Researchers just found out two important things about phantom limb pain. First is that a “reorganization” of the wiring of the brain is the primary cause of phantom limb pain. The second is that using artificial intelligence techniques, phantom limb pain may be treated.
Researchers used a brain-machine interface to train ten amputees with phantom limb pain to control a robotic arm with just the use of their brains. What they learned is that when the individuals tried to control the prosthetic arm by associating the movement with their missing arm, their phantom limb pain increased. But, when they were trained to associate the movement of the prosthetic arm associated with their existing arm, their pain decreased.
Their results, reported in the journal Nature Communications, demonstrate that in patients with chronic pain associated with amputation or nerve injury, there are “crossed wires” in the part of the brain associated with sensation and movement, and that by mending that disruption, the pain can be treated.
The authors also note that the findings could also be applied to those with other forms of chronic pain.
“Even though the hand is gone, people with phantom limb pain still feel like there’s a hand there – it basically feels painful, like a burning or hypersensitive type of pain, and conventional painkillers are ineffective in treating it,” said study co-author Dr. Ben Seymour, a neuroscientist based in Cambridge’s Department of Engineering.
“We wanted to see if we could come up with an engineering-based treatment as opposed to a drug-based treatment,” he added.
The cause of phantom limb pain is theorized to be associated with faulty “wiring” of the sensorimotor cortex, which is the part of the brain responsible for processing sensory inputs and executing movements.
Seymour and his colleagues, led by Takufumi Yanagisawa from Osaka University, used a brain-machine interface to “decode the neural activity of the mental action needed for a patient to move their missing hand, and then converted the decoded phantom hand movement into that of a robotic neuroprosthetic using artificial intelligence techniques.”
“We found that the better their affected side of the brain got at using the robotic arm, the worse their pain got,” said Yanagisawa. “The movement part of the brain is working fine, but they are not getting sensory feedback – there’s a discrepancy there.”
The researchers then altered their technique to train the “wrong” side of the brain. When they were trained in this technique, the patients found that their pain significantly decreased. As they learned to control the arm in this way, it takes advantage of the ability of the brain to restructure and learn new things.
Noting that the effects are temporary and require expensive equipment, Seymour said, “Ideally, we’d like to see something that people could have at home, or that they could incorporate with physio treatments. But the results demonstrate that combining A.I. techniques with new technologies is a promising avenue for treating pain, and an important area for future UK-Japan research collaboration.”