September 2024: Speech Alternatives

For me and, I think, for Lois, my loss of speech is much crueler than my loss of mobility. She gets no conversation from me. Silent meals and silent drives between Portland and the beach are particularly trying. My laconic emails and slowly typed computer-generated speech do not fill the void very much.

I am now typing part time with the hand mouse and practicing the rest of the time with the new head mouse, which consists of a detecting lens mounted on top of my speech computer and connected via a USB port. A button is mounted on my work tray and wired to the detector. Pushing it gives a left mouse click; there is no way to right click. The detector tracks a reflective silver – colored dot, about a quarter inch in diameter, adhered to the bridge of my glasses. I can move the cursor around the screen by moving my head up and down, left and right. The head mouse is still a little slower and clumsier than the hand mouse, but I can use it for a couple of hours without too much fatigue. I practice with it because I fear that I will eventually lose control of the hand mouse.

I do regular neck exercises, so I am aware of slight weakness of my neck, but this is nowhere near as bad as my hands. Usually, people with ALS retain eye movements until very late in the illness, so the eye tracker is the next option if the head mouse fails, but so far, my eyes have been too dry for me to use the eye tracker.

All these methods are enhanced by software that includes word prediction and the ability for me to store frequently used phrases. I currently prefer a program called Grid 3 that allows me to speak with an automated voice. I can text, email, command Alexa, and make a phone call.

There are options being developed that would help if I could not use the hand mouse, head mouse, or eye tracking. I have had some sessions with Dr. Betts Peters and her colleagues from OHSU. They attached EEG electrodes to my scalp and had me watch letters flash on a monitor. I could think about the letter I wanted and spell words. This is slow and not yet ready for practical use.

Another intriguing new approach is the Synchron Stentrode. The company has attached electrodes to the kind of stents that are now routinely inserted by catheter into arteries or veins. The Stentrode is inserted into the jugular vein in the neck and threaded through the veins of the brain into the large venous superior sagittal sinus where it can be implanted next to the motor strip, location of the cell bodies of the upper motor neurons. The wires from the 16 electrodes of the Stentrode run through veins to a Bluetooth transmitter placed under the skin just below the collar bone. When the subject tries to move a paralyzed limb, the computer receives electrical signals. With training and practice, the subject can learn to move a cursor on a screen and do everything that I do using the Grid 3 software.

Implanting electrodes in the motor strip and using the recorded brain signals to function like a computer mouse has been done for a few years and is getting more sophisticated.  Elon Musk has used this approach with his Neuralink device. Until now this has required a craniotomy, a neurosurgeon cutting a hole in the skull and placing a foreign object in the brain.

When I read the report of four people with ALS or PLS who were successfully treated with the Stentrode, I was excited because this approach avoids the craniotomy. (Mitchell P, Lee SCM, Yoo PE, et al. Assessment of Safety of a Fully Implanted Endovascular Brain-Computer Interface for Severe Paralysis in 4 Patients: The Stentrode With Thought-Controlled Digital Switch (SWITCH) Study. JAMA Neurol. 2023;80(3):270–278.) I learned that Synchron is planning a larger trial in 2025. I signed up with the company registry, hoping that Portland would be one of the sites for this trial.

In my post of August 31, 2023, I wrote about experiments at UCSF and Stanford in which the implanted electrodes allowed the subject to generate computer-aided speech in real time, rather than moving a cursor on screen. Last month researchers from UC Davis reported an impressive improvement in this approach (Card NS, Wairagkar M, Iacobacci C, et. al. An accurate and rapidly calibrating speech neuroprosthesis N Engl J Med. 2024 Aug 15;391(7):609-618.)

They implanted four electrode chips with a total of 256 electrodes in the brain of a 45-year-old man who had ALS. Within months, he was speaking through a computer 32 words a minute from a 125,000-word vocabulary with 98 per cent accuracy. The man’s pre-operative speech function was better than mine, so it is not clear that I would get the same results, but I would be tempted to try it; however, it is unavailable in Portland, and I doubt that I am strong enough to travel to California for an extended stay or to undergo a five-hour craniotomy.

Speech facilitated by brain-computer interface is improving rapidly. I am hoping that I will benefit soon.