Part II: Inserted electrodes opens up a new window into the brain

Mr. ZHANG, the 72-year-old paraplegic patient in the ward of the Department of Neurosurgery in the Second Affiliated Hospital, received interactive training with robotic arms and manipulators for more than 4 months, and the process of mind control fared increasingly well.

At this moment, Mr. ZHANG, peacefully sitting in his chair, is helping himself to a bottle of diet coke. Two years ago, he suffered a car accident and his limbs ended up utterly paralyzed. Fortunately, his brain was not damaged. He is fixing his attention and envisioning the action of reaching for coke in his brain. EEG signals are transmitted to decoders via electrodes implanted in his cerebral cortex. Then a robotic arm begins to rotate and reach for the bottle of coke on a nearby table. It adjusts its five fingers to hold the bottle tightly and pulls it back bit by bit until the straw in the bottle is exactly placed in Mr. Zhang’s mouth.

This is the intriguing part of the brain-computer interface. It can open up a new transmission channel in the case of spinal injuries and damage to the original neural pathway of the human body.

“At present, the brain-computer interface is primarily applied to the reconstruction of limb motor functions in patients with severe and irrevocable motor dysfunction such as ALS and paraplegia in clinical practice,” said WANG Yueming, deputy director of the Frontier Science Center for Brain and Brain-computer Integration and a professor from the Qiushi Academy for Advanced Studies.  

Previously, researchers at the University of Pittsburgh and the California Institute of Technology successfully utilized the brain-computer interface and manipulators to enable paralyzed patients to drink beer and eat chocolate independently. In these cases, electrodes were implanted into the patients’ cerebral cortex.

Mr. ZHANG is China’s first patient to have brain-computer interface electrodes implanted in his brain. “Two Utah microelectrode arrays are implanted into his cerebral cortex. Each electrode is 4 square millimeters and has 100 electrode pins, each of which corresponds to one or more neuron cells. At the other end of electrodes is a computer which can record neural signals emitted by the brain in real time,” introduced ZHANG Jianmin, director of the Department of Neurosurgery in the Second Affiliated Hospital.

Picking up a cup to drink water is extremely simple for a normal person. However, this action, in fact, involves an elaborate process of brain control: the movement of the arm must be accurate and stable, and the pressure exerted by fingers on the bottle must be appropriate…All of these call for an accurate and smooth signal command.

“EEG signals can also be collected by covering an electrode on the scalp. For example, one of the applications of the brain-computer interface is to wear a customized ‘hat’ and direct a model car to make simple forward and braking movements via EEG signals. However, such signals are inaccurate and unclear. When electrodes are directly inserted into the motor neuron cells of the brain, the signals obtained will turn out to be more direct, stable and abundant,” WANG Yueming added.

“When electrodes are directly inserted into neuron cells, it is like watching a football match in a stadium. You can see with your own eyes whether the player is volleying or heading the ball into the net. However, collecting EEG signals in other ways can be compared to ‘listening to’ a match outside the stadium. You can only get a general picture through cheers or boos,” said WANG Yueming.

Researchers at Zhejiang University developed the artificial intelligence algorithm to decode the EEG signals sent by Mr. ZHANG and then direct the movement of the robotic arm. For example, Mr. ZHANG can move the robotic arm left and right via motion imagination now. Only after repeated feedback learning can the decoder perform its task accurately. “This process entails immense interactive adaptation between a person and a computer,” said WANG Yueming, “When the robotic arm can red Mr ZHANG’s EEG signals, he is also learning how to use this new tool.”