Summary: Researchers have developed a new method to improve neuroprostheses that promises greater precision and practicality for everyday tasks. By decoding neural signals tied to hand postures rather than movement speed, the team showed that virtual hand control in rhesus monkeys could closely replicate delicate human-like hand actions.
This advance indicates that posture-related brain activity is a particularly informative signal for brain-computer interfaces (BCIs). Leveraging those signals could help restore fine motor function—such as grasping small objects or manipulating tools—for people with paralysis or neurodegenerative conditions, improving day-to-day independence.
Key Facts:
- Prioritizing neural signals representing hand posture increased precision in prosthetic hand control.
- Trained rhesus monkeys could operate a multi-dimensional virtual hand with movements that matched actual hand postures recorded earlier.
- The approach points to new opportunities for improving fine motor performance in future hand prostheses and BCIs.
Source: DPZ
Everyday tasks like carrying shopping bags or threading a needle rely on a spectrum of power and precision grips. The crucial role of our hands becomes obvious when they are no longer under voluntary control—due to spinal cord injury, paralysis, or progressive diseases such as ALS.
For decades, scientists have pursued neuroprosthetics—artificial limbs and hand devices controlled directly by brain signals—to restore mobility and functional independence. A core challenge has been achieving the level of fine motor control required to interact with the varied objects and tasks of daily life.
Neuroprosthetic systems bridge damaged nerve pathways using brain-computer interfaces that decode cortical activity, translate it into commands, and drive a prosthetic device. Historically, many BCI designs for the arm and hand emphasized velocity-related signals that represent how fast or in what direction a movement is executed.
The new study, led by researchers at the Neurobiology Laboratory of the German Primate Center, explored a different approach: extracting and using neural signals that encode hand and finger postures. “How well a prosthesis performs depends primarily on the neural information the interface reads,” says Andres Agudelo-Toro, first author of the study. “We tested whether posture-related signals could provide better control than signals tied mainly to movement velocity.”
Rhesus monkeys (Macaca mulatta) were chosen for the experiments because their nervous systems, visual processing, and manual dexterity closely resemble those of humans, making them suitable models for studying grasping behavior. In an initial training phase, two animals learned to move a virtual avatar hand displayed on a screen while performing the same grasps with their real hands. A data glove with magnetic sensors recorded the real hand configurations to create a training dataset of posture transitions.
After mastering the task, the monkeys were trained to control the virtual hand by imagining or intending the grips without moving their real hands. The researchers recorded population activity from cortical areas specialized for controlling hand movements and updated the BCI algorithm to emphasize posture transitions—the sequences of hand and finger configurations that produce different grips.
“Instead of treating only the movement endpoint as important, we adapted the decoding protocol so that the execution path—the actual posture transitions—also shapes how the prosthetic hand moves,” explains Agudelo-Toro. This change produced notably improved performance: the avatar hand executed complex grips, including fine precision grips, with high accuracy that closely matched the real hand recordings.
Senior author Hansjörg Scherberger highlights the implications: “Our findings show that posture-related activity in the grasping circuit is a dominant and usable signal for controlling a hand prosthesis. Incorporating this signal into BCIs can substantially improve the fine motor abilities of neural prostheses.”
Beyond demonstrating accurate control of a multidimensional hand prosthesis in a primate model, the study establishes posture decoding as a practical information channel for future neuroprosthetics. This suggests a path forward for BCI design that could enable users to perform everyday object interactions with greater dexterity and reliability.
Funding: The research received support from the German Research Foundation (DFG, grants FOR-1847 and SFB-889) and the European Union Horizon 2020 project B-CRATOS (GA 965044).
About this neuroprosthetics and neurotech research news
Author: Susanne Diederich
Source: DPZ
Contact: Susanne Diederich – DPZ
Image: Image credited to Neuroscience News
Original Research: Open access.
“Accurate neural control of a hand prosthesis by posture-related activity in the primate grasping circuit” by Andres Agudelo-Toro et al., published in Neuron.
Abstract
Accurate neural control of a hand prosthesis by posture-related activity in the primate grasping circuit
Brain-computer interfaces have the potential to restore hand movement for people with paralysis, but current devices still lack the fine control required to interact reliably with everyday objects. While prior hand BCI work often emphasized velocity-based control in analogy to reach-related activity, accumulating evidence indicates that posture information predominates in hand-specific cortical areas.
To test whether posture signals can directly and causally control a prosthesis, the authors developed a BCI training paradigm focused on reproducing posture transitions. Monkeys trained under this protocol were able to operate a multidimensional hand prosthesis with high precision, including delicate precision grips. Analysis showed that posture-related neural activity in the targeted grasping circuit was the principal contributor to control. These results demonstrate posture-based neural control of a complex hand prosthesis and open the door for future interfaces to exploit this rich information channel.