SNIPPET: A single subanesthetic ketamine infusion (0.5 mg/kg) significantly reduced depressive, anhedonic, and ruminative symptoms in treatment-resistant depression patients within 24 hours, while broadly increasing resting state functional connectivity within the Default Mode and Frontoparietal brain networks — though symptom improvement correlated with pre-treatment connectivity, not post-infusion changes.

Ketamine Rewires Brain Connectivity in Treatment-Resistant Depression — But Not How You'd Expect

THE PROTOHUMAN PERSPECTIVE#

Treatment-resistant depression affects roughly 30% of everyone diagnosed with major depressive disorder. That's not a niche statistic — it means millions of people cycling through SSRIs, SNRIs, therapy modalities, and lifestyle interventions that simply don't land. For those of us interested in neural optimization, this study from Translational Psychiatry (2026) matters because it challenges a straightforward assumption: that ketamine works because it changes brain connectivity. The data suggests something more nuanced. Your brain's connectivity before treatment may matter more than how connectivity shifts after. That's a fundamentally different framing — one that points toward predictive biomarkers, not just reactive treatments. If we can identify who will respond based on pre-existing neural architecture, we move from trial-and-error prescribing toward genuine precision psychiatry. And that, more than any single drug, is what shifts the trajectory of human mental performance.

THE SCIENCE#

What Ketamine Does to Resting State Brain Networks#

Resting state functional connectivity (rsFC) is the synchronized activity between brain regions when you're not doing anything in particular — just existing. I think the word "resting" is doing too much work here, because these networks are anything but idle. The Default Mode Network (DMN) handles self-referential thinking, rumination, and autobiographical memory. The Frontoparietal Network (FPN) manages executive control, attentional switching, and cognitive flexibility.

In treatment-resistant depression, these networks talk to each other abnormally. Sometimes too much (the DMN stuck in a ruminative loop), sometimes too little (the FPN unable to intervene). This is well-established across multiple neuroimaging reviews ^[1][2].

The new study, published March 2026 in Translational Psychiatry, examined 24 participants with TRD (16 women; mean age 44.35 ± 15.86 years) who received a single IV infusion of racemic ketamine at 0.5 mg/kg ^[1]. Ninety-six-channel EEG was recorded 24 hours before and 24 hours after infusion. A healthy control group of 34 participants (25 women; mean age 32.49 ± 14.07 years) provided baseline comparison data without any intervention.

The Connectivity Paradox#

Here's where it gets complicated.

Twenty-four hours post-infusion, depressive symptoms dropped significantly. Anhedonia improved. Ruminative symptoms — that grinding, recursive self-focus — decreased measurably on the Ruminative Response Scale ^[1]. So far, so expected. Ketamine works fast; this aligns with prior RCT data showing peak antidepressant effects at approximately 24 hours, with a 40–46% response rate in that window ^[3].

But the symptom reduction did not correlate with changes in rsFC. Let me say that again: the connectivity shifted, the symptoms improved, and these two things were statistically unrelated.

What did predict symptom improvement was pre-ketamine rsFC — the connectivity landscape that existed before the drug was administered ^[1]. This is a genuinely counterintuitive finding. It suggests ketamine's antidepressant action, at least in this 24-hour window, may depend less on what it does to connectivity and more on what connectivity patterns are already in place to be acted upon.

Using exact low-resolution electromagnetic tomography (eLORETA), the researchers found broad increases in theta and beta-band rsFC within the DMN, within the FPN, and between these two networks post-ketamine ^[1]. The authors attribute this to ketamine's known synaptogenic effects — its ability to rapidly promote new synaptic connections, likely mediated through NMDA receptor antagonism, subsequent AMPA receptor activation, and downstream BDNF-TrkB signaling that drives synaptic plasticity ^[4].

The Synaptogenesis Question#

This reminds me of something from the synaptic pruning literature — different context, but the pattern holds. More connectivity is not inherently better. In healthy neural development, the brain prunes excess connections to optimize signal-to-noise. So when ketamine produces a broad increase in rsFC across both DMN and FPN, the question becomes: is this therapeutic rewiring, or is it neurological noise?

The study authors are honest about this tension. They note that ketamine's synaptogenic effects can be short-lived ^[1], which raises the possibility that this connectivity surge is transient — a kind of neural "bloom" that may or may not consolidate into lasting therapeutic architecture. Duman et al. (2022) have shown that ketamine rapidly increases spine density in prefrontal neurons in preclinical models, but these structural changes begin to decay within days without repeat dosing ^[4].

I'm less convinced by the assumption that synaptogenesis alone explains the clinical picture. The dissociation between symptom improvement and connectivity change is the most interesting finding here, and it doesn't fit neatly into a "more synapses = less depression" framework. What does this actually feel like for the patient whose connectivity is shifting but whose symptom relief traces back to where their brain started?

EEG vs. fMRI: A Methodological Note#

Most prior ketamine-connectivity studies used fMRI, which captures hemodynamic changes with high spatial resolution but poor temporal resolution ^[2][5]. This study's use of 96-channel EEG with eLORETA source localization offers something different: direct measurement of neural oscillatory activity with millisecond precision. The trade-off is spatial resolution, and eLORETA has known limitations in localization accuracy. But for capturing the temporal dynamics of connectivity changes — especially in theta and beta bands that correlate with different cognitive and emotional processes — EEG has real advantages.

Sumner et al. (2024) found neurophysiological evidence that frontoparietal connectivity changes and GABA-A receptor modulation underpin ketamine's antidepressant effects ^[6]. The current study extends this by showing those connectivity changes happen broadly, not selectively, which complicates the mechanistic story.

COMPARISON TABLE#

THE PROTOCOL#

Important: Ketamine for TRD is a medical intervention, not a biohacking experiment. The following reflects current clinical protocols informed by the evidence. This is not a self-administration guide.

Step 1. Establish treatment resistance with a qualified psychiatrist. The clinical definition typically requires failure of at least two adequate antidepressant trials at therapeutic doses for sufficient duration (≥6–8 weeks each). Document your treatment history thoroughly.

Step 2. Obtain baseline neuropsychiatric assessment. Request standardized measures: Montgomery-Åsberg Depression Rating Scale (MADRS) or Hamilton Depression Rating Scale (HAM-D), plus the Ruminative Response Scale if rumination is a primary feature. Based on this study's findings, baseline resting state connectivity may predict your response — ask about qEEG assessment availability.

Step 3. The standard IV ketamine protocol uses racemic ketamine at 0.5 mg/kg infused over 40 minutes. This is the dose used in the current study and aligns with the majority of clinical trials ^[1][3]. Ensure cardiac monitoring, pulse oximetry, and blood pressure measurement throughout.

Step 4. Plan your post-infusion window deliberately. The data shows peak symptom effects at approximately 24 hours post-infusion ^[1][3]. This is not the time to make major life decisions or assess your "new normal." Arrange support. Avoid driving for at least 2 hours post-infusion.

Step 5. Assess response at 24 hours and again at 7 and 14 days. Effects typically persist for 1–2 weeks after a single infusion ^[3]. If response is positive but waning, discuss a maintenance protocol. Smith-Apeldoorn et al. (2022) reviewed maintenance ketamine protocols and found evidence supporting repeated dosing, though optimal frequency remains an open question ^[7].

Step 6. Consider integrating psychotherapy in the post-infusion period. The synaptogenic window — when new neural connections are being formed — may represent a period of enhanced neuroplasticity where therapeutic learning could consolidate more effectively. This is theoretically grounded but not yet confirmed in large trials.

Step 7. Track your response systematically. Use validated scales (not just "I feel better"), and if possible, request follow-up EEG to monitor connectivity changes. The current study's finding that pre-treatment connectivity predicts response suggests longitudinal EEG monitoring could become a standard predictive tool.

What is resting state functional connectivity and why does it matter in depression?#

Resting state functional connectivity measures how synchronized different brain regions are when you're not engaged in a specific task. In depression — particularly treatment-resistant cases — the Default Mode Network (which handles self-referential thought and rumination) and the Frontoparietal Network (which manages executive control) show abnormal communication patterns. These aren't just correlates; they may reflect the mechanistic architecture of depressive experience itself.

How quickly does ketamine reduce depressive symptoms?#

Based on the current study and prior RCT data, a single IV infusion at 0.5 mg/kg produces significant symptom reduction within 24 hours ^[1][3]. The effects typically peak at that 24-hour mark and can persist for 1–2 weeks, though this varies substantially between individuals. Honestly, we don't yet have reliable predictors of duration — that's one of the gaps this line of research is trying to close.

Why didn't connectivity changes correlate with symptom improvement?#

This is the study's most provocative finding, and the honest answer is we're still working it out. One possibility: ketamine's synaptogenic effects produce a broad, non-specific increase in connectivity that reflects a general neural "opening up" rather than targeted therapeutic rewiring. The symptom relief may depend on the specific connectivity architecture that was in place before treatment — essentially, what the drug has to work with ^[1].

Who is a candidate for ketamine treatment?#

Ketamine is currently indicated for treatment-resistant depression — defined as failure to respond adequately to at least two antidepressant trials. It's administered under medical supervision, not self-prescribed. Patients with uncontrolled hypertension, active psychosis, or substance use disorders involving ketamine are typically excluded. The FDA-approved nasal spray formulation (esketamine/Spravato) has specific prescribing restrictions through a REMS program.

What are the risks of ketamine infusion for depression?#

Short-term side effects include dissociation, nausea, elevated blood pressure, and dizziness — most resolving within 2 hours. The dissociative experience itself is variable; some patients find it distressing while others describe it as neutral or even meaningful. Long-term risks with repeated dosing are less well-characterized, particularly regarding bladder toxicity and cognitive effects, which are documented in recreational users at higher doses. I'd want to see more long-term safety data before anyone commits to indefinite maintenance protocols.

VERDICT#

Score: 7/10

This study adds a genuinely useful piece to the ketamine-for-depression puzzle, specifically by showing that pre-treatment connectivity may be a better predictor of response than post-treatment connectivity changes. That's a finding with real clinical implications for patient selection. The catch, though: n=24 is small, there's no placebo arm for the ketamine group, the healthy controls didn't receive any intervention (so we can't rule out test-retest effects entirely), and the single 24-hour post-infusion timepoint misses the full trajectory of connectivity changes. The EEG methodology is a strength — more accessible and temporally precise than fMRI — but eLORETA's spatial limitations mean we're working with approximations of source localization. I'd want to see this replicated with a larger sample, a placebo control, and multiple post-infusion timepoints before it changes clinical practice. The synaptogenesis interpretation is plausible but remains speculative without direct molecular evidence in this cohort. Good science. Not yet the full picture.