JOVANA
Explore Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Where to Stimulate: Targets, Circuits and Connectomes

Psychiatric illness lives in distributed circuits, not single nuclei — so modern targeting is shifting from an atlas coordinate to a connectome, and from one target per disease to a network view.

From a nucleus to a network

The intuition inherited from Parkinson's DBS — find the culpable nucleus, stimulate it — travels poorly to psychiatry. Depression and OCD are better described as dysfunctions of distributed limbic and cortico-striatal circuits, so the electrode's real job is to modulate a network node whose influence propagates along known connections. This is the premise of network-targeted stimulation: you are not treating a point, you are perturbing a graph.

The depression targets

Several targets have been pursued for treatment-resistant depression, each with a rationale and none yet decisive. The subcallosal cingulate (Area 25) sits at a convergence of mood-regulating pathways and is chronically over-active in depression. The ventral capsule / ventral striatum engages reward and motivation circuitry. The lateral habenula — an 'anti-reward' hub that is hyperactive in depression — and the medial forebrain bundle are more recent, mechanistically-motivated bets. Each is a hypothesis about which node most efficiently moves the mood network.

Note what unites the setbacks and the hopes: the disagreement is rarely about whether stimulation does something — it usually does, engaging the local and network mechanisms of DBS — but about whether that something is the right something, reliably, in this patient. That is a targeting-and-selection problem as much as a stimulation problem.

OCD, and one target for more than one disease

The ventral capsule/ventral striatum is also the best-established OCD DBS target, and the subthalamic nucleus is an alternative. That the same anatomical region is used for both OCD and depression is a clue, not a coincidence: it sits on a cortico-striato-thalamo-cortical loop whose dysregulation shows up as different symptoms depending on which sub-circuit dominates. It also warns us that 'target' and 'diagnosis' are not in one-to-one correspondence — the same contact can produce very different stimulation-induced behavioural states depending on exactly which fibres it recruits.

Targeting by connectome, not coordinate

The most important shift in modern targeting is to stop chasing a fixed stereotactic coordinate and instead place the volume of tissue activated so that it maximally overlaps a connectivity profile associated with response. Using each patient's own tractography, one asks: where must I stimulate so that the activated fibres reach the response-associated network? This connectomic view reframes 'a good target' as 'a good connectivity fingerprint,' and it partly explains why fixed-coordinate trials underperformed individually-tuned case series.

\hat{R} \;=\; \big\langle\, \mathrm{VTA}(u,\vec r)\,,\ C_{\text{target}}(\vec r)\,\big\rangle \;=\; \int_{\Omega} \mathbb{1}\!\left[\,E(\vec r) > E_{\text{th}}\,\right]\, c(\vec r)\; d\vec r

A connectomic targeting objective. The activated volume is where the induced field E exceeds an activation threshold E_th; weighting it by a per-voxel connectivity map c(r) to a response network gives a predicted response R̂ to maximise over lead placement and stimulation settings u.

The main neuromodulation targets and the circuits that connect them — subcallosal cingulate, ventral capsule/ventral striatum, lateral habenula and their cortico-striatal partners — the map this guide argues you should read as a network, not a list of coordinates.