This double-blind placebo-controlled study (n=48) investigated the antidepressant efficacy of ketamine (14 or 35mg/70kg) in patients with depression and found evidence that its rapid antidepressant effects at 40 and 240 minutes post-treatment were facilitated by glutamatergic neurotransmission in the prefrontal cortex.
Background
Low-dose ketamine has been found to have robust and rapid antidepressant effects. A hypoactive prefrontal cortex (PFC) and a hyperactive amygdala have been suggested to be associated with treatment-resistant depression (TRD). However, it is unclear whether the rapid antidepressant mechanisms of ketamine on TRD involve changes in glutamatergic neurotransmission in the PFC and the amygdala.
Methods
A group of 48 TRD patients were recruited and equally randomized into three groups (A: 0.5 kg/mg-ketamine; B: 0.2 kg/mg-ketamine; and C: normal saline [NS]). Standardized uptake values (SUV) of glucose metabolism measured by 18F-FDG positron-emission-tomography before and immediately after a 40-min ketamine or NS infusion were used for subsequent region-of-interest (ROI) analyses (a priori regions: PFC and amygdala) and whole-brain voxel-wise analyses and were correlated with antidepressant responses, as defined by the Hamilton depression rating scale score. The 18F-FDG signals were used as a proxy measure of glutamate neurotransmission.
Results
The ROI analysis indicated that Group A and Group B, but not Group C, had increases in the SUV of the PFC (group-by-time interaction: F = 7.373, P = 0.002), whereas decreases in the SUV of the amygdala were observed in all three groups (main effect of time, P < 0.001). The voxel-wise analysis further confirmed a significant group effect on the PFC (corrected for family-wise errors, P < 0.05; post hoc analysis: Group A<Group C, Group B<Group C). The SUV differences in the PFC predicted the antidepressant responses at 40 and 240 min post-treatment. The PFC changes did not differ between those with and without side effects.
Conclusion
Ketamine's rapid antidepressant effects involved the facilitation of glutamatergic neurotransmission in the PFC.
Ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, produces rapid antidepressant effects at low doses and acts directly on the glutamatergic system. Prior work has suggested that treatment-resistant depression (TRD) is characterised by hypoactivity in the prefrontal cortex (PFC) alongside hyperactivity in limbic regions such as the amygdala, but it is unclear whether ketamine's rapid clinical benefits in TRD involve changes in glutamatergic neurotransmission specifically within these regions. Li and colleagues designed a randomized, double-blind, placebo-controlled study using 18F-fluorodeoxyglucose positron-emission tomography (18F-FDG-PET) to probe rapid changes in brain glucose uptake as a proxy for glutamate-related activity. The primary hypothesis was that low-dose ketamine would rapidly increase PFC activity; the amygdala was also examined as a key limbic node in mood circuitry. Scans and clinical ratings were scheduled to capture effects within the first hour after infusion and to relate imaging changes to fast antidepressant responses.
This was a double-blind, randomized, placebo-controlled trial in 48 adults (aged 21–65) meeting DSM-IV-TR criteria for major depressive disorder who met prespecified criteria for treatment-resistant depression (failure to respond to at least three antidepressants across adequate doses/durations plus at least one failed trial during the current episode). Key medical, neurological and substance-use exclusions applied; informed consent and ethical approval were obtained. Participants continued their background medications (the study was an add-on design). Subjects were randomised 1:1:1 to receive a 40-minute intravenous infusion of ketamine 0.5 mg/kg (Group A), ketamine 0.2 mg/kg (Group B), or normal saline (Group C), with 16 subjects per group. Depressive symptoms were assessed with the 17-item Hamilton Depression Rating Scale (HDRS-17) at baseline and at 40, 80, 120 and 240 minutes post-infusion; responders were defined as ≥50% reduction in HDRS-17. The positive symptoms subscale of the Brief Psychiatric Rating Scale (BPRS) and systematic side-effect monitoring were used to track psychotomimetic and other adverse effects. Neuroimaging comprised two resting 15-minute 18F-FDG-PET acquisitions (pre- and post-infusion) together with structural MRI for normalisation. The protocol involved two FDG injections and timing intended to capture glucose uptake before and after the 40-minute infusion; standardised uptake values (SUVs) were calculated as a semi-quantitative index of glucose metabolism and were interpreted here as a proxy for glutamatergic neurotransmission (FDG uptake is largely driven by astrocytic metabolism in response to neuronal glutamate release). Analysis included region-of-interest (ROI) extraction for the PFC and amygdala using automated anatomical labelling (AAL) templates and normalisation of PFC SUV to global mean SUV, together with voxel-wise whole-brain analyses using SPM8. The primary statistical approaches were two-way repeated-measures ANOVA (time × group) with age and sex as covariates, post hoc t-tests (adjusted for multiple comparisons), Pearson correlations relating SUV changes to HDRS-17 percent change, and linear regression models including age, sex, baseline HDRS-17 and SUV changes to predict antidepressant response. Standard thresholds for significance (P < 0.05, with family-wise error correction applied for voxel-wise maps) were used.
Ketamine, which is an N-methyl-D-aspartate (NMDA) antagonist, is an anesthetic agent with a rapid onset and a short duration of action. The onset of ketamine peaks in 1-5 min, and the duration of action is less than 20 min. A growing body of evidence has indicated that the glutamatergic system, which ketamine acts directly on, could be critically involved in the pathophysiology of depression. Low-dose ketamine results in rapid improvement of depressive symptoms (within 1 h) in patients with treatment-resistant depression (TRD). Robust antidepressant effects of ketamine have been shown in several randomized controlled studies. Similarly, a recent placebo-controlled study investigating anhedonia, another core symptom of depression, has determined that a single ketamine injection for TRD can significantly reduce the level of anhedonia in 40 min. The prefrontal cortex (PFC) and the subcortical and limbic brain regions (e.g., the amygdala) are critically involved in the neurocircuitry of depression. Enhanced activity in response to negative stimuli in the emotion-processing subcortical and limbic regions and deficiencies in the top-down regulatory activities of various parts of the PFC are the characteristic features of major depressive disorder. TRD patients have a hypoactive PFC and a hyperactive amygdala; this functional dysregulation between the PFC and the amygdala has been shown to persist even after a patient's depression has been temporarily controlled with aggressive treatment. Recently, it has been further demonstrated that pronounced hypofrontality has a key role in the neuropathology of TRD. However, it remains unclear whether the rapid antidepressant mechanisms of ketamine on TRD involve changes in glutamatergic neurotransmission in the PFC. This randomized, double-blind, placebo-controlled study utilized 18 F fluorodeoxyglucose positron emission tomography ( 18 F-FDG-PET) to investigate cerebral glucose uptake at baseline levels (215 to 0 min) and following ketamine or placebo injection (45-60 min after the injection). The experiment was designed to capture the earliest and perhaps the most important (based on the aforementioned preclinical and clinical studies) biological effects of ketamine on the brain. Notably, 18 F-FDG-PET measures the uptake of glucose primarily metabolized by astrocytes in response to glutamate release from neuronal cells. The 18 F-FDG signals thus provide a proxy for the quantification of glutamatergic neurotransmission in the human brain. This study was designed to investigate the rapid effects of ketamine, specifically on the glutamatergic system of mood circuitry in TRD. We hypothesized that the rapid antidepressant effects of lowdose ketamine involve PFC activation. Activity in the amygdala, which is a core region of the limbic system that is involved in the mood-related neurocircuitry, was also investigated.
Create a free account to open full-text PDFs.
Berman, R. M., Cappiello, A., Anand, A. et al. · Biological Psychiatry (2000)
Lally, N., Nugent, A. C., Luckenbaugh, D. A. et al. · Translational Psychiatry (2014)
Nutt, D. J. · Journal of Psychopharmacology (2015)
Murrough, J. W., Iosifescu, D. V., Chang, L. C. et al. · American Journal of Psychiatry (2013)
Papers in Blossom that reference this study
Matsingos, A., Wilhelm, M., Noor, L. et al. · Frontiers in Psychiatry (2024)
Li, W. C., Chen, L. F., Su, T. P. et al. · Journal of Affective Disorders (2023)
Muthukumaraswamy, S., Forsyth, B., Lumley, T. · Expert Review of Clinical Pharmacology (2021)
Bahji, A., Vazquez, G. H., Zarate, C. A. · Journal of Affective Disorders (2021)
Kohtala, S., Alitalo, O., Rosenholm, M. et al. · Pharmacology and Therapeutics (2020)
Su, T. P., Chen, M. H., Li, C. T. et al. · Neuropsychopharmacology (2017)
Lener, M. S., Niciu, M. J., Ballard, E. D. et al. · Biological Psychiatry (2017)
All 48 randomised participants (16 per group) completed the first-stage protocol. Baseline demographics, illness duration, comorbidities and baseline HDRS-17 scores were similar across groups. Clinically, both active ketamine groups showed superior rapid antidepressant responses compared with placebo, with significantly more responders in Groups A and B than Group C at 40 and 80 minutes (P < 0.05). The largest mean HDRS-17 improvement occurred at 40 minutes post-infusion for the active ketamine arms (post hoc: 0.5 mg/kg > placebo; 0.2 mg/kg > placebo). BPRS positive-symptom scores did not differ between groups over follow-up. Adverse events were described as mild and self-limiting. Group A (0.5 mg/kg) reported more floating sensations than Group B (0.2 mg/kg); otherwise side-effect rates were similar and no additional medical treatment was required. The presence of floating sensations did not associate with larger PFC SUV changes. ROI analysis showed a significant group-by-time interaction for the PFC (F = 7.373, P = 0.002): both ketamine doses produced increases in PFC SUV that were not observed in the placebo group. Reported effect sizes for PFC change were Cohen's d = 0.316 for Group A and d = 1.019 for Group B. For the amygdala there was no group or interaction effect, but a strong main effect of time (F = 40.695, P < 0.001), with SUV decreasing after treatment across all three groups. Changes in PFC SUV correlated with clinical improvement (percent HDRS-17 change) (r = −0.433, P = 0.002), whereas amygdala SUV changes did not correlate with symptom change (r = 0.070, P = 0.705). PFC and amygdala SUV differences were negatively correlated in the ketamine groups, consistent with reciprocal regulation of these regions; this correlation was not present in the placebo group. Whole-brain voxel-wise analyses confirmed a significant main effect of group in the PFC and additionally identified group effects in the supplementary motor area, dorsal anterior cingulate cortex (dACC), and bilateral post-central gyrus. Post hoc contrasts indicated greater post-treatment increases in SUV in the PFC, SMA and parts of the post-central gyrus for both ketamine groups versus placebo. The extracted text reports that Group A showed larger SUV increases than Group B in the dACC and bilateral post-central gyrus; the reporting on right PFC differences between doses in the extracted text is inconsistent and not clearly presented. Exploratory paired tests showed bilateral PFC increases and decreases in limbic structures (amygdala, hippocampus, parahippocampus) and cerebellum in the ketamine groups; the placebo group exhibited increases in posterior cortical regions and decreases in limbic and brainstem areas. Linear regression identified the change in PFC SUV as the strongest predictor of antidepressant response at both 40 minutes (b = −0.403, P = 0.031) and 240 minutes (b = −0.549, P = 0.008). Being in a ketamine treatment group also predicted response at 40 minutes (b = 0.305, P = 0.047) but not at 240 minutes. Changes in amygdala SUV, age and sex were not significant predictors in these models.
Li and colleagues interpret their findings as evidence that rapid antidepressant effects of low-dose ketamine in TRD are associated with facilitation of glutamatergic neurotransmission in the PFC. Both ROI and voxel-wise analyses converged on the PFC as a ketamine-sensitive locus, and the magnitude of PFC SUV increase related to clinical improvement within the first day. The observed negative correlation between PFC and amygdala SUV changes in the ketamine groups is taken to support a functional reciprocity in PFC–amygdala circuitry, whereby augmentation of PFC activity may help suppress limbic hyperactivity implicated in depression. The authors position these results within prior imaging and preclinical work showing ketamine-induced PFC changes and downstream synaptic mechanisms; they note consistency with earlier PET and functional studies in healthy volunteers and psychiatric samples. Placebo-associated brain changes differed from ketamine effects, with the placebo group showing posterior cortical activations and amygdala decreases that the authors suggest may reflect expectancy and perceptual processing rather than ketamine-specific glutamatergic modulation. In addressing potential confounds, the investigators argue that psychotomimetic side effects do not account for the PFC findings: side effects were mostly mild, 0.2 mg/kg did not produce more adverse effects than placebo, and PFC changes were similar across ketamine doses and independent of floating sensations. They also cite preclinical timing data (mTOR signalling and synaptic changes) that are temporally compatible with rapid PFC effects. Key limitations acknowledged by the authors include the add-on design (participants remained on their existing medications, so observed effects may reflect interactions with background treatments), the possible suboptimal sensitivity of the HDRS-17 for very short-term symptom change, and the fact that the present report covers only the first-hour stage—whether PFC activation relates to sustained antidepressant effects beyond the acute period remains unresolved and requires further study.
The double-blind, placebo-controlled data reported by Li and colleagues indicate that low doses of ketamine produce rapid antidepressant effects in treatment-resistant depression that are associated with increased glutamate-linked glucose metabolism in the prefrontal cortex. The authors conclude that facilitation of PFC glutamatergic neurotransmission is a key component of ketamine's early antidepressant mechanism.
Eligible subjects were adult patients between 21 and 65 years of age with a Diagnostic and Statistical Manual of Mental Disorders-IV-Text Revision (DSM-IV-TR) diagnosis of major depressive disorder. To ensure that patients were resistant to treatment medications, all TRD patients were required to have a history of failing to respond to at least three different antidepressants with adequate dosage and treatment duration and of failing at least one trial of adequate antidepressant treatment during their current depressive episode. The recruited TRD patients did not have major medical or neurological illnesses or a history of alcohol or substance abuse. Other exclusion criteria are described in the Supporting Information. This study was performed in accordance with the Declaration of Helsinki and was approved by the local Ethics Review Committee. Informed consent was provided by all of the participants.
Patients underwent a detailed psychiatric and medical history-taking, a diagnostic interview, and brain structural imaging by magnetic resonance imaging (MRI) at baseline. Following screening and baseline ratings, a subsequent 40min acute treatment phase was conducted, which involved constant intravenous infusion of ketamine or placebo via an infusion pump. The study involved two stages, as it aimed to study the rapid effects of ketamine on the human brain in the first hour after ketamine administration and in the third day. In this article, we report the results of the first stage; the second stage is still ongoing, and its results will be reported elsewhere. For the first stage, a total of 48 r Ketamine Modulates the PFC r r 1081 r TRD patients were randomized in a 1:1:1/A:B:C ratio to each of the respective experiment groups: A: 0.5 mg/kg, B: 0.2 mg/kg, and C: normal saline (Supporting Information Fig.). Two low-dose ketamine regimens were selected to test and to confirm our hypothesis because both 0.5 mg/kg and 0.2 mg/kg have been shown to be effective in the treatment of TRD. The entire study procedure was double-blinded, and participants and staff were all blind to the treatment assignment. An independent research nurse who was not involved in the study was responsible for the random allocation and was not allowed to release any information to others, including the study nurse who applied the intravenous injection. Depressive symptoms were rated using the 17-item Hamilton Depression Rating Scale (HDRS-17)at baseline (immediately before the first 18 F-FDG-PET scan) and at 40, 80, 120, and 240 min post-ketamine administration. The positive symptoms subscale of the brief psychiatric rating scale (BPRS) was used] to measure potential psychotomimetic effects of ketamine at each time point. The present study primarily focused on the rapid antidepressant responses and neuroimaging findings during the first day of treatment. Primary outcomes, such as the percent change between baseline and 40-min HDRS-17 scores, were correlated with imaging results. Responders were defined as having at least a 50% decrease in their HDRS-17 score from baseline. Side effects following injections, including floating sensations, dissociative symptoms, dizziness, nausea, and somatic discomfort, were recorded.
The MR images were acquired with a 3.0 GE Discovery 750 whole-body high-speed imaging device and highresolution structural T1-weighted images were acquired. Two volumes of 18 F-FDG PET scans (i.e., pre-and postketamine injections; the brain acquisition time for each PET volume is 15 min) of at rest glucose utilizations were acquired on a PET/CT scanner (Discovery VCT; GE Healthcare, USA) with 3D brain mode in the morning (9000 h to 1200 h). The 1st, 15-min PET scan was acquired 45 min after an intravenous (IV) injection of about 222 MBq of 18 F-FDG. Subsequently, another IV bolus of approximately 222 MBq of 18 F-FDG was injected. At 5 min after the 2nd FDG injection, a 40-min IV infusion of normal saline or low dose ketamine (0.2 mg/kg or 0.5 mg/kg of ketamine) was given in a resting state. After 40 min of rest (staying awake in a deemlight room), the 2nd, 15-min PET scan was acquired. Other details regarding MRI and PET acquisition please refer to the Supporting Information.
The PET data were analyzed using Statistical Parametric Mapping version 8 software (SPM8; Wellcome Department of Cognitive Neurology, Institute of Neurology, University Col-lege London, London, England) implemented in Matlab 7.1 (The Mathworks Inc., Sherborn, MA). A group-specific MRIaided 18 F-FDG template was createdand used to normalize each subject's PET images, followed by smoothing with a 3D Gaussian kernel (full width half maximum [FWHM] 5 8 mm). The smoothed and normalized PET images in the standardized brain space were created and subjected to further analysis. Because we were primarily interested in the relative changes in brain metabolic activity in the first hour after IV infusion treatment (drug or placebo), the standardized uptake value (SUV), which is a validated semi-quantitative method, was used to correct for FDG activity at the time of injection for each of the two FDG scans (before and 40 min after injections). Specifically, the parametric whole-brain SUV image was calculated by dividing the concentration of radiotracer activity in the normalized PET image (MBq/kg) by the FDG dose at the injection time (MBq) divided by the body weight (kg). The total FDG dose at the 2nd injection time was approximated with consideration of the residual dose from the 1st FDG injection.
To test our hypothesis, region of interest (ROI) analysis was performed. The mean SUV in the a priori regions (PFC and amygdala) was extracted from unsmoothed SUV images in the standard stereotactic space by using PMOD version 3.0 (PMOD Technologies, Zurich, Switzerland) as described previously. Normalized SUV values (values were normalized by dividing the PFC mean SUV by the global mean SUV) were reported. The delineations of the PFC and amygdala were performed by an automated anatomical labeling template (AAL template) to prevent bias from inter-or intra-rater reliability issues that arise in manual delineation. The delineated PFC combined ten anatomical regions of the AAL template (i.e., F1, F2, F1O, F2O, F3T, F3O, F3OP, GR, F1MO, and F1M) in bilateral hemispheres to maximize the coverage of the PFC; this delineation of the PFC covered the clusters that are most involved in depression and TRD, including the dorsolateral PFC.
Voxel-wise analysis was performed using SPM8 to confirm the results of the ROI analysis and to explore any regional changes that were specific to ketamine treatment. We compared the between-group SUV changes by twoway ANOVA, with time (i.e., the 1st and 2nd SUV images) as the within-subject factor, ketamine groups (i.e., A: 0.5 mg/kg, B: 0.2 mg/kg, and C: normal saline) as the between-subject factor, and age and sex as covariates of no interest. The group-by-time interaction and the main r Li et al. r r 1082 r effects of group and time that were significant at P < 0.05 after voxel-level family-wise error (FWE) correction were reported. For the post hoc analysis, independent t-tests were performed for the between-group comparisons. The statistical thresholds were set at P < 0.0167 (50.05/3, corrected for the three groups). For exploratory purposes, we conducted a separate paired t-test to visualize the relative changes between the 1st and 2nd SUV images in the three groups, with the significance thresholds set at a voxellevel uncorrected value of P < 0.05 and cluster-level FWEcorrected value of P < 0.001.
Statistical analysis of demographic and clinical data was performed using SPSS 16.0 (SPSS Inc, Chicago, IL). Oneway analysis of variance (ANOVA) (or Student's t-test) and Fisher's chi-square test (or Yate's correction) were used to compare the continuous and categorical variables between groups (i.e., ketamine groups). A value of P < 0.05 was used to indicate statistical significance. To investigate the relationship between improvements in depression (i.e., percent change in HDRS-17 scores after treatment) and differences in before-versus-after SUV in the PFC and amygdala, Pearson's correlation analysis was performed, and the correlation coefficient (r) was reported. P < 0.05 (2-sided tests) was considered to be statistically significant. To determine whether any changes in the measures of the PFC and amygdala (i.e., SUV of the PFC and amygdala) were a result of the interaction between time (i.e., before and after ketamine treatment) and ketamine group (i.e., A, B, and C), two-way repeated measures ANOVA was conducted, with time as the within-subject factor and ketamine groups as the between-subject factor, whereas age and sex were treated as covariates. The interaction between group and time and the main effects of group and time were reported, with P < 0.05 (2-sided test) considered to be statistically significant. The Least Significant Difference (LSD) method was used for post hoc analyses. Finally, linear regression was performed, with age, sex, baseline HDRS-17 scores, before-versus-after SUV changes in the PFC and amygdala, and ketamine groups (low-dose ketamine or placebo group) included as independent factors. The percent change in the HDRS-17 score at 40 min and 240 min after treatment was the dependent variable. A P-value < 0.05 (2-sided test) was deemed statistically significant.
All 48 subjects (16 subjects in each group) were recruited from December 2012 to April 2014 and participated in the entire study. Baseline demographic and clinical features (i.e., age, gender, duration of illness, psychiatric comorbidities, and HDRS-17 scores) were similar among the three groups (Table). The active ketamine infusion groups (Groups A and B) exhibited better antidepressant responses immediately fol-lowing ketamine treatment than the placebo group (Group C) (Table). There were significantly more responders in the active ketamine infusion group (Groups A and B) than in the control group (Group C) at 40 min and 80 min following ketamine treatment (P < 0.05). The results of the mean HDRS-17 changes indicated that the best outcome was at 40 min after active ketamine injection (40 min vs. baseline, Table) (P < 0.05; post hoc analysis: 0.5 mg/ kg>placebo, 0.2 mg/kg > placebo). There was no between-group significance found regarding changes in BPRS positive symptoms during the followup (Table). The side effects following ketamine treatment were all mild and self-limiting and required no additional medical treatment. However, statistically, more cases of floating sensation occurred in Group A (0.5 mg/ kg) than in Group B (0.2 mg/kg) (Table).
The ANOVA model for the PFC showed a significant group 3 time interaction (F 5 7.373, P 5 0.002), in which both of the low-dose ketamine groups were more associated with PFC activation than the placebo group (Fig.). The ANOVA analysis did not reveal a significant main effect of group (F 5 0.223, P 5 0.793) or a significant interaction between group and time (F 5 0.776, P 5 0.467) for the amygdala; however, there was a significant main effect of time on the SUV of the amygdala (F 5 40.695, P < 0.000) (Fig.). In brief, the SUV in the PFC increased only in the active ketamine groups (i.e., Group A and B; Effect size for Group A: Cohen's d 5 0.316, for Group B: 1.019) but not in the placebo group (Group C), whereas the SUV in the amygdala decreased significantly after treatment in all three groups (Fig.). The antidepressant outcomes, which were based on the changes in the HDRS-17 scores, correlated well with differences in the SUV in the PFC (r 5 20.433, P 5 0.002) (Fig.) but not in the amygdala (r 5 0.070, P 5 0.705) (Fig.). The PFC changes did not differ between those with and without floating sensations (with vs. without: 0.0186 6 0.0092 vs. 0.0198 6 0.0156, t 5 20.210, P 5 0.835). Furthermore, the SUV differences in the PFC were negatively correlated with the differences in the amygdala (Fig.), which supports a functionally reciprocal relationship between the PFC and the amygdala brain circuit in depression. This functional regulation in the PFC-amygdala circuitry was observed in patients who received active ketamine treatment, but not in those who received a placebo (Fig.).
The results of the whole-brain analysis showed that there was a significant main effect of the group on the PFC, the supplementary motor area (SMA), the dorsal anterior cingulate cortex (dACC), and the bilateral post-central gyrus (PCG) (Fig., boxed on the left). The post hoc analyses showed that after treatment, both of the ketamine groups had significantly more increases in SUV in the PFC, SMA, and parts of the PCG compared with the placebo group. However, Group A (0.5 mg/kg) had a significantly greater increase in SUV in the dACC and bilateral PCG than Group B (0.2 mg/kg) (Fig.; 0.5 > 0.2 mg/ kg) and a significantly lower increase in SUV in the right side of the PFC (Fig.; 0.2 < 0.5 mg/kg). The ANOVA analysis did not reveal a significant main effect of time or a significant interaction between group and time. The results of the group-by-time analysis confirmed that the PFC was the main locus of low-dose ketamine IV treatment. The exploratory results revealed significant increases in SUV in the bilateral PFC, and decreases in SUV in the amygdala, parahippocampus, hippocampus, and cerebellum were observed in both the 0.5 mg/kg (Supporting Information Fig.) and 0.2 mg/kg (Supporting Information Fig.) ketamine groups (Supporting Information Table). Group A also demonstrated an increase in SUV in the posterior cingulate cortex, precuneus, and thalamus. In the placebo group, significant increases in SUV were observed in the precuneus, posterior cingulate cortex, parietal cortex, temporal cortex, and occipital cortex, whereas significant decreases in SUV were found in the amygdala, hippocampus, cerebellum, and pons (Supporting Information Fig.and Table).
Finally, linear regression models demonstrated that the most significant factors predicting antidepressant effects 40 min after administration of ketamine were the changes in the SUV in the PFC (b 5 20.403, P 5 0.031) and being in a ketamine group (b 5 0.305, P 5 0.047), whereas the SUV changes in the amygdala (P 5 0.182), age (P 5 0.149), and sex (P 5 0.111) were not significant. Similarly, the most significant factor predicting antidepressant effects after 240 min was the change in SUV in the PFC (b 5 20.549, P 5 0.008), but not the ketamine group (P 5 0.974), SUV changes in the amygdala (P 5 0.170), age (P 5 0.370), and sex (P 5 0.132).
Our study was a randomized placebo-controlled study that investigated the effects of ketamine on brain mechanisms during the first hour of treatment. The rapid changes in glutamate neurotransmission following the NMDA-antagonist ketamine treatment were investigated in a double-blind randomized study design. As hypothesized, the ROI analysis and the voxel-wise analysis found that the rapid antidepressant effects of low-dose ketamine involved activation of the PFC. The evidence supporting this hypothesis included the finding that the increase in SUV of the PFC was specific to the low-dose ketamine treatment and that it was significantly related to the rapid antidepressant responses in the first day after ketamine treatment.
Although most previous studies administered intravenous ketamine at a dose of 0.5 mg/kg, some research has demonstrated that other dosing regimens, such as 0.2 mg/ kg, are also effectivein treating TRD. Consistent with the existing evidence from western countries, we found that both of the ketamine doses (0.2 and 0.5 mg/kg) resulted in rapid antidepressant effects during the first one to two hours (Table); the long-term antidepressant effects of ketamine warrant further studies. Most importantly, our neuroimaging data suggested that increased glutamatergic neurotransmission in the PFC was critical to the ketamine-specific rapid antidepressant effect. The finding that both of the active ketamine groups, but not the placebo group, demonstrated an increase in the PFC SUV supports this claim as this increase was well correlated with antidepressant response (Fig.). It is important to note that the 18 F-FDG-PET signals could be considered to be a specific proxy measurement of glutamate neurotransmission. The PET signals were used to investigate the rapid effects of ketamine on the glutamatergic system of mood circuitry in TRD. Structural and functional deficits of the PFC have been found to be associated with TRD. Higher degrees of PFC abnormalities, as observed in TRD patients, may result in the failure to control limbic hyperactivity (e.g., ROI analysis showing the effects of ketamine on the PFC and the amygdala among different groups. A: Normalized PFC glucose uptake (PFC SUV/whole-brain SUV) was significantly elevated in the 0.5 mg/kg and 0.2 mg/kg groups but not in the placebo group. The ROI analysis demonstrated a significant increase in the amygdala in all three groups. The bars are presented as the mean (standard deviation [SD]). B: The PFC SUV changes after treatment; the changes were significantly correlated with improvement in depression as documented by percent changes in HDRS-17 scores. C: No correlation existed between the amygdala SUV changes and clinical improvement. Pvalues of less than 0.05 were considered to be statistically significant (*P < 0.05, **P < 0.005).
A recent open-labeled study reported that ketamine could change glucose metabolism in various parts of the PFC in TRD patients. Augmenting PFC activity (e.g., by non-invasive brain stimula-tion) has long been sought as an option for treating TRD. Interestingly, the present study found that 0.5 mg/kg of ketamine was not associated with a greater increase in PFC activity than Correlation between SUV changes in the PFC and the amygdala. This correlation was observed most significantly in the patients receiving active ketamine treatment (bottom left figure) (P < 0.05). However, this correlation was not observed in the placebo group, suggesting that the antidepressant mechanisms of the placebo differ from the antidepressant mechanisms of ketamine. 0.2 mg/kg, supporting the notion that the antidepressant action of ketamine occurs only in a small-and low-dose range. Our major finding, that the PFC is involved in ketamine's rapid antidepressant effects, was not unexpected as a single-dose intravenous injection of low-dose ketamine (0.3 mg/kg) had previously been shown to induce PFC activation and cerebellar deactivation within 20 min in healthy individuals. Additionally, a placebo-controlled neuroimaging study reported that low doses of ketamine resulted in rapid changes in glucose metabolism in many mood-related brain areas in patients with bipolar disorders. In another 18 F-FDG PET study with 17 healthy volunteers, a metabolic increase in the PFC was also observed during the infusion of a sub-anesthetic dose of ketamine. The ketamine-induced PFC functions may, in turn, lead to a decrease in limbic hyperactivity and thus improve depressive symptoms. This is supported by the fact that the changes in the SUV of the PFC were negatively correlated with SUV changes in the amygdala, which was observed only in the ketamine groups (Fig.). The results from previous neuroimaging investigations have demonstrated that beliefs and expectations can markedly regulate functional activity in brain regions that are involved in perception and various aspects of emotion processing. In the placebo group (Fig.and Supporting Information Fig.), no prominent PFC changes were observed, whereas a decrease in the SUV in the amygdala and an increase in the brain regions involved in the sensory and visual pathways, such as the occipital cortex, the temporal cortex (the ventral stream of the visual pathway), the parietal cortex (sensory perception and the dorsal stream of the visual pathway), and the posterior cingulate cortex/precuneus (associated with awareness and behavioral adaptation), was found. It has been reported that the most important brain change involved in the placebo effects of emotional processing is the functional decrease in the bilateral amygdala. In addition, perceptions of drug preparations could play an important role in the psychological expectancy and placebo responses. Voxel-wise analysis showing a significant main effect of ketamine among the different groups and the post hoc analysis. Left panel: The significant main effect of the group was on the PFC, the supplementary motor area (SMA), the dorsal anterior cingulate cortex (dACC), and the bilateral post-central gyrus (PCG). Right panel: The post hoc analyses showed that both of the ketamine groups had significantly greater increases in the SUV in the PFC, SMA, and parts of the PCG than the placebo group. The 0.5 mg/kg group (Group A) had a significantly greater increase in the SUV in the dACC and bilateral PCG than the 0.2 mg/kg group (Group B), but had a significantly lower increase in the SUV of the right-side of the PFC than the 0.2 mg/kg group. The significance thresholds for voxel-wise analysis were set at a voxel-level family-wise error (FWE)-corrected level of P < 0.05 by two-way group-by-time ANOVA. No significant main effect of time and no significant interaction between group and time were found. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] should be associated with a higher perceived strength of treatment compared to oral intake of conventional antidepressant pills, the findings related to changes in the posterior brain regions (e.g., occipital, temporal, parietal cortex) could be attributed to individual expectation and memory.
Ketamine on Imaging Findings? As we studied the rapid effects of ketamine (i.e., 45-60 min after the injection) on brain glucose metabolism, there could have been potential psychotomimetic effects that occurred following ketamine injection. Some side effects, such as floating sensations, dissociative symptoms, dizziness, nausea, chest discomfort, and crying, were reported in the present study. However, the only difference between groups was a higher incidence of a floating sensation in Group A, the 0.5 mg/kg ketamine group. The 0.5 mg/kg group was also associated with a greater increase in the SUVs in the bilateral sensory cortex and the attention system, the dorsal ACC, than the placebo and 0.2 mg/kg group, which may partly explain the occurrence of the side effects in the 0.5 mg/kg group (Fig.). Group B, the 0.2 mg/kg group, was not associated with more side effects than the placebo group. As the PFC changes and the PFC-amygdala correlation were identified in both the 0.5 mg/kg and 0.2 mg/kg ketamine groups, it is plausible that the observed clinical effects and the associated changes in the PFC in those receiving active ketamine treatment were not caused by direct psychotomimetic effects. In fact, patients who experienced more side effects (e.g., floating sensation) did not experience greater increases in the SUV of the PFC than those who did not experience side effects. BPRS changes over time were also similar across the three groups. A singledose intravenous injection of low-dose ketamine (0.3 mg/ kg) has previously been reported to induce PFC activation within 20 min in healthy individuals, and this finding was not simply induced by psychotomimetic effects of ketamine. The results from preclinical data also support the belief that the effect of ketamine on the PFC is potentially the crux of its functioning as an antidepressant; the rapid antidepressant effects of ketamine have been found to be dependent on the mammalian target of rapamycin, which peaks at 30-60 min following ketamine injection and leads to increased synaptic signaling proteins and an increased number and function of new spine synapses in the PFC of rats within 24 h.
Some limitations should be considered in the interpretation of our study results. First of all, despite its randomized and placebo-controlled design, the current study could be considered an add-on ketamine study because the original medication regimen that the TRD patients had failed to respond was not discontinued during the ketamine treatment and PET scans. Therefore, the observed responses to ketamine could be the result of a combination or a regulation of the effects of the medications that the patients were already using. However, because we have previous solid data to support the antidepressant effects of ketamine among Asian patients with TRD, the add-on design was more ethically appropriate and could provide more naturalistic data. Additionally, the study's main finding of the ketamine-induced increases in the PFC SUV occurred within the first hour of ketamine treatment, whereas the concomitant medications were not altered. Furthermore, there were no differences in the medications used across all groups (Table). Second, although the Hamilton depression rating scale has been used in many of the previous ketamine studies, it may not be ideal for rating shortterm changes in depression. Third, the results of the present study provide evidence to support the involvement of PFC activation on initial antidepressant effects of low-dose ketamine in human TRD subjects. However, whether the increase in PFC activity is also related to the antidepressant effects of ketamine at other time points (e.g., day three) or to the persistent resolution of TRD remains elusive and requires further research.
The results of this double-blind placebo-controlled study provide evidence that the rapid antidepressant mechanisms of low doses of ketamine involve the facilitation of glutamatergic neurotransmission in the PFC.