Back

Plasma Proteomics Identifies a Microtesla Magnetic Therapy Response Signature in Long COVID

Brady, N. R.; Canori, A.; Maltz, D. S.; Kirsher, D.; Zhou, W.; Becker, J.; Putrino, D.; Gurfein, B. T.

2026-08-27 neurology
10.64898/2026.08.25.26361319 medRxiv
Show abstract

Cognitive impairment is a disabling feature of Long COVID with no established disease-modifying therapy, and little is known about the biological changes accompanying clinical improvement. Microtesla Magnetic Therapy (MMT) is a low amplitude radiofrequency electromagnetic field intervention delivered to the whole brain. In a randomized, sham-controlled feasibility trial, at home MMT was feasible, safe, and well tolerated, with evidence of clinical improvement among treated participants. We explored molecular changes associated with response using SomaScan 11K plasma proteomics on paired baseline and week 4 samples. Participants were classified post hoc within each treatment arm using a clinician-selected response phenotype integrating cognitive and symptom domains. These groups were used for proteomic, pathway, and OrganAge analyses. MMT response was associated with selective proteome remodeling and an exploratory 17 protein response pattern in which Hedgehog interacting protein (HHIP), a Hedgehog signaling antagonist, was most strongly associated with response. Directional pathway analysis identified patterns consistent with lower inflammatory and injury biology and higher repair and adaptive remodeling. OrganAge analysis showed trends toward lower Brain and Organismal OrganAge with MMT. These findings prioritize HHIP and the exploratory 17 protein response pattern for prospective validation and support evaluation of plasma proteomics for monitoring treatment response.

Matching journals

The top 14 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.