Verified by 
Artificial Intelligence
Can AI Explain How Simple Organisms Make Decisions?
Physicists use AI machine learning to model chemotaxis in single-cell organisms.
Posted May 11, 2021 Reviewed by Kaja Perina
How do living organisms that lack a brain or nerve cells make decisions? In a new study published in May 2021 in the Proceedings of the National Academy of Sciences of the United States of America (PNAS), scientists used artificial intelligence (AI) machine learning to model how small organisms developed the ability to navigate their environment.
The research was conducted by lead author Andras Zött in collaboration with Gerhard Kahl, Maximilian Hübl, and Benedikt Hartl at the Institute for Theoretical Physics at Technische Universität Wien in Austria. The team of scientists sought to computationally model how simple, single-celled organisms move and navigate without a brain or nervous system.
In biology, there are two main classes of cell types—prokaryotes and eukaryotes. Prokaryotes such as bacteria and archaea lack a membrane-bound nucleus and organelles. Eukaryotes can be unicellular or multicellular and are larger and more complex than prokaryotes. They have both membrane-bound nucleus and organelles. Plants, algae, fungi, protists, and animals are eukaryotic. Examples of protists include amoebas, paramecium, algae, and slime molds. Single-celled or unicellular organisms are either classified as one or the other.
“Various microorganisms and some mammalian cells are able to swim in viscous fluids by performing nonreciprocal body deformations, such as rotating attached flagella or by distorting their entire body. In order to perform chemotaxis (i.e., to move toward and to stay at high concentrations of nutrients), they adapt their swimming gaits in a nontrivial manner,” wrote the scientists. “Here, we propose a computational model, which features autonomous shape adaptation of microswimmers moving in one dimension toward high field concentrations.”
In microbiology, chemotaxis is the movement and migration of cells towards or away from chemical stimuli. It is a biological process that enables microorganisms to increase their bioavailability and plays a role in physiological processes and disease states.
“We apply a genetic algorithm to the internal decision-making machinery of a model microswimmer and show how it learns to approach nutrients in static and dynamic environments,” wrote the researchers. “Strikingly, the emerging dynamics resembles the well-known run-and-tumble motion of swimming cells.”
The project was in two phases. In the first phase, the scientists applied a genetic algorithm called NEAT (NeuroEvolution of Augmented Topologies) to show that a microswimmer is capable of learning to swim without a chemical field in a viscous environment and to build artificial neural networks (ANNs) to describe the decision making tied to the motion. The researchers applied a reinforcement learning algorithm to optimize the microswimmer’s movement strategy in a viscous fluid.
“We demonstrate that complex locomotion and navigation strategies in chemical environments can be achieved by developing a surprisingly simple internal machinery, which in our case, is represented by a small artificial neural network,” wrote the researchers.
In artificial intelligence computing, artificial neural networks are a type of model with architecture that is somewhat inspired by the biological brain. ANNs have networked artificial nodes that are analogous to neurons that have weights. The connections, also called edges, are like synapses in the brain. The weight of the neuron and edges are adjusted during the learning process. The architecture of an artificial neural network contains layers of nodes. Specifically, an ANN will have a layer of input nodes, a layer of output nodes, and one or more layers of hidden nodes.
“The run-and-reverse behavior in our system is an emergent feature which sustains in the absence of a chemical field (as observed, for example, for swimming bacteria) without explicitly challenging the microswimmer to exploit search strategies in the absence of a field during training,” the researchers wrote. “From an evolutionary point of view, it makes sense that bacteria have learned this behavior in complex chemical environments.”
In the second phase, the scientists address the “challenging problem of finding a policy which allows the microswimmer to navigate on its own within a complex environment.” The team discovered that “ANNs of both spatial and temporal sensing methods are able to generalize their capability to predict the chemical gradient over a much wider range of parameters.”
“Our work demonstrates that the evolution of a simple internal decision-making machinery, which we can fully interpret and is coupled to the environment, allows navigation in diverse chemical landscapes,” wrote the researchers. “These findings are of relevance for intracellular biochemical sensing mechanisms of single cells or for the simple nervous system of small multicellular organisms such as Caenorhabditis elegans.”
Copyright © 2021 Cami Rosso All rights reserved.
