United States: In a major development, the researchers of Tel Aviv University and Tel Aviv Sourasky Medical Center could successfully read a patient’s thoughts with the help of a speech-brain-computer interface.
More about the finding
As per the press release, the participating patients were silent and imagined saying a syllable. The researchers implanted depth electrodes deep inside the patient’s brain, transmitting thoughts in the form of electrical signals.
Further, the computers detected those signals and transformed them into vocalized forms, as interestingengineering.com reported.
As per the experts, this successful attempt has provided a rare glimpse into the depths of the human brain and helped a paralyzed man express his thoughts in the real world.
More about the experiment
Firstly, the study was a pilot project by Ariel Tankus, PhD, in the School of Medical and Health Sciences of TAU and Tel Aviv Sourasky Medical Center (Ichilov Hospital), Ido Strauss of the SMHS of TAU and the Functional Neurosurgery Unit of Ichilov hospital.
They involved an epileptic patient since they belonged to a group of patients, which is a subset of epileptics, that do not respond well to medications and thus call for neurosurgical techniques, according to the study leader, Tankus.
Additionally, he belonged to a sub-division of this group, where the emphasis appears to be on the deeper part of the brain instead of its exterior cortex.
The focus relates to “the source of the ‘short’ that’s sending powerful electrical waves through the brain,” as Tankus explained.
To reach the point sources of the problem, the doctor involved has to implant electrodes into the deep structures of the brain.
Conduction of the experiment
Doctors implant electrodes into the deep structures of their patients’ brains to find the exact location of the source of the problem.
Thus, the patient that was employed in this study already had these ‘brain readers’ installed and was in the hospital expecting a subsequent seizure, which such an epileptic has to experience,
While the patient is having a seizure, the medical practitioners can determine where the focus is to operate, as interestingengineering.com reported.
After this, the electrodes were placed, and the patient agreed to the procedures; the researchers asked the patient to say aloud /a/ and /e/ syllables.
To obtain the results for these syllables, they employed deep learning and machine learning to capture the brain signals connected with them and also taught these AI models where certain cerebral cells are located.
According to the press release, they were able to distinguish the electrical activity corresponding to the desire to say /a/ and /e/. The computer was able to recognize these electrical patterns.