Opparounds

Connectivity- vs Scalp-Based Targeting of Accelerated Transcranial Magnetic Stimulation for Depression: A Randomized Clinical Trial

Taylor JJ, Kare MR, Haj-Darwish D, et al. · JAMA Psychiatry · June 2026

This randomized trial tested whether targeting accelerated TMS (aTMS) using a patient's own resting-state functional connectivity beats the standard scalp-based Beam F3 method in treatment-resistant depression. Forty adults with moderate-to-severe, moderately-to-severely treatment-resistant MDD underwent a 41-minute connectivity scan, then were randomized to left-DLPFC stimulation targeted either to the region most connected to a validated depression circuit, or to the scalp landmark clinics use today.

At one month, median MADRS reduction was 24 points with connectivity-based targeting versus 18 points with scalp-based targeting (P = .02), an effect size of 0.8 with a number needed to scan of five. Individualized targets were reproducible within a given patient (about 4.5 mm) but varied meaningfully between patients (about 13 mm) — the mechanistic argument for why the target itself matters.

Read the paper →  doi:10.1001/jamapsychiatry.2026.1100

Antidepressant Treatment Duration and Risk of Sudden Cardiac Death: A Nationwide Cohort Study

Mujkanovic J, Warming PE, Kessing LV, Køber LV, Winkel BG, Lynge TH, Tfelt-Hansen J · Heart Rhythm · March 2026

This Danish cohort followed 4.3 million adults to ask whether how long someone has been on an antidepressant — not just whether they're on one — changes sudden cardiac death (SCD) risk. Exposure came from two decades of pharmacy registry data, categorized by cumulative duration and recency of use.

After adjusting for age, sex, and cardiovascular comorbidity, patients with 1 to 5 years of antidepressant exposure had a 56% higher SCD risk (HR 1.56), and those with 6 or more years had more than double the risk (HR 2.17, both P < .001) versus unexposed people. The signal held across SSRIs and TCAs alike, arguing against a single-drug idiosyncrasy and for a duration-dependent, class-spanning effect.

Read the paper →  doi:10.1016/j.hrthm.2026.03.006

A fever can push clozapine into a toxic range without a single missed dose

Clozapine is metabolized almost entirely by CYP1A2, and CYP1A2 is exquisitely sensitive to inflammatory cytokines. During a systemic infection, IL-6 and TNF-alpha transcriptionally suppress CYP1A2, and clozapine clearance can drop sharply within days — independent of smoking status, dose, or adherence. Case series describe plasma levels rising two- to threefold during febrile respiratory infections, high enough to produce sedation, sialorrhea, ataxia, or seizures that get misread as the infection itself, or worse, as relapse requiring a dose increase.

The distinguishing feature is timing: toxicity that tracks with CRP elevation and fever, not with any change in the regimen. Because the cytokine effect can outlast the fever by days, a level drawn right at symptom onset can look reassuring even as the peak is still ahead. The practical move: any clozapine patient with a systemic infection and new sedation or autonomic signs gets a level and a hold, not a dose increase — and the level should be repeated as the fever resolves rather than trusted as a single data point.

Epistemic trust is what determines whether insight travels home

Mentalization-based treatment's account of change has narrowed from 'better mentalizing reduces symptoms' to something more specific: therapy works by restoring epistemic trust — the patient's felt sense that another mind's input about them is relevant and worth updating on. Patients with histories of relational trauma or invalidation often default to epistemic vigilance, screening out corrective information regardless of its accuracy, which is why an accurate interpretation can land and still change nothing.

Epistemic trust isn't installed by being right; it's earned when the patient experiences being accurately mentalized — recognized as a specific person with a specific mind, not processed through a generic template. Once that trust opens, ordinary social learning resumes: the patient can use new input from people outside the room, which is the actual mechanism by which gains outlast termination.