Acute ketamine challenge increases resting state prefrontal-hippocampal connectivity in both humans and rats

Aberrant functional coupling between the prefrontal cortex and hippocampus (PFC-HC) has been implicated in schizophrenia and in individuals at clinical high risk. Resting-state functional magnetic resonance imaging (rs-fMRI) studies indicate reduced PFC-HC connectivity in chronic schizophrenia, while task-based studies sometimes show increased coupling. Advances in rodent functional imaging have made it possible to probe homologous PFC-HC circuits across species, but it remains unclear whether the same pharmacological manipulation produces comparable changes in humans and rats. Grimm and colleagues therefore tested whether an acute ketamine challenge—commonly used as a pharmacological model of NMDA receptor hypofunction—modulates resting-state PFC-HC connectivity in a consistent manner across species. The primary aim was to compare rs-fMRI measures of PFC-HC coupling after subanesthetic ketamine in healthy human volunteers and in Sprague-Dawley rats, to evaluate the potential of this circuit phenotype as a translational biomarker for psychiatric research.

Human participants were 24 healthy adults (12 females, mean age 25 years, mean weight 70 kg) who completed a subject- and observer-blind, placebo-controlled, randomized three-period crossover study. Each subject received saline (placebo), ketamine (0.5 mg/kg), and scopolamine in counterbalanced order across three sessions; for this report only the ketamine versus placebo data were analysed. Infusions were delivered by an intravenous pump; resting-state fMRI began approximately 20 minutes after the end of the ketamine infusion. Blood samples for ketamine and norketamine were taken at 10 and 80 minutes after infusion end and analysed by mass spectrometry. Eighteen male Sprague-Dawley rats (373–447 g) were allocated to two groups (N=9 per group) and received either S-ketamine (25 mg/kg subcutaneously) or saline vehicle. The order of injections and time of day were randomized. rs-fMRI acquisition started 30 minutes after injection. At the end of the experiment (within 50–80 minutes post-injection) blood was collected for measurement of ketamine and norketamine. Human fMRI data were acquired on a 3-T scanner (EPI: TR=1790 ms, TE=28 ms, 34 slices, 332 volumes). Rat imaging used a 9.4-T system with medetomidine sedation following brief isoflurane, and EPI parameters optimised for the rat brain (TR/TE 1700/17.5 ms, 29 coronal slices, 300 acquisitions over 8.5 min). Preprocessing in humans used SPM8 and the CONN toolbox: realignment, slice-timing correction, normalization to MNI space, smoothing (8 mm FWHM), band-pass filtering (0.01–0.1 Hz), motion regression and aCompCor nuisance regression. For humans, seed regions were right and left DLPFC (BA9/46), and hippocampal target masks were from the Harvard-Oxford atlas (50% probability threshold). Second-level statistics used flexible factorial models in SPM8; drug and order were modelled as fixed effects and ketamine versus placebo was the primary contrast. ROI-based family-wise error (FWE) correction at P FWE < 0.05 was applied. Rat data processing also used SPM8 with FieldMap correction, motion regression, physiological noise removal, slice-timing correction, band-pass filtering (0.01–0.1 Hz) and spatial normalisation to a rat template. Seed time courses were extracted from left and right prelimbic cortex (PrL), mean-seeded, smoothed (0.8 mm) and voxel-wise correlations were Fisher Z-transformed. Second-level analyses comprised two-sample t-tests (25 mg/kg versus saline) with hippocampal masks and cluster- and peak-level FWE correction (P FWE < 0.05). The extracted text notes that lower ketamine doses in rats (5 and 10 mg/kg) were tested with identical procedures but produced no significant PrL–hippocampal connectivity effects (data not shown).

Across species, acute ketamine increased PFC–hippocampal functional connectivity measured with rs-fMRI. In humans, ketamine significantly increased the correlation between DLPFC seeds and the left hippocampus. Two reported peaks surviving left-hippocampal ROI correction were: T=3.69, P FWE = 0.027 (x= -30, y=-30, z=-14) and T=3.97, P FWE = 0.013 (x=-26, y=-10, z=-22). The extracted text also states that other DLPFC–hippocampus pairings showed no significant ketamine effect (P FWE > 0.05), though the laterality descriptions in the extraction are not fully consistent. Plasma concentrations (within-subject average) were ketamine 364.88 ng/ml (SD=228.48) and norketamine 43.24 ng/ml (SD=16.1); no significant correlations were observed between these plasma levels and DLPFC–hippocampus connectivity at the reported peak voxels (all P > 0.19). In rats, ketamine significantly elevated correlations between the left PrL and both left and right hippocampus, and between the right PrL and the left hippocampus; no significant effect was found for right PrL to right hippocampus. Measured plasma levels in ketamine-treated rats were ketamine 1693.87 ng/ml (SD=475.92) and norketamine 1333.01 ng/ml (SD=585.78), with a high correlation between ketamine and norketamine levels (r=0.712). No statistically significant correlations were found between plasma concentrations and the connectivity measures at clusters surviving FWE correction (for the strongest rat effect, P = 0.215 for ketamine; P = 0.241 for norketamine). The extracted text also reports that lower rat doses (5 and 10 mg/kg) did not produce significant PrL–hippocampal connectivity changes under identical analysis pipelines.

Grimm and colleagues interpret the parallel increase in PFC–hippocampal coupling after acute ketamine in humans and rats as evidence of a robust cross-species pharmacological effect, supporting the use of rs-fMRI as a translational biomarker. They consider NMDA receptor antagonism with consequent disinhibition of GABAergic interneurons and increased extracellular excitatory neurotransmitters (for example glutamate) as plausible mechanisms underlying the observed hyperconnectivity. Supporting observations include ketamine-induced increases in gamma oscillations and regional metabolic and blood-flow changes in hippocampal and cortical regions, which relate to BOLD connectivity measures. The authors note the ketamine-associated increase was more prominent in the dorsal hippocampus across species, consistent with higher NMDA receptor expression along the dorsal–ventral hippocampal axis. Animal work is cited indicating altered directional drive between medial prefrontal cortex and dorsal CA1 after ketamine, which could imply hippocampal-originating contributions to the increased coupling observed. They also discuss how changes in AMPA-mediated plasticity and LTP in the PFC–HC circuit may relate to cognitive impairments seen after acute ketamine. Several species- and protocol-related caveats are acknowledged. Rat imaging required sedation (medetomidine following isoflurane), although the investigators argue medetomidine preserves global network properties; rats received a single subcutaneous bolus of S-ketamine at higher plasma exposure than humans who received a racemic ketamine infusion. Because only a single dose per species was studied, a quantitative exposure–response alignment across species was not possible and the absence of an exposure–response correlation may reflect this limitation. Lateralisation was discussed: humans showed left-hippocampal lateralisation of the ketamine effect while rats showed bilateral hippocampal responses, and the authors propose differences in receptor distribution, connectivity or species-specific lateralisation as possible explanations. Finally, the authors contrast their findings with resting-state reductions in PFC–HC connectivity reported in chronic schizophrenia. They suggest acute ketamine may model a hyperglutamatergic or prodromal-like state rather than chronic disease, and caution against equating ketamine effects directly with schizophrenia. Future directions proposed by the authors include multimodal PET–fMRI to measure local NMDA receptor density, broader graph-theoretic mapping of ketamine-induced connectivity changes across networks, and longitudinal studies probing the transition from prodromal to chronic illness.

The study concludes that acute ketamine induces comparable systems-level increases in PFC–hippocampal coupling in humans and rats. Grimm and colleagues argue that these cross-species pharmaco-fMRI effects support the utility of homologous fMRI measures as translational biomarkers and may aid the development of novel treatment strategies by providing a drug-responsive circuit phenotype relevant to psychiatric disorders.