Neuroprosthesis: The solution to communication for the paralysed?
By Shivam Bhindi
Every four hours, someone is paralysed by a spinal cord injury, permanently changing not only their life, but those surrounding them. Paralysis is the loss of muscle function in one, or many, parts of your body and it occurs when there is a problem with the messages that pass between your brain and your muscles. For those suffering from more serious paralysis, something as simple as communication is stripped from them without warning. Simple everyday tasks, such as having a conversation, become impossible and this can be emotionally devastating for the sufferer.
Some current methods of communication for the paralysed include electronic devices that allow computer speech to be generated through movement of a facial feature; this can include eye trackers and twitching of the cheek, as famously used by Stephen Hawking. However, despite being a viable option, the letter-by-letter approach makes these processes painfully slow, especially due to the fast nature of our speech, which averages at 150-200 words a minute.
Another option is neuroprosthesis – devices which use electrodes that aim to restore partial movement that has been lost due to spinal cord injury. A recent breakthrough in this field was in a study conducted by neurosurgeon Edward Chang –instead of the spelling approach commonly used, Chang aimed to translate one word at a time. Instead of restoring movement in the hand for typing, Chang intended to translate the signals arriving at the vocal tract to allow paralysed people a faster and easier method of communication.
In order to translate these signals controlling the vocal tract, the model needed to learn the association between patterns of complex brain activity and what wants to be said. These patterns were detected by studying patient volunteers at the UCSF epilepsy centre, all of whom had normal speech. These patterns were modelled statistically, allowing them to successfully decode speech signals. This, however, did not necessarily mean it would work in those who had their vocal tract paralysed for a prolonged period of time. Would the brain-vocal tract connection still be intact after such inactivity? The only way to find out was through a trial.
The first participant suffered from a devastating brainstem stroke 15 years prior to the trial, destroying the connection between his brain, vocal tract and limbs. At that point, the patient used a pointer attached to a cap, poking letters on a screen as a way to communicate. After surgically implanting an array of electrodes into the patient’s speech motor cortex, his brain was monitored for many months. In these months, the patient attempted to say 50 simple words numerous times, allowing scientists to record the brain signals associated with each word. Eventually, as he tried to say different words, and the computer had successfully associated them with a signal, they appeared, one by one, on a screen, proving that the theory works. Furthermore, they switched to asking questions, the response being either a sentence or a string of different words. Surely enough, each of the responses was displayed on the screen, with up to a 93 percent accuracy at a rate of 18 words a minute.
Despite this being considerably slower and less accurate than normal speech, this trial presents a massive stepping stone for the field of neuroprosthesis. Additionally, this technology could be expanded in multiple ways, for example, by helping those who have suffered from severe strokes to communicate in a more natural way. From here on, scientists now know that text can be generated by monitoring brain activity with an electrode, and it is certain that this technology will develop in the near future to a standard which allows those who struggle to communicate to retain some aspects of a normal lifestyle.
References:
2. https://medlineplus.gov/paralysis.html
3. https://www.sciencedaily.com/releases/2021/07/210714174148.htm
4. https://www.macmillandictionary.com/dictionary/british/neuroprosthesis