While we tend to focus on SBM on exposing and routing out pseudoscience in medicine, it is also important to point out why science-based medicine is so effective. It may not be fashionable to say so, but good ol’ reductionist science is a powerful tool that can have a massive impact in people’s lives.
One story I have been following for years is the incremental advances being made in brain-machine interfaces (BMI), and here I am happy to report on yet another small but significant incremental advance. That is something else worth reminding the public about – science generally advances through accumulated baby steps, and rarely through “stunning breakthroughs”. But breakthroughs are better for headlines, even if they make for sloppy science communication.
A recent study in Nature reports the case of a patient suffering with advanced amyotrophic lateral sclerosis (ALS), which has caused him to be completely paralyzed, enough to be considered in a “locked-in” state. He cannot even move his eyes. ALS is a neurodegenerative disorder that affects motor neurons, causing muscle twitching, weakness, and eventual paralysis. The mind, however, is intact. While there are variable outcomes, in most cases patients get to the point where they are too weak to breath on their own. At that point they may choose to die comfortably or to be placed permanently on a ventilator to breathe for them.
For those with advanced cases of ALS a variety of devices have been developed to allow them to continue to have some function and to communicate, despite losing the ability to speak. Stephen Hawking was perhaps the most famous example of this. He used, by the end, a camera that tracked subtle facial movements to indicate his intentions. Intel actually helped develop a software system to anticipate which words he was trying to communicate and help with error correction. Patients who retain even the tiniest movement in a finger can use that to control a mouse. Some patient blow into a straw to make their desires known.
While these technologies are indispensable, they can be slow and tedious, and some patients do not have the residual muscle movements to use even these techniques. This is why the ultimate communication device would be one that communicates with the brain directly, and does not require any muscle movement at all. This is possible with a brain-machine interface. This technology has been slowly developing over the last few decades.
The good news is that research has already established all the necessary proofs of concept. Electrodes can accurate read brain signals and translate them into intentions. Patients can learn to control their thoughts in order to use the signals to control a mouse cursor or other interface. Brain plasticity allows the interface, with several weeks or months of training, to become progressively easier and quicker. Software has been developed to interpret these brain signals with increasing sophistication.
The primary limiting factor at this time is the hardware technology – the ability to place electrodes in intimate contact with cerebral cortex and to have that connection remain stable (despite the brain’s subtle pulsations) without forming scar tissue that would degrade the connection. Some of the research has been done with scalp electrodes (Hawking tried an electrode cap, but this did not work for him), which is safe and easy, but the resolution of the signals is poor and this limits the communication. Brain surface electrodes work much better, but they are invasive. Long term there is a risk of infection and scaring.
For this reason there are various teams developing the electrode technology, which is now arguably the most important piece of the BMI technology. There are microwire electrodes like thin hairs that can be placed within brain tissue, and there are electrodes that can be placed inside the blood vessels in the brain.
The Nature study used two microelectrode (not wire) arrays on the motor cortex, connected to a percutaneous (through the skin) connector and amplifier. The subject had previously lost all ability to communicate through eye movements or other methods. At first they instructed the patient to imagine moving their hand or doing other movements, but this did not result in any detectable alteration in brain firing. On day 86 after implantation they embarked on a new strategy – using auditory feedback. The brain signals were translated into an auditory tone, and on the first day of trying the patient was able to modulate the tone by voluntarily altering his brain activity.
By day 98 the patient was able to modulate the tone so that it matched a target. By day 106 the patient was able to modulate the tone to freely select letters to form words and sentences. The patient was able to achieve an accuracy of about 90% and was able to spell at 1.08 letters per minute. This is tediously slow, but obvious better than no communication at all.
Again, this is not a breakthrough but another in a long line of incremental advances, inching toward the day when brain-machine interfaces will be robust enough to allow for something approaching normal communication and other applications. The advance here is demonstrating that a patient can communicate through a BMI without any corresponding motor movement at all. This study also reinforces the fact that patients need feedback in order to learn to modulate their brain activity. In this case auditory feedback was used, but in some cases visual feedback can be used (such as seeing a cursor move on a screen).
In short, this technology works. At this point continued incremental advances are all that is necessary – improved electrodes, increased number of electrodes, improved software algorithms, improved robotics, and optimized training techniques. Predictive algorithms can significantly improve the speed of communication. “Self-driving” technology could aid in the control of robotic limbs to manipulate one’s environment and provide some mobility and independence. A computerized environment could allow significant control over lighting, electronic devices, doors, etc. With enough electrodes and connected devices, near normal functionality could be restored. This of course will take decades of research and development, but there does not appear to be any hard obstacles in the way.