A large number of experiments have shown that it is possible to read information from the brain. It is now feasible to automatically interpret patterns that occur in brain waves and use them to control robotic arms or wheelchairs. Even language spoken in thought or what has just been heard can be reconstructed relatively well by measuring and translating electrical activity in the motor or auditory cortex in conjunction with machine learning. The situation is different with pain: It is much more difficult to use electrodes in the brain to detect whether a patient is currently perceiving pain.
While many sensory impressions are processed in permanently assigned brain regions, the processing of pain is distributed over many parts of the brain and is accordingly difficult to detect. One speaks here of the so-called pain matrix of the brain. A research team at New York University School of Medicine has now achieved a breakthrough in this area. Not only can you detect pain with the aid of an implant, you can also switch it off fully automatically in a second step.
To read brain activity, they chose a well-known point on the pain matrix, the anterior cingulate cortex, a part of the belt curl that belongs to the limbic system. Here the brain also processes a lot of other information at the same time. In order to filter out the patterns relevant for pain recognition from this clutter of signals, they used a state space model. This is a mathematical method in which dynamic systems are represented in the form of matrices and vectors. In contrast to many similar experiments, no artificial intelligence is used here, but a statistical model.
Cybernetic control circuit instead of continuous fire
If the system detects pain, it immediately sends a signal to another brain region in the prefrontal cortex. Stimulating them means that the pain is less noticeable. However, in their animal experiments, the researchers do not use electrodes for electrical stimulation, but rather optogenetic stimulation. This means that the desired neurons are first made sensitive to light by genetic engineering and can then be stimulated by light signals. This has the advantage, among other things, that no direct electrical contact with the relevant neurons is necessary and scarring and encapsulation does not occur so easily.
In fact, the researchers were able to show that this system can recognize acute pain and automatically suppress it. They gave genetically modified rats heat pain in one paw, which then withdrew their paw 40 percent more slowly than when the system was switched off. The result also shows that the pain is apparently not completely but only gradually suppressed.
Although deep brain stimulation has been used for more than 10 years to treat Alzheimer’s disease, depression and neuropathic pain that can no longer be treated in any other way, these systems stimulate the desired brain regions continuously and not only at the moment when the occurrence of pain is measured. For this purpose, an additional cybernetic control circuit was added to the rat brain, which would also be conceivable in humans at some point. A pain-relieved cyborg soldier is not yet within reach – but one who can endure pain better. More medically relevant than the suppression of acute pain is the question of whether such a system can also regulate chronic pain. That would be a game changer for countless pain patients, who can often only be helped to a limited extent by medication or whose medication sometimes has severe side effects.
Nociceptive and Neuropathic Pain
The research team also addresses this question in its paper, which is called Preprint in the journal Nature Biomedical Engineering has appeared. They induced an artificial allodynia in the rats – an oversensitivity to harmless touch, which is then perceived as pain. Painful hypersensitivity to touch, light or noise can often be accompanied by neuropathic pain such as migraines. In the experiment, the researchers were able to show that their system is able to alleviate such pain.
“Hypersensitivity is often associated with migraines, but it is not the migraine itself,” said neurophysiologist Markus Dahlem in an interview with Technology Review. The CEO of the migraine app developer M-sense finds the approach of reading pain from the anterior cingulate cortex very impressive, but in his opinion it is only suitable for detecting nociceptive pain. These are those that are reported to the brain via nerve tracts, for example because a part of the body is injured or inflamed. Such pain can also become chronic, according to Dahlem, but what we think of when we speak of chronic pain is mostly neuropathic pain, which arises in many different ways.
In addition, the question arises as to how it can be reliably established in animal experiments that the animals really feel different types of neuropathic pain at the moment. But if neuropathic pain is recognized and automatically suppressed with this system, there is still a long way to go from animal experiments to prototypes for human brains. These would also have to be changed optogenetically, which is only not a problem in experimental animals. It may be a few more years before chronic pain sufferers can begin to hope for a brain implant that will automatically regulate their pain.
Nevertheless, the idea of additional cybernetic control loops in the brain is fascinating. Controlling not only machines but other parts of the brain with data from the brain in real time invites you to do all kinds of thought experiments.