Biosimilar memristor-based neuron developed by UNN scientists
The prototype developed at Lobachevsky University is based on microelectronic memristors and fully imitates the electrical activity of nerve cells, thus approaching the energy efficiency and speed of natural cells. Memristors perform the function of ion channels: they ensure signal transmission and reproduce the dynamics of brain cells.
Memristors can function as both neurons and synapses, making it possible to implement key elements of artificial neural networks on their basis. This opens the way to the development of miniaturised neural interfaces that extend or restore the functions of biological neural networks.
"Damaged neurons regenerate very slowly. Memristors can be used to produce devices that speed up regeneration. They will process brain signals and stimulate nerve cell activity in real time. For example, a neurochip with dozens of biosimilar neurons shows promise for repairing damaged neural networks in the spinal cord. The high speed of signal processing in memristors makes it possible to develop neurointerfaces that will predict epileptic seizures before the onset of pathological neuronal activity," explained the author of the study, Ivan Kipelkin, a junior researcher at the UNN Laboratory for Stochastic Multistable Systems.
Lobachevsky University scientists have created a memristor neuron based on one of the known simplified mathematical models, which will allow researchers to achieve more economical and efficient hardware implementations of such devices.
Prototypes of memristor microcircuits are developed at the UNN Memristor Nanoelectronics Laboratory together with technology partners, the Sedakov Research Institute for Measuring Systems and the Molecular Electronics Research Institute.
"The quality of our memristors is at the level of world laboratories, with theoretical models that we immediately test in practice. The next step is to build a neural network of 28 memristor neurons on a single chip to model the functionality of the spinal cord," said Ivan Kipelkin.
The research was carried out by the UNN interdisciplinary team as part of the research program of the National Centre for Physics and Mathematics (Sarov). The results were published in Frontiers in Neuroscience, a leading journal in the field of neuroscience.
