A peer-reviewed study published on 16 November 2026 reports unusually durable prosthetic-control signals from implanted electromyography. Researchers tested continuous, simultaneous control of two virtual fingers using signals from regenerative peripheral nerve interfaces, or RPNIs.
The human participant, who had a transradial amputation, reached 56% of the task performance measured with a natural hand. Most notably, the decoding system showed no statistically significant decline over 283 days without recalibration (p=0.78). A rhesus macaque was also studied, providing supporting preclinical evidence.
The result addresses a persistent practical obstacle for advanced prostheses: control systems based on biological signals can drift, forcing repeated recalibration and limiting everyday reliability. Long-lived implanted signals could make future multi-degree-of-freedom hands more dependable and easier to use.
The evidence is nevertheless preliminary. This was one human participant performing a narrow laboratory task with a virtual interface, not a trial of a complete physical prosthetic hand in daily life. The study does not establish clinical effectiveness, commercial availability, durability across a wider population, or any workplace or economic advantage.
The appropriate interpretation is therefore a meaningful engineering advance in the control layer, not a demonstration of human-equivalent or superhuman prosthetic capability. Larger studies and integration with physical devices are required before the result can support clinical or commercial claims.