Novosibirsk State Technical University (NSTU) is developing an approach to taking into account the individual characteristics of users to improve brain—computer interfaces. The innovative project in the field of neurotechnology is primarily aimed at helping people with disabilities.
The key idea of the project is to create a personalized control system for external devices based on the analysis of brain activity. The system adapts to the individual neurophysiological characteristics of each user.
The brain—computer interface (BCI) is a technology for direct information exchange between the brain and an electronic device without the need for direct physical interaction with it. Brain signals are recorded and converted into commands for controlling computers, robotic prostheses, or other electronic devices. BCI is an important area in the field of medicine and neurorehabilitation, as it allows patients with motor disorders to control external devices, and is used as auxiliary systems for communication, typing, and calling for help.
As explained by Alexey Kozin, senior lecturer at the Department of Data Collection and Processing Systems at NSTU-NETI, the system under development is based on an approach based on stable visually evoked potentials (SSVEP) — stable rhythms of brain activity that occur when a person looks at a light source flickering at a certain frequency. By focusing on the desired source, the user generates a clear signal in his brain, which the system reads and converts into a command.
According to the developer, traditional brain—computer interfaces based on the SSVEP paradigm use the same set of frequencies for all users. However, the brain response of different people to the same photostimulation frequency is significantly different: one person will have the best response at 10 Hz, another at 15 Hz, and the third has the clearest signal at twice the frequency. Personalization is important for the effective operation of the BCI.
"Each person's reactions are unique. The purpose of my work was to come up with a method that would take them into account. We select for each user a set of the most "responsive" frequencies, to which his brain gives the strongest and most stable signal. This significantly increases the speed, accuracy and reliability of control," said Alexey Kozin.
As part of the project, a hardware part was developed — a photostimulator for 5-9 commands (a device with a set of flickering LED panels). Each LED panel flashes at a user-selected frequency from 0.5 to 128 Hz. Focusing on one of them causes a corresponding response in the user's brain. Special software allows you to flexibly adjust the parameters of the frequency, brightness and color of the photo stimuli for each user. In addition, the author of the project has completed the printing and assembly of a robotic arm manipulator, designed a control unit for it. The robot arm executes commands recognized by the brain—computer interface: moves left/right, forward/backward, opens and closes the grip. The arm is controlled directly by commands from the neural interface. In real time, EEG signals are classified wirelessly and the corresponding commands are sent to the robot arm controller.
Thus, a hardware and software complex has been developed that combines all the necessary steps: recording signals from a headset, analyzing the user's brain response to a set of frequencies and compiling his "personal profile", configuring a photo simulator to optimal frequencies for the user, recognizing commands in real time and transmitting them to control a robot or other devices.
The project is primarily aimed at helping people with disabilities. The development makes brain—computer interface technology more efficient and accessible for practical applications in areas such as neuro-rehabilitation, assistive devices, and smart home environments. As the author of the project notes, since the concept of "brain—computer" interfaces is also penetrating the gaming industry, BCI personalization can also be applied in this area.
The developer's plans for the future include improving the software part of the personalized brain—computer interface.