From speech-to-text software to hearing and mobility aids, nearly every person will need some form of assistive technology during their lifetime. Today, more than 2.5 billion people worldwide require these products to help them lead healthier and happier lives, according to the World Health Organization. But existing hardware still excludes some disabled people or those with chronic or degenerative illnesses—not everyone can control a joystick, for instance. Newer solutions that attempt to link a person’s thoughts up to a computer are still in early development, and it’s taken users some time to learn how to use brain-computer interfaces.
Enter a team of researchers from Singapore and China who are trying to fill in the gaps with a mouth guard. Their device, which they detailed in a study published Monday in Nature Electronics, interprets a user’s bite pattern into one of 21 different signals. The scientists then programmed these bites to direct a remote-controlled wheelchair and play a virtual piano in order to showcase the potential for this new class of assistive technology.
Xiaogang Liu, a materials science researcher at the National University of Singapore who led the study, told The Daily Beast in an email that existing assistive technologies leave plenty to be desired. Voice recognition only works in low-noise settings, eye-tracking software requires a camera be mounted in front of a user, and newly developed brain-computer interfaces need complex arrangements of wires and can involve invasive surgeries. A mouth guard would solve these issues in a low-cost, accessible fashion, Liu said.
ADVERTISEMENT
“Compared to existing assistive technologies, we believe interactive mouth guards are suitable for most people, including those who wear dentures,” he explained.
The mouth guard consists of six pressure sensors embedded in a sturdy plastic shell that can withstand over 20,000 bites before needing to be replaced, Liu said. The six sensors are aligned in two rows, allowing a user to bite down on one or more sensors at a time. A computer then recognizes the pattern and classifies it as one pattern. The patterns correspond to discrete directions too.
When two volunteers tested the mouth guard’s ability to steer a wheelchair, for example, different bite patterns turned the wheelchair, accelerated it forward, and made it stop. They then played the first few bars of “Happy Birthday” on a digital piano using different bite patterns to correspond to the various notes in the song.
In a final test, one participant used the mouth guard to spell out the word “love” and the value 3.14 using a virtual keyboard. These applications measured the mouth guard’s “sophistication and dexterity,” Liu said, adding that future tests of the device may incorporate more complex games and real-world scenarios.
There did seem to be a learning curve: Early on, it took the participants almost 90 seconds to play the tune, but 70 training sessions later, they could complete the task in a third of the time. Still, the training time was shorter than what brain-computer interfaces typically require, Liu said. His team is developing the second generation of “smart mouth guards,” he said, which will use machine-learning algorithms to train the device alongside the participant to cut training times.
While these mouth guards can’t be a solution for everyone, they may be able to offer one more option to assistive technology users, helping people who could not otherwise access these devices.
