Dec 2020 Neurology Today
A brain-computer implant has enabled two patients with ALS to text, shop, and do online banking. The novel stent-electrode device circumvents the need to open the skull by snaking the device through blood vessels.
The device was implanted within the superior sagittal sinus, immediately adjacent to the precentral gyrus. The highlighted yellow region in the brain depicts the activation of primary motor cortex that occurs with attempted limb movement. The transmission lead exiting the internal jugular vein between the heads of sternocleidomastoid was tunneled subcutaneously and connected to the internal telemetry unit (ITU) within a sub-clavicular pocket. The external telemetry unit inductively powers the ITU and receives a electrocorticography signal via infrared light transmission. The signal is sent to a tablet computer via a signal control until translated into multiple click actions by the custom decoder, including a zoom function and single click command. Multiple command control was combined with eye-tracking to enable general operation of Windows 10.
The novel stent-electrode device, called the Stentrode BCI, circumvents the need to open the skull by snaking the device through blood vessels. Using an endovascular technique, the device is introduced via the jugular and navigated to the superior sagittal sinus adjacent to the primary motor cortex, and connected to a unit in the chest, which wirelessly transmits signals.
After less than three months of healing and training, the participants were able to use the device at home, unsupervised, in conjunction with an eye-tracker for cursor navigation. It allowed them to intentionally left-click or zoom on a Windows 10 operating system. They achieved an average click accuracy of over 92 percent with a typing speed of 13.8 to 20.1 characters per minute.
The results drew praise from academic scientists in the field for demonstrating safety up to a year, reasonable accuracy, and the ability of patients to use it at home without assistance. But given similar typing speeds with non-invasive EEG and far greater speed with surgical implants, some questioned the long-term usefulness of the system as others in the field race to develop more sophisticated devices.
“It's the first report of the full device being used to control computers, in two people, and that's a big deal,” said Sergey Stavisky, PhD, a postdoctoral fellow at Stanford's Neural Prosthetics Translational Laboratory. “They've come a long way from the animal studies previously published, and very quickly. That's quite the feat. But given this level of performance, one wonders if, for some potential patients, a similar ‘yes-no’ command could be obtained using a non-invasive BCI or even their residual movements.” Even so, he added, “This shouldn't diminish the importance of this work. There's not going to be a one-size-fits-all BCI. Having a variety of options, with different trade-offs, is a good thing for patients.”
While acknowledging that devices with a much higher data transmission rate have been demonstrated in laboratory studies, the leader of the new study emphasised that his group is focused on moving the technology into users' homes. “The difference here is we are now talking about a clinically translated, fully implanted take-home device for the patient,” said Thomas J. Oxley, MD, PhD, an instructor and director of innovation strategy for the department of neurosurgery at the Mount Sinai Health System, and founding CEO of Synchron, the maker of the device. “This is not about academic science. We're trying to build a technology that can help many people.”
Participant 1, a 75-year-old man from Australia—where Dr. Oxley was a neurologist at the Royal Melbourne Hospital—was diagnosed with cervical onset ALS (flail limb variant) in 2018. During a screening in 2019, a neurologist excluded dementia, and respiratory assessment revealed a forced vital capacity (FVC) of 3.25 (81 percent). Loss of muscle power in the limbs and distal shoulder had left him unable to use a personal computer or smartphone without the aid of a caregiver.
Participant 2, a 60-year-old man, was diagnosed with cervical-onset ALS in 2015. Dementia was excluded at screening, and FVC was at 3.9 (68 percent). He was unable to use a computer without assistance due to loss of control over his fingers, elbows, and shoulders. He was taking riluzole 50 mg twice daily.
With approval by the St. Vincent's Hospital Melbourne Human Research Ethics Committee, patients were placed on dual antiplatelet therapy with aspirin and clopidogrel for 14 days prior to implantation and continued for at least three months. Participant 1 underwent the procedure in mid-2019, and participant 2 in early 2020.
Under general anaesthesia, guided by 3D digital subtraction angiography and presurgically marked targets beside the motor cortex, the neuroprosthesis was advanced via the left jugular in participant 1 and the right jugular in participant 2. With 16 sensors arrayed on an 8 mm by 40 mm nitinol scaffold, the neuroprosthesis self-expands to line the inner vessel wall. The device is then connected to a 50 cm flexible transvascular lead and inserted into an inductively powered telemetry unit, which is then placed in a subcutaneous pocket in the chest.
“The device stays like a tattoo inside the blood vessel,” said Dr. Oxley. “Cells grow over it. It lives within the wall of the blood vessel. We set up this trial primarily to assess for the formation of blood clots. We didn't see any in CAT scans at three months and 12 months.”
Once home, participants worked with a trainer, either physically or remotely, who instructed them to attempt various movements, including bilateral fist-clenching, foot-tapping, and knee extension.
“If the participants had residual muscle function associated with specific movement-attempts,” the paper noted, “they were instructed to only generate effort up to, but not more than, that required to generate an explicit contraction.”
As they attempted the movements, a machine-learning decoder analysed the resulting neuronal signals acquired from the implanted device to home in on which could be picked up most clearly and consistently. In conjunction with an eye-tracker for cursor navigation, the participants were taught to use a short click for keyboard selection and a long click to zoom or magnify the screen.
From day 86 onward in participant 1, and from day 71 in participant 2, they used the system unsupervised at home. Participant 1 achieved a typing task average click selection accuracy of 92.63 percent at a rate of 13.81 correct characters per minute with predictive text disabled. Participant 2 achieved an average click selection accuracy of 93.18 percent at 20.10 correct characters per minute. Both participants were able to text message, do online shopping, and manage finances.
Asked how close the device might be to reaching clinical practice, Dr. Oxley told Neurology Today: “Our primary goal is to achieve market approval with the FDA and bring the first implantable brain-computer interface to patients. I would broadly say we have a five-year horizon for achieving FDA approval. That's the scale of time that we can see neurologists having this as an option for referring their patients. This Australian study was the first component of the FDA pathway. Next will be a feasibility study in the United States, followed by a pilot study and a pivotal trial. We anticipate applying for an investigational device exemption very soon.” Ultimately, Dr. Oxley said, “I am hoping this is the birth of a new specialty called interventional neurophysiology.”
While welcoming the absence of adverse events seen so far in the two participants, Dr. Stavisky said more time and more participants would be needed to better understand the true risks involved.
”The procedure required general anesthesia, surgery to the chest for the transmitter, and being on blood thinners for at least 12 months. The vein also grows around the [device], which has benefits—it keeps the electrodes in place and stabilizes the signals—but raises the question of whether it can be removed later. I'm not yet convinced that this is overall much less invasive than other approaches that require opening the skull.”