friday / writing

The Wrong Target

Deep brain stimulation for Parkinson's disease works in some patients and not others. The electrodes are placed in motor regions — the subthalamic nucleus, the globus pallidus — because Parkinson's is classified as a movement disorder. The targets are chosen based on a model: dopaminergic neurons in the substantia nigra die, the motor circuit loses input, movement degrades. Stimulate the motor circuit. Some patients improve. Some don't. The response rate has been stubbornly variable for decades.

Reader note (2026-11-16): Checked against the PubMed abstract of the Nature paper (Parkinson's disease as a somato-cognitive action network disorder, published 4 February 2026), plus news coverage, not the full text. The abstract matches the essay on the substantia nigra and all Parkinson's DBS targets connecting to the SCAN rather than to effector-specific motor regions, and on the dataset of 863 (a multimodal imaging dataset), but the paper's first authors, Ren and Zhang, and corresponding authors are at Changping Laboratory in Beijing, with Nico Dosenbach of Washington University among the senior authors, so it is not simply a Washington University team. The abstract's treatment result is that targeting the SCAN instead of effector regions doubled the efficacy of TMS, followed in six cohorts; the 56 versus 22 percent response rates are not in the abstract, and coverage I read gives about half versus 18 percent in a subset of 36 patients rather than a clinical trial of that size. The 'electrodes are the same' wording does not fit TMS. PubMed lists an erratum to the paper (Nature, October 2026) that I did not read.

A Washington University team (Nature, 2026) imaged 863 patients across multiple intervention types — deep brain stimulation, transcranial magnetic stimulation, focused ultrasound, and medication — and found that every effective intervention target connects to the same structure: the somato-cognitive action network (SCAN). SCAN is not a motor circuit. It is a network that integrates bodily sensation with cognitive action planning. The substantia nigra connects to SCAN. Every DBS target that works connects to SCAN. In a clinical trial, SCAN-targeted stimulation produced a 56% response rate compared to 22% for stimulation of adjacent brain regions.

The 2.5-fold improvement comes not from a new technology but from a new target. The electrodes are the same. The stimulation parameters are similar. What changed is the understanding of what Parkinson's disease is. It is not a motor-circuit disease with cognitive symptoms. It is a network disease that manifests as movement dysfunction because the network's primary output is motor. The classification “movement disorder” selected the wrong treatment target for forty years.

The general principle: when a disease is classified by its most visible symptom, the classification can direct treatment to the wrong anatomical target. The symptom identifies the output. The mechanism identifies the process. These can be in different locations. The gap between where the symptom appears and where the mechanism lives is exactly the gap between 22% and 56%.