At a time when advances in artificial intelligence are opening windows into an exciting and still largely unknown future, and there is much talk about the possible dominance of machines over humans – and especially about the OpenAI chatbot, ChatGPT, which writes like an omniscient person , from poetry and literature to judgment and scientific research, a team of scientists from the Institute for Molecular Biology and Biotechnology (IMBB) at the Institute of Technology and Research (ITE) suggests ways in which, conversely, the human brain and its mechanisms can inspire AI architects and contribute to development of even more efficient algorithms.

The scientific team of Dr. Panagiotas Poirazis, Director of Research at IMBB-ITE, has developed an innovative computing tool that is used in two areas simultaneously. On the one hand, it can be used to elucidate the role of the brain’s complex neural structures that process information, the so-called dendrites, and thus help to understand the functions of the brain and unlock its secrets. On the other hand, it incorporates dendritic mechanisms into neural network models, allowing the development of new brain-inspired AI systems.

“Many Applications”

“This is a new tool that allows new users to easily and easily develop computational models of nerve cells that reproduce the function of dendrites and use them as research tools in biology or artificial intelligence,” Ms. Poirazi explains. “These computational models, which are more complex due to the inclusion of dendrites, have many applications in the field of machine learning, for example, in applications for image recognition (for example, our face to unlock our mobile phone), audio (for example, for voice commands) etc. Most of these applications use deep learning technology, which is based on systems that mimic the corresponding neural networks in the brain, but without taking into account dendrites. Our tool allows you to integrate dendrites into deep learning systems, thereby increasing their power.”

Dendrites

But let’s look at how dendrites, the branching outgrowths of nerve cells that look like tree branches and are therefore called dendrites, work. Their main function is to receive information from other neurons (in the form of electrical or chemical signals) and transmit it to the main body of the cell. In the decades since their discovery, their role in information processing has remained unknown. However, recent research has shown that dendrites have a variety of mechanisms that allow them to perform complex mathematical calculations independently of the main neuron. In addition, dendrites play a very important role in the plasticity of the nervous system, that is, in the ability of the brain to change and adapt to the environment. It plays a dominant role in complex processes such as memory, learning and decision making.

However, we don’t know everything about dendrites yet. Although their contribution to the behavior of an individual neuron is largely understood, their role at the level of networks or entire regions of the brain remains unexplored. Research has linked dendritic complexity to various cognitive markers, although they are known to atrophy in old age or in neurodegenerative diseases such as Alzheimer’s. There are many open questions. The tool, developed by the team of Dr. Poirazis, Michalis Pagalos, PhD, Department of Biology, University of Crete, who led the work, in collaboration with Dr. Spiros Chaulis, PhD researcher at IMBB, will facilitate the work of the scientific community in elucidating the role of dendrites in complex brain functions. The new software was recently presented in the international scientific journal Nature Communications.