session-related confounding but could have introduced additional variance. Lastly, no correlations were observed between the change in VT and the VT/fp and DVR-1 metrics, potentially due to the variance introduced by the DVR-1 and VT/fp normalisations (Supplemental Data).
The mean BOLD timeseries of each fMRI scan was extracted from the same pre-defined ROIs used for PET analysis, including the dlPFC, vmPFC, ACC, PCC, hippocampus and amygdala. Brain FC was computed as the pairwise Pearson's correlation between each ROI and the ACC, used as a seed region. The ACC was chosen as the seed region for the FC, consistent with previous investigations,to test the complementary network effects of ketamine alongside its effects on glutamate concentration.
To compute whole-brain intrinsic activity, ALFF measures were calculated using the Analysis of Functional Neuroimages (AFNI v.20.1.06) 3dRSFC module. ALFF is the power of the BOLD signal in the 0.01-0.1 Hz low-frequency range. The data in these analyses were band-pass filtered using a range of 0.01-0.1 Hz. As ALFF is sensitive to the raw values of the BOLD time series (which are arbitrary values), some normalisation of ALFF measures is standard practice.In this case, Z-normalisation was used after ALFF analysis, where the mean of each participant's ALFF image was divided by its standard deviation, to give an overall Z-score brain map for each participant (Supplemental Data).
A one-sided linear mixed-effect model (LMM) was used for within-subject analysis, measuring changes in WEMWEBS, POMS, fp, VT, DVR-1 and VT/fp measures of [ 11 C]-UCBJ after ketamine administration (e.g. ROI_ VT ~ Time + (1|Subject)). See Supplemental Data for analysis of ketamine effects across multiple regions. An interaction term was added to the model to test for the between-subject difference in [ 11 C]-UCBJ metrics change before and after ketamine administration between participants who had their Scan 2 at 1-2 days following ketamine and those who had it at 7-8 days. For all models, the effect size was estimated using Cohen's d. Also, the confidence of the result was estimated by computing the Bayes Factor (BF). The same statistical approach was adopted to test pre-to post-ketamine differences in ACC Glutamate concentrations (Two-sided). Pair-wise Spearman's correlation tests were used to analyse the correlation between WEMWBS and CADSS scores, and between these psychological and the neural measures. The same test was used for correlations between [ 11 C]-UCBJ VT, DVR-1, and VT/fp metrics and their difference following ketamine and to test for the correlation between glutamate concentrations and [ 11 C]-UCBJ within the ACC. To account for FDR inflation due to multiple comparisons, the p-values resulting from Spearman's tests were adjusted independently using the Benjamini-Hochberg adjustment. To measure the effects of ketamine on ALFF, we implemented a voxel-wise paired analysis using FSL's FEAT with a second-level mixed-effects model (FLAME1). Results were thresholded at Z = 3.1, with a cluster p < 0.05. To test for FC difference following ketamine administration, a permutation test (1000 permutations) was performed using the R coefficients resulting from Pearson's correlation test between the mean pre-and post-ketamine fMRI scans. The p-values were corrected for multiple comparisons using Bonferroni correction across the five comparisons (i.e. α < 0.001). To explore the correlation between the structural and functional measures acquired with PET and fMRI, respectively, a whole-brain data-driven approach was adopted. Voxel-wise Pearson's correlations were performed between [ 11 C]-UCBJ VT and ALFF datasets, following methods in Shatalina et al.,to explore correlations between PET and fMRI metrics in a whole-brain, data-driven approach. Statistical analyses were conducted in MATLAB 2023A and R 2024.
We observed statistically significant reductions in POMS at 7 days (M: -15 ± 20; β = -14.82; p = 0.031), and 4 weeks (M: -14 ± 15; β = -13.91; p = 0.011) and a trend towards increases in WEMWBS at 7 days following ketamine (M: 6 ± 9; β = 5.64; p = 0.072), suggesting improvements in mood and wellbeing.
Our primary PET outcome measure was [ 11 C]-UCBJ VT (See Materials and Methods). Assessments of the normality of the PET data and results from the normalised DVR-1 and VT/fp [ 11 C]-UCBJ metrics are reported in Supplemental Data. With all subjects included in the analysis (n = 11), we observed no statistically significant increase in [ 11 C]-UCBJ VT following ketamine administration in any of the analysed ROIs (Supplemental Table). Notably, there was a trend towards an increase in [ 11 C]-UCBJ VT in all ROIs, approaching statistical significance in the amygdala (Figure(f); M: 7.33 ± 13.44%; β = 1.17; p = 0.066; d = 0.495; BF = 0.837) and the ACC (Figure
Ketamine produced a statistically significant increase in glutamate levels within the ACC (Figure
To investigate the effects of ketamine on ACC-related network activity, the ACC was used as a seed region for FC analysis using our pre-defined ROIs. There was a
No statistically significant relationships were found between changes in [ 11 C]-UCBJ VT and long-term changes in WEMWEBS and POMS in any of the analysed ROIs (Supplemental Table).
No statistically significant relationship was found between changes in glutamate and [ 11 C]-UCBJ VT within the ACC following ketamine administration (n = 9; ρ = 0.01, p = 0.980).
Ketamine produced a statistically significant reduction in ALFF only in a small region of the occipital cortex (n = 11; FLAME-1, Z > 3.1, cluster p < 0.05; Figure
A single sub-anaesthetic ketamine dose did not produce a statistically significant increase in [ 11 C]-UCBJ 1-8 days post administration within DMN regions in healthy male participants. Additionally, no differences were found between participants scanned 1-2 days after ketamine administration and those scanned 7-8 days after, in any of the analysed PET metrics and ROIs. However, we observed a robust and sustained increase in glutamate concentrations within the ACC following ketamine administration, particularly pronounced at 7-8 days after drug. Further, we observed significant changes in functional coupling between the ACC and the dlPFC and between the ACC and the amygdala. Our exploratory, whole-brain, multimodal PET/fMRI analysis revealed that alterations in synaptic plasticity, as captured via [ 11 C]-UCBJ VT, were inversely related to changes in intrinsic regional brain activity measured via ALFF, emerging in brain areas belonging to the DMN. Taken together, our findings suggest a complex and heterogeneous spatiotemporal effect of ketamine on neural plasticity, glutamate dynamics and functional brain activity and connectivity across different scales. Notably, we detected significant improvements in mood that were maintained for up to 4 weeks following a single dose of ketamine in our healthy adult sample. However, these changes did not correlate with changes in [ 11 C]-UCBJ VT and the interpretability of these behavioural findings is restricted by the small sample size and the fact that our study enrolled only male healthy volunteers. While the change in [ 11 C]-UCBJ VT following ketamine was not statistically significant, there was a trend towards an increase in our pre-defined ROIs. The trend was further explored with an additional analysis across the pre-defined ROIs and whole-cortex (See Supplemental Data). When analysing all pre-defined ROIs collectively in a single model, we observed a statistically significant increase in [ 11 C]-UCBJ VT after ketamine (p < 0.001). The whole-cortex analysis yielded similar significant results (p < 0.001). However, when we calculated volumeweighted averages of either the pre-defined ROIs or all cortical regions, the results mirrored our initial findings, showing a non-significant trend (p = 0.092). An in-depth discussion of this analysis is presented in the Supplemental Data. The [ 11 C]-UCBJ DVR-1 and VT/fp normalised metrics showed no change following ketamine exposure, but those metrics had validity issues. Notably, the VT trend was also present in the CS reference region used for DVR-1, which undermines the interpretability of the DVR-1. Also, the [ 11 C]-UCBJ VT and VT/fp did not covary following ketamine, and fp itself did not differ after ketamine, suggesting increased variance introduced by the VT/fp normalisation (Supplemental Data). Overall, our results do not provide conclusive evidence of a group-level increase in synaptic plasticity within the DMN induced by a single psychedelic dose of ketamine (1 mg/kg) in healthy human brains at 1-8 days post-administration, potentially due to high individual variability in drug neuroplastic response. Nonetheless, the presence of a trend is noteworthy, warranting replication in a larger sample size. Our results broadly align with the study by Holmes et al.,which also provided inconclusive evidence regarding a unidirectional effect of ketamine (0.5 mg/kg) on synaptic SV2A within DMN regions, as measured via [ 11 C]-UCBJ, in both healthy and depressed individuals 1 day post-drug administration. Several factors could account for the absence of sizeable changes in [ 11 C]-UCBJ following ketamine. While we employed double the dosage of Holmes et al.,approximating those commonly used in pre-clinical studiesand the highest titration dose used in psychiatric practice,the absence of a robust effect on [ 11 C]-UCBJ suggests that changes in synaptic plasticity in humans (if they could be imaged at all) might require multiple doses to be detected via PET. This is supported by research showing that some patients require multiple ketamine doses to achieve therapeutic benefit, associated with significant structural and functional alterations.Another possibility is that ketamine's effects on plasticity may be more pronounced in patients with depression, as supported by some clinicaland preclinical observations.Also, [ 11 C]-UCBJ binding reductions were detected in advancedbut not early-stage schizophrenia,suggesting that [ 11 C]-UCBJ may be more effective as a biomarker in cases of substantial neural remodelling, such as observed in severe patient populations. Lastly, multiple physiological factors may have influenced the [ 11 C]-UCBJ VT results (see Supplemental Data for discussion on sleep). Within the ACC, ketamine produced a statistically significant increase in glutamate concentrations, with the effect particularly pronounced in the Day 7-8 group. The glutamate 'surge' induced by psychedelic doses of ketamine is generally thought to be acute,not extending beyond the drug's presence in the body, as previous studies did not detect significant changes in glutamate concentrations at 1-2 days post-ketamine in either healthy subjects or depressed patients.However, using high-resolution MRS, Li et al.found an increased glutamine/glutamate ratio measured in the ACC 1 day post ketamine administration in healthy volunteers, while another study found a significant glutamate increase at 7 days.Alongside our findings, the persistence of elevated glutamate from 1 to 8 days post-administration challenges the model of ketamine-induced glutamate increase as purely transient. Yet, differences in scanning sequences and acquisition methodologies across studies might account for the variability in the results. The predominant framework, largely based on pre-clinical evidence, proposes that ketamine induces an acute disinhibition of frontocortical pyramidal neurons, via suppression of inhibitory interneurons following ketamine-mediated N-methyl-D-aspartated (NMDA) receptor antagonism, triggering downstream signalling cascades that can promote synaptic remodelling over hours to days.In our limited sample of healthy male participants, we did not observe an association between the ketamine-related increase in ACC glutamate and synaptic plasticity within the same region. Nevertheless, the observed persistence of increased glutamate concentrations beyond acute drug exposure is noteworthy and may point to a more enduring shift in excitatory/inhibitory (E/I) balance within the ACC following ketamine. This is relevant as dysregulation of cortical E/I balance has been proposed to contribute to depressive symptomatology, and ketamine may transiently recalibrate these processes through glutamatergic mechanisms.Importantly, the ACC has been repeatedly implicated as a locus of ketamine's antidepressant action across human neuroimaging and animal work, including evidence linking ketamine-related modulation of ACC activity and circuitry to antidepressants and anti-anhedonic effects.In this context, our ACC-seeded fMRI functional connectivity analysis indicates that ketamine significantly altered ACC coupling with both the dlPFC and amygdala. Specifically, ketamine shifted the baseline anticorrelation between ACC and amygdala towards positive coupling, while shifting the weak baseline correlation between ACC and dlPFC activity towards anticorrelation. Although these findings are not a test of antidepressant efficacy in this healthy-volunteer study, the reconfiguration of ACC connectivity with fronto-limbic regions is broadly consistent with previous evidence and suggests ketamine's therapeutic effects may involve perturbing maladaptive ACC-PFC hypersynchrony observed in depressive states and reconfiguring ACC interactions with limbic circuitry.Nonetheless, given the small sample and correlational nature of our results, these interpretations remain provisional and require replication in adequately powered clinical studies and, ideally, causal designs. At the whole-brain level, we observed a trend towards a decrease in intrinsic brain activity, measured via ALFF, which was statistically significant only in a small region of the occipital cortex. Previous studies using fractional ALFF, instead of raw ALFF, during ketamine infusion reported either globalor occipital-specificreductions, aligning with our observation. Notably, one of these studies indicated that the effect diminished by 1 day post-administration, which may account for the lack of statistical significance observed in our data.ALFF quantifies the power of infra-slow (0.01-0.1 Hz) BOLD signal fluctuations at the voxel level. Multimodal evidence indicates BOLD signal changes correlate with high-frequency synchronised neural activity, as measured via integrated fMRI/electroencephalography (EEG). In landmark studies, it was shown that regional ketamine-induced BOLD signal changes and connectivity patterns follow different temporal profiles of EEG desynchronisation of oscillatory activity.However, no clear relationships between the EEG metrics and fractional ALFF were detected during ketamine administration, suggesting a complex relationship between those metrics under pharmacological perturbation.While the reduction of ALFF observed in our study might be a product of ketamine-induced perturbation of synchronous neural activity in the infra-slow range or at higher frequencies, attenuated in effect size due to the temporal distance from the infusion, more research is needed to elucidate the physiological source of this effect. Our exploratory multimodal analysis revealed intriguing neural dynamics associated with ketamine-induced alterations of synaptic plasticity. A negative relationship was found between changes in [ 11 C]-UCBJ VT and ALFF following ketamine, meaning that participants showing an increase in synaptic plasticity following ketamine also showed reductions in intrinsic brain activity. Interestingly, these effects emerged specifically in regions belonging to the DMN, despite our data-driven, whole-brain approach. At baseline, no correlation was observed in the same regions between [ 11 C]-UCBJ VT and ALFF, but there was a positive correlation between these metrics in other lower-order brain regions (Supplemental Data). It has been proposed that the DMN involvement with narrative-self-referential processes makes it responsible, when overactivated, for ruminative and negative thought loops, characteristics of depressive symptomatology.Hence, similar to classic psychedelics,the perturbation of DMN functioning produced by ketamine may recalibrate the network activity by flattening its apical position in the cortical hierarchy and promoting its integration with bottom-up processes. Our results also suggest that ketamine's modulation of DMN activity is sustained in the long term and is paralleled by increases in synaptic plasticity. We speculate that the observed effect of ketamine may offer a precise spatiotemporal window for therapeutic intervention. However, our interpretation is tentative and specific hypothesis-driven approaches are required to confirm the validity of our findings. Importantly, our experimental design presents several limitations. First, the limited sample size (n = 11) significantly constrained our statistical power and included only male subjects, limiting generalisability. Additionally, none of the participants were naïve to classic psychedelics, although most participants were ketamine-naïve and all had abstained from ketamine and classic psychedelics for at least 6 months before study entry. Participants were scanned at varying time points post-ketamine (1-8 days) to capture a comprehensive view of ketamine's temporal effects on [ 11 C]-UCBJ. However, this approach reduced the specificity of our group-level analyses. Limitations inherent to the [ 11 C]-UCBJ tracer also present challenges for examining the subtle and dynamic neuroplastic effects induced by pharmacological interventions in humans in vivo. Indeed, it is likely that the spatial resolution (~2-3 mm) of [ 11 C]-UCBJ PET imaging is insufficient to accurately capture the dynamic synaptic alterations induced by ketamine in the living brain. The [ 11 C]-UCBJ tracer targets a pre-synaptic protein that is ubiquitously expressed across neuronal populations, failing to capture the neural population-specific effects captured by animal research.Also, most of the evidence reports that ketamine primarily exerts post-synaptic effects on neuroplasticity, such as increased dendritic spine density, with only a fraction of these spines maturing into functional synapses.Preclinical research indicates that sub-anaesthetic ketamine doses increase spine density by 12%-50%, with approximately 20% of these new spines developing into functional synapses.This translates to an estimated 3%-10% increase in synaptogenesis, which is within the range of effects observed in our study and by Holmes et al.in severe depression. However, it has also been reported that ketamine has inhibitory effects on neurotransmitter recycling and vesicle trafficking, adding further complexity to interpreting SV2A as a plasticity marker.Lastly, increases in SV2A could also result from a rise in the number of vesicles at the presynaptic terminal, potentially due to Hebbian learning, or from changes in the number of SV2A molecules per vesicle.Therefore, while [ 11 C]-UCBJ is among the most precise tracers currently available for in vivo imaging of synaptic plasticity, the specific process it captures remains uncertain. Despite these limitations, our comprehensive multimodal assessment yielded promising results, serving as a proof-of-concept for using integrated PET/MRI to characterise ketamine's pharmacodynamic effects. While unimodal effect sizes for the main effects of ketamine on PET metrics were limited, the MRS and fMRI analyses provided novel insights into sustained ketamine effects on glutamate concentrations and network dynamics across the DMN, respectively. Our data-driven approach also highlighted the DMN regions as a key substrate for ketamine's cross-modal effects, linking changes in synaptic plasticity markers with intrinsic brain activity, offering valuable guidance for future hypotheses with meaningful clinical potential in developing targeted interventions.
Create a free account to open full-text PDFs.
Erritzoe, D., Timmermann, C., Godfrey, K. et al. · Nature Mental Health (2024)
Marguilho, M., Figueiredo, I., Castro-Rodrigues, P. · Journal of Psychopharmacology (2022)
Mueller, F., Musso, F., London, M. et al. · NeuroImage (2018)
Alexander, L., Jelen, L. A., Mehta, M. A. et al. · Neuroscience and Biobehavioral Reviews (2021)
Gattuso, J. J., Perkins, D., Ruffell, S. G. D. et al. · International Journal of Neuropsychopharmacology (2022)