The primary reason that I and others favor science-based medicine, as opposed to the alternatives, is that science works. As Carl Sagan said, “Science delivers the good.” Science has other virtues – it is transparent and self-corrective also.
Recently two unrelated news items have provided an opportunity to compare a scientific vs a pseudoscientific approach to the same problem – that of communicating to patients who are locked-in.
Locked-in describes those who suffer from an injury or neurological disease that mostly paralyzes them, so that they cannot move or communicate. One scenario that leads to a locked-in state is a brainstem stroke, where patients are paralyzed below the eyes – they can only blink and move their eyes, but nothing else. Widespread trauma can lead to a similar situation. ALS, which leads to progressive loss of motor neurons, can also result in total or near total paralysis.
Traditionally communication with patients who are locked in has been limited to blinking once for “yes” and twice for “no,” or perhaps twitching a finger to indicate the same. The advent of computers has lead to more sophisticated communication, such as selecting letters by looking at them displayed on a computer screen, the subject’s choice noted by a built-in camera; or by moving a joy-stick with one finger. Physicist Stephen Hawking has managed to write books using similar methods.
But these methods are slow and cumbersome. They are invaluable to the patient, and I don’t want to diminish their importance, but the point is that there is tremendous room for improvement. The ultimate goal, of course, would be to produce real-time natural communication through either speech or writing.
Recently the case of Rom Houben has come to media attention. He is a man in his 40’s who has been in an apparent coma for the last 23 years following a car accident. However it was recently discovered that he is not in a coma or, more precisely, persistent vegetative state as was previously diagnosed but is locked in. Sometime during the last 23 years his brain recovered enough so that Mr. Houben regained consciousness. But he was mostly paralyzed by his injuries and had no way of indicating he was conscious to those taking care of him.
Eventually, a neurologist, Steven Laureys, a specialist in disorders of consciousness, reevaluated Mr. Houben with a PET scan and a more detailed clinical assessment and found that he appeared to be conscious, and therefore locked-in. Dr. Laureys reports that, although he was unable to move his eyes or hands, he could move his toe and they were able to communicate with him in the traditional but tedious yes/no method.
Then Linda Wouters came into the picture. She is a speech therapist who uses facilitated communication (FC), a technique of holding the hand of a client who cannot communicate and “helping” them type out words on a board or computer screen. However, FC has been thoroughly studies and discredited. It turns out that alleged communication through FC is simply the ideomotor effect (subconscious movements responsible for ouija board and dowsing movements) – the facilitator is doing all the communicating, even if they are not aware of it.
Video of Wouters writing with Houben’s hand tell the tale. In some videos Houben is not even looking at the screen, and may not even be awake. Wouters claims she is detecting minute movements in Houben’s hand who is guiding her to the letters – meanwhile she is flying across the computer screen with Houben’s hand. Her claims are implausible to the point that they should be rejected, barring rigorous and solid evidence to substantiate them, which is lacking.
The claims of FC are implausible, and the scientific evidence demonstrates strongly that FC does not work – it is a sad self-deception. The FC claims of Wouters are highly implausible, and the video evidence strongly suggests that she is doing all the communicating, not Houben.
What this means is that Wouters, with FC, has stolen Houben’s ability to communicate with the world. Far from providing a method of communication to a person who is locked-in, it has robbed that person of any hope of communicating, and has also inflicted a cruel fiction on them and their family. If what we strongly suspect about FC in this case is true, then Wouters is now communicating in Houben’s name.
That is the legacy of FC, a disproved method that has descended into the abyss of pseudoscience.
Brain Machine Interface
By contrast, there are those who are trying to use science to truly give the ability to communicate to those who have lost it through neurological disease or injury.
Current methods for communicating with the paralyzed primarily involve exploiting what little voluntary muscle activity they have (whether in the eyes, hands, or feet) in order to control a computer to produce words. However, some paralyzed patients lack even the minimal motor activity necessary to control such devices.
Another option is to use brain-wave measurements (electroencephalogram – EEG – measurements) to control a communication device. This could theoretically allow for a person with zero motor function to communicate by thoughts alone.
There has been steady progress in brain machine interfaces (BM() in the last decade. They involve using scalp surface electrodes, or electrodes implanted on or in the brain itself, to read the electrical signals from the brain. Those signals are then interpreted by a computer and sorted into one of a few distinct states. Computer software can then use those states to designate different letters, move a cursor around a screen, or operate equipment.
The key features of such systems are the number of different states that can be distinguished and the accuracy of the computer’s interpretation. So far such systems can only distinguish a small number of states (3-4), require extensive training, and have accuracies in the 60-90% range.
Like many new technologies, this is modest, and not any better than motor controlled systems (for those who have any motor control). It is often the case, however, that new technology does not have any advantage over older but more mature technology, except that it has more potential to improve in the future.
Recently another team has published the results of their research with BMI. A team led by Frank Guenther at Boston University have implanted electrodes in the speech area of the brain of a man who is locked in. These electrodes connect to radio transmitters which send signals to the external receiving device – so there is no need to have any wires going through the skin (a setup for infection). The implanted transmitters can be recharged through an external coil.
The computer that receives the signals interprets them as intended speech, and then makes sounds based upon its interpretation. The patient can then use the sounds as feedback, to fine tune their control.
They have now published their initial results. The patient has been trained to produce three distinct vowel sounds with his mind alone. This is an important proof of concept, but is very modest in practice. It is not sufficient for speech communication.
Other researchers have achieved similar results – distinguishing just a few states. So far it seems like the most practical application is moving a cursor across a screen. Up, down, left, and right are all that is needed in order to land the cursor on the desired letter or word.
Guenther and his team hope to increase the number of electrodes and therefore the precision of their system, to add more phonemes to its repertoire. They optimistically project that within 5 years their system will allow for some type of speech. We’ll see. I have learned to be skeptical of the 5 year optimistic projections of researchers – essentially they are saying that they will make impressive progress within the next funding cycle.
But that aside, this approach have obvious potential and progress is being made. Whether it is in 5 years or 20 or 50, progress is encouraging and it is likely we will cross significant functional thresholds in the future.
Or maybe, before this approach has a chance to mature, it will be supplanted by some other method that no one has thought of yet. The science will take time to work itself out.
The differences between FC and BMI are glaring. FC takes place all in the mind of the facilitator, who claims to have a highly developed skill to interpret movements that no one can see or detect. The claims are unbelievable, were made prior to meaningful research, and persist after they have been shown to be false by rigorous science.
BMI, on the other hand, is following a slow and steady research path with transparency. The claims that are being made are cautious and responsible. Sure, researchers are optimistic about the potential of their research – we expect that from researchers who need to be constantly on the prowl for competitive funding. But the only claims that matter are those they publish in the peer-reviewed literature, and they are dry and precise.
The bottom line is that getting the science right is what matters. Doing good science will help patients who are locked in. Pseudoscience will only exploit and further victimize them.
We owe it to them to remain dedicated to quality science as the standard of care in medicine.