Rumination, characterized by repetitive negative thinking that is difficult to disengage from, is a prominent feature of major depressive disorder and a transdiagnostic risk factor for psychopathology. Neuroimaging research has linked rumination to altered dynamics among large-scale brain networks, particularly the default mode network (DMN), which supports self-referential processing, and the frontoparietal network (FPN), which supports cognitive control. Previous studies have associated repetitive negative thinking with greater time spent in a hybrid state involving both networks. Ketamine produces rapid antidepressant effects and has been reported to reduce rumination, but the neural mechanisms underlying these effects remain incompletely understood. This study investigated whether ketamine acutely alters occupancy of a rumination-related DMN–FPN coactivation pattern and whether pretreatment with lamotrigine, a putative inhibitor of presynaptic glutamate release, attenuates these effects.

In this randomized, double-blind, placebo-controlled fMRI study, 75 healthy adults aged 18 to 45 years with no current or past psychiatric disorders were assigned in a 1:1:1 ratio to receive placebo followed by placebo, placebo followed by ketamine, or lamotrigine followed by ketamine. Lamotrigine 300 mg or matching placebo was administered orally two hours before scanning. Participants in the ketamine groups received intravenous racemic ketamine at approximately 0.12 mg/kg during the first minute, followed by 0.31 mg/kg/hour for approximately 55 minutes. Resting-state fMRI data were collected before infusion, during infusion, and 24 hours after infusion. Whole-brain coactivation pattern analysis identified recurring brain states and quantified their occupancy, defined as the total number of imaging volumes assigned to each state, and persistence, reflecting consecutive volumes spent in that state. Trait rumination was assessed using the German version of the Response Styles Questionnaire during an optional assessment after trial completion. Following imaging exclusions, 62 participants were included in the primary neuroimaging analyses and 54 in analyses involving rumination.

Seven recurring coactivation patterns were identified. The pattern of primary interest was a hybrid state involving simultaneous activation of the DMN and FPN. At baseline, greater occupancy of this state was associated with higher self-reported rumination, although the estimate was imprecise and its confidence interval included zero (β = 0.272; 95% CI, −0.018 to 0.564).

During infusion, participants in both ketamine groups spent substantially less time in the hybrid DMN–FPN state than participants receiving placebo. Between-group effect sizes at this time point were large for ketamine without lamotrigine pretreatment versus placebo (Cohen’s d = −1.60; 95% CI, −2.84 to −0.87) and for lamotrigine plus ketamine versus placebo (d = −1.21; 95% CI, −1.96 to −0.63). However, both ketamine groups already had lower occupancy at baseline. Comparisons of baseline-to-infusion change between groups produced smaller, less conclusive estimates: d = −0.58 (95% CI, −1.25 to 0.03) for ketamine versus placebo and d = −0.40 (95% CI, −1.00 to 0.18) for lamotrigine plus ketamine versus placebo. Thus, the large differences observed during infusion should be distinguished from the more uncertain evidence for differential change from baseline.

Among participants receiving ketamine without lamotrigine pretreatment, higher trait rumination was associated with larger reductions in hybrid DMN–FPN state occupancy from baseline to infusion (β = −0.602; 95% CI, −1.180 to −0.334). Associations in the placebo and lamotrigine-pretreated groups were inconclusive, with confidence intervals including zero. These within-group findings do not establish that the strength of the association differed between treatment groups. Occupancy in both ketamine groups returned toward baseline by 24 hours, when no meaningful between-group differences were observed. Exploratory analyses also found reduced occupancy of a canonical somatomotor coactivation pattern during ketamine infusion, with recovery toward baseline at follow-up.

Lamotrigine pretreatment was associated with numerically smaller ketamine-related changes in both the DMN–FPN and somatomotor patterns. However, direct comparisons between the ketamine-only and lamotrigine-pretreated groups were imprecise and did not provide clear evidence that lamotrigine attenuated these effects. The findings therefore do not establish glutamatergic mediation. Additional analyses showed that associations and group differences were substantially attenuated after accounting for state persistence, suggesting that differences in occupancy were largely related to how long participants remained in these network configurations rather than simply how often they entered them.

These findings suggest that ketamine transiently alters occupancy of a brain state previously linked to rumination. The association between higher trait rumination and larger occupancy reductions in the ketamine-only group raises the possibility that this network state is relevant to ketamine’s previously reported antiruminative effects. However, the study involved healthy volunteers rather than patients with depression, and the reported analyses did not assess treatment-related changes in rumination or depressive symptoms. Trait rumination was measured during an optional post-trial assessment, with timing that varied across participants. Other limitations include the relatively small sample, baseline differences in state occupancy, potential functional unblinding, and possible contributions from ketamine’s psychotomimetic effects and changes in vigilance. Interpretation of lamotrigine as a selective probe of glutamatergic mechanisms is also uncertain.

The results should therefore be viewed as preliminary mechanistic evidence, not as validation of a clinical biomarker or demonstration that changes in this brain state mediate antidepressant response. Studies in patients with depression that directly measure network dynamics, rumination, and clinical outcomes are needed to establish the translational significance of these observations.

Reference: Meiering M et al. Transl Psychiatry. 2026;16(1). Abstract