The circuit that makes accelerated TMS work: a fronto-insular pathway, mapped from mouse to human
- In an optogenetic mouse model of accelerated intermittent theta burst stimulation (aiTBS) targeting prelimbic cortex, the antidepressant-like effect was reproduced — and traced to a specific downstream pathway, not just the stimulated site.
- aiTBS drove cell-type-specific changes in intratelencephalic prefrontal projection neurons: upregulated synapse-related gene expression, increased dendritic spine density, and larger excitatory currents — a structural plasticity signature, not transient excitability.
- Projection-specific optogenetic and chemogenetic manipulation showed that activation of a prefrontal→insular (fronto-insular) network was both *necessary and sufficient* for the behavioral effect — silencing it abolished the response.
- The mechanism was validated in humans: intracortical stereo-EEG and resting-state fMRI confirmed that TMS-evoked responses propagate through fronto-insular connectivity, anchoring the animal circuit in the human brain.
We have been delivering accelerated TMS — the SAINT/aiTBS family of protocols — to treatment-resistant patients for several years without a mechanistic account of why it works. This paper closes part of that gap. It is the rare neuromodulation study that runs the full arc from molecular plasticity in defined cell types to a causal circuit, and then checks that circuit in the human cortex.
What the data shows
The authors built an optogenetic analogue of aiTBS in mice, stimulating prelimbic cortex (the rodent homologue of the human prefrontal target) and reproducing rapid antidepressant-like behavior. They then asked what changes inside the neurons. The answer was concrete: intratelencephalic projection neurons showed elevated expression of synapse-construction genes, denser dendritic spines, and amplified excitatory postsynaptic currents. This is the cellular grammar of durable plasticity — the cortex is being rewired, not merely excited for a few hours.
Whole-brain c-Fos mapping and fiber photometry then pointed downstream to the insula. Using projection-specific optogenetics and chemogenetics, the team demonstrated that the prefrontal→insular pathway was necessary and sufficient: drive it and you get the effect, block it and the effect disappears. That double dissociation is what separates a correlation from a mechanism.
Why mapping the insula matters clinically
The insula is the cortex's interoceptive hub — it integrates the body's internal state into emotional experience. Implicating it as the obligatory relay for aiTBS reframes the antidepressant effect as a recalibration of how the brain reads the body, not simply a "boost" to a sluggish prefrontal cortex. For clinicians, this is the beginning of a rational target-selection logic: response may depend on the integrity of fronto-insular connectivity in a given patient, which is measurable with resting-state fMRI before treatment. It also suggests why interoceptive and somatic symptoms — appetite, fatigue, the felt sense of dread — often shift early in responders. We are not yet personalizing TMS by connectome, but this is the kind of evidence that will let us.
Accelerated TMS does not just wake up the prefrontal cortex — it recruits a prefrontal-to-insula circuit that is necessary and sufficient for the antidepressant effect.
The causal mechanism rests on an optogenetic mouse model; the human validation establishes the circuit exists and conducts TMS-evoked signal, but does not prove the same necessity-and-sufficiency in patients. Antidepressant-*like* behavior in rodents is not clinical remission.