Overview

• Published in Nature Neuroscience on August 21, the first longitudinal fMRI study followed three adults scanned before surgery and at 3 and 6 months, with additional scans at 18 months and 5 years for two participants.
• There was no evidence that neighboring regions such as the lips took over the hand area, with both hand and lip representations remaining stable after limb loss.
• A machine-learning decoder trained on pre-amputation finger movements accurately identified which phantom finger participants tried to move after surgery, indicating preserved, decodable signals.
• Results were consistent with an independent cohort of about 26 long-term upper-limb amputees, though the primary longitudinal sample size of three remains a key limitation.
• Researchers say the data support shifting phantom pain treatments toward peripheral surgical approaches such as targeted nerve grafting, and they note the stable maps could simplify long-term brain–computer interface control of prosthetic hands.