To the trauma of losing the arm, many amputees have to add the phantom limb syndrome: although they no longer have it feel they are still there, they feel the tickling, the cold, the heat and also the pain. A pain that is sometimes constant, which assaults when one sleeps, forcing some to take morphine for years. For others, even that does not work. Now, a team of neuroscientists has used augmented reality and machine learning to get 14 of those amputees to feel less of their phantom pain.
The head of the neurorehabilitation laboratory at Chalmers University of Technology (Gothenburg, Sweden), recalls that there are several theories about phantom limb pain: basically, it is a series of rapid changes in the brain following amputation that entangle the circuits that intervene in pain. The scientists have done now is to undo those changes – to bring the brain to the moment before the amputation, when there was no pain.
That brain reversion has been followed by 14 amputees. All had phantom pain for 10 years on average, although some had been in pain for more than 30 years. All of them had also tried other techniques such as mirror therapy or motor imaging or directly implants, but without success. Four of them were taking potent analgesics every day for several years. The neuroscientist selected the most difficult cases in several clinics. They wanted to focus on patients with chronic phantom limb pain who had not responded to any treatment.
Since September 2014 those 14 participants have participated in 12 sessions. Before each of them they had to rate their phantom pain using several metrics, from intensity scales to pain frequency, to the different qualities of pain: burns, cuts, cramps etc. During the sessions, they were placed in front of a computer. An augmented reality system allowed them to see themselves on the screen and where their stump should appear was a virtual arm.
They placed a series of electrodes on the real stump that collected the activity of the muscles. Although biceps and triceps are far from the hand, the algorithms allowed that if the patient thought to open the hand, the virtual hand would open. And if I wanted to close it, a slight activation of the muscle would cause the limb to close on the screen. After training the algorithm (machine learning) and patients the amputees were able to manage their hand in a virtual world in which they appeared as an avatar. In another part of the session, they used their digital arm to play with a video game of auto racing.
What they were looking for was rob circuits and neurons of pain. The method, called motor execution of the phantom limb, was intended to undo the neural reorganization that occurred in the area of the motor and sensory cortex involved in feeling and moving the limb after amputation and expanding the area that until now was limited to active muscles Of the shoulder.
The results, published in The Lancet, are very promising. Of the 12 patients who felt constant pain, six began to have it intermittently. All the amputees also reduced their sense of pain. In the intensity, frequency or quality scales of pain, the mean reduction was about 50%. Intrusion cases, such as night pain or during daily activities, were also reduced by half. Of the four who had to take medicines all the time, two were able to reduce the dose. Most encouraging was that the improvement was maintained, albeit reduced, six months after the sessions.
Neuroscientist convinced that with more sessions, the reduction would have been greater. In fact, the phantom pain waned as the sessions progressed, without seeming to stabilize. One of the first patients whose pain was relieved to a normal life has the system at home and connects to the computer when he feels he is coming back.
The researchers themselves acknowledge that their results, while good, need to be taken with caution. This is a reduced sample and there was no other group of amputees with an alternative treatment to control the experiment. Therefore, their next objective is to expand the sample with participants from six countries, with amputees of lower extremities and counting on volunteers for the control group. Another formula that has been designed to test and improve his method is to make it accessible to everyone. That is why, although they are about to take a commercial version, both algorithms, software and hardware are open source and its use is free.