Scientists Develop a Tongue-Controlled Robotic Glove That Could Help Paralyzed People Use Their Hands Again

Imagine controlling a robotic hand simply by moving your tongue. For people who have lost hand movement because of a spinal cord injury, this technology could offer a new way to perform everyday activities, such as holding a bottle, picking up objects, or using a toothbrush.

Researchers have developed a hybrid hand-assistance system that combines a soft robotic exoskeleton with functional electrical stimulation (FES). The system allows people with severe paralysis to control finger movements using a non-invasive tongue-based interface. Unlike conventional systems that rely on muscle activity or brain signals, this approach uses the tongue to issue commands, making it particularly useful for people whose hand and arm movements are severely limited.

The research was led by Oguzhan Kirtas and his team, who tested the technology in five people with cervical spinal cord injuries. The results demonstrated that the system could support several everyday grasping tasks, even in individuals who had lived with paralysis for decades.

Why Restoring Hand Movement Matters

Spinal cord injuries affect more than 20 million people worldwide. When an injury occurs in the cervical region of the spinal cord, which runs through the neck, it can severely limit movement in the arms and hands. This condition is known as tetraplegia.

For many people living with paralysis, regaining hand function is one of their most important goals. The ability to hold a cup, use a toothbrush, or pick up a small object can make a significant difference in independence and quality of life.

Although rehabilitation robotics, brain-computer interfaces, and electrical stimulation have made progress in restoring movement, important challenges remain. Some systems require detectable muscle activity, while others depend on complex equipment or invasive procedures. Controlling both the opening and closing of the fingers safely can also be difficult.

The new system addresses these problems by combining robotic assistance, electrical stimulation, and tongue-based control in a single approach.

How the Tongue-Controlled Robotic System Works

The technology has three main components: a non-invasive tongue interface, a soft robotic hand exoskeleton, and a functional electrical stimulation system.

1. A Tongue Interface for Giving Commands

The first component allows users to control the system through tongue movements. The researchers developed a non-invasive tongue-robot interface known as nTRI.

The interface uses inductive sensors mounted on a small printed circuit board inside the mouth. A frame-integrated activation unit can be moved with the tongue to activate different sensors and select commands.

Unlike certain earlier tongue-control systems, this design does not require a device to be attached directly to the tongue or a tongue piercing. This could make it more practical and comfortable for long-term use.

The interface allows users to select different commands for controlling the robotic hand. By moving their tongue to the appropriate positions, they can choose grasping patterns and trigger the hand-opening process.

Because the tongue often retains its movement after a high-level spinal cord injury, it can provide a useful control method for people who cannot reliably move their hands or arms.

2. A Soft Robotic Glove That Moves Five Fingers

The second component is a lightweight, tendon-driven soft hand exoskeleton. It weighs approximately 125 grams on the hand and is designed to assist finger movements without the bulk of a rigid robotic structure.

The system uses tendons to bend the fingers. Importantly, it can control the flexion of all five fingers independently, allowing it to support different grasping patterns.

When a user selects a command through the tongue interface, the exoskeleton pulls the required tendons to close the fingers around an object.

This can help users perform activities that require different hand positions, rather than being limited to a single basic grasp.

The researchers demonstrated five types of grasps designed for activities of daily living. These included tasks involving common objects such as bottles and toothbrushes.

3. Electrical Stimulation Helps Open the Hand

Closing the fingers around an object is only half the challenge. A person must also be able to release the object safely.

To address this problem, the researchers incorporated functional electrical stimulation, or FES. This technique uses electrical pulses delivered through electrodes placed on the skin to activate muscles and produce movement.

In the new system, the robotic exoskeleton handles finger flexion, while FES helps produce finger extension, allowing the hand to open.

This division of work is an important feature of the design. Instead of using additional robotic motors to open every finger, the system uses the person's own muscles through electrical stimulation.

The researchers limited FES to brief periods during hand opening, with stimulation lasting no more than two seconds in the reported trials. This approach was intended to reduce the amount of stimulation required and help limit muscle fatigue.

It also avoids relying on bulky extension mechanisms that can add resistance when the fingers bend.

The researchers reported that the system enabled hand opening without causing finger hyperextension during the experiments.

What Happened During the Tests?

The team evaluated the system in five participants with cervical spinal cord injuries. Before using the exoskeleton, participants generally could not grasp the test objects independently. Two participants occasionally managed a few lightweight objects by using compensatory wrist movements.

With the tongue-controlled exoskeleton and FES-assisted hand opening, the participants achieved a 96% success rate in the tested grasping tasks. The version using a passive finger-opening mechanism achieved an 89% success rate.

These results suggest that combining robotic finger flexion with electrically stimulated finger extension could provide practical assistance with everyday activities.

The system also improved the participants' assisted grasping forces. Compared with testing without the exoskeleton, average pinch forces were approximately 2.7 times higher with the FES-opening configuration, while grasp forces were approximately 5.9 times higher.

The measured pinch forces ranged from 2.6 to 7.0 newtons, while grasp forces ranged from 8.6 to 17.1 newtons. These measurements indicate that the system could generate useful forces for handling many lightweight everyday objects.

However, these results reflect the combined assistance provided by the device, not a recovery of the participants' natural hand strength.

The participants also reported favorable experiences with the FES-assisted version. They rated it more positively for wearability, ease of use, and willingness to use it than the passive-opening configuration.

Reported discomfort, pain, and fatigue levels were generally low. Notably, even participants who had experienced paralysis for 20 to 30 years were able to release their grasp using electrical stimulation.

The average time to complete tasks was approximately 12 to 13 seconds. This includes more than the device's operating time, since participants also needed to react to commands, approach objects, and coordinate with the experimental setup.

Why This Approach Could Be Different

Existing robotic gloves and FES systems each have advantages, but they also face limitations.

A robotic exoskeleton can provide controlled movement without requiring continuous muscle activation. However, adding separate motors for every movement can make the device heavier and more complicated.

FES can activate the user's own muscles, but prolonged stimulation can cause fatigue. It also requires careful electrode placement and adjustment to the individual.

The new approach separates the two functions: the robotic glove bends the fingers, while electrical stimulation helps straighten them. This arrangement reduces the need for extra actuators and limits FES use to a brief phase of the task.

The tongue interface also provides a control option for people who cannot reliably operate conventional switches, touchscreens, or muscle-controlled devices.

Together, these features could help make assistive hand technology more compact and easier to operate.

What Are the Limitations?

Despite the promising results, the technology is still at an early stage of development.

The study involved only five participants, so larger trials are needed to determine how consistently the system works across a wider population. The reported success rates should not be interpreted as proof that the device will work equally well for everyone with paralysis.

The current setup also relies on external equipment, including electrical stimulators. This limits portability and would need to be addressed before the system could become convenient for everyday use outside a laboratory.

Users may still require assistance when putting on or removing the equipment. In addition, FES electrodes must be positioned and calibrated for each individual, which can make the initial setup time-consuming.

The researchers also noted that fatigue was assessed through participants' reported experiences rather than objective measurements of muscle performance. Further studies will be needed to establish how the system performs during longer periods of repeated use.

Future versions could incorporate additional sensors to improve object-contact detection, simplify control, and support more grasping patterns.

A Potential Step Toward Greater Independence

The tongue-controlled hybrid hand exoskeleton offers a new way to combine human movement, electrical stimulation, and robotic assistance. By allowing users to issue commands with their tongue, bend their fingers through tendon-driven robotics, and open their hands with FES, the system addresses several important challenges in assistive technology.

Its ability to support everyday grasping tasks in people with severe and long-term spinal cord injuries is particularly encouraging.

The device does not yet restore independent hand function without assistance, and further testing is necessary before its benefits can be established in everyday settings. Nevertheless, the findings suggest that a relatively lightweight robotic glove, controlled through a non-invasive tongue interface, could help people with paralysis perform tasks that are otherwise difficult or impossible.

With further development, this technology could contribute to assistive devices that are more comfortable, practical, and accessible, helping people with severe paralysis gain greater independence in their daily lives.

Reference: Kirtas, O., Veltink, P.H., Kæseler, R.L. et al. A tongue-controlled FES-robotic exoskeleton restores hand function in users with cervical spinal cord injury. npj Robot 4, 53 (2026). https://doi.org/10.1038/s44182-026-00117-9