🧫 Bacteria have a memory and can learn without a brain

The idea that living beings without a brain can learn and retain information seems counterintuitive. Yet a recent study from Carnegie Mellon University, published in PRX Life, proves that Escherichia coli bacteria are capable of storing data about their past and adjusting their growth accordingly.

This discovery challenges the classical view that their behavior depends only on current conditions, and opens avenues for how we should approach infections.


An experiment allowed researchers to track individual bacteria. By rapidly alternating between rich and poor nutrients, they measured the growth of each bacterium in real time. The results indicate that bacteria do not all react the same way: those exposed to frequent changes adapt faster than those raised in a stable environment.

The key element of this adaptation lies in the memory of environmental frequencies. Bacteria distinguish between fast or slow cycles of nutrients, a level of memorization more advanced than previously demonstrated. For Josiah Kratz, first author of the study, this means that bacteria can discriminate between different frequencies and adjust their behavior based on their history.

The transmission of these memories occurs across bacterial generations. In E. coli, a generation lasts between 30 minutes and one hour. Proteins produced during stress, such as nutrient deprivation, are inherited by descendants for up to two generations. Thus, a bacterium that has never experienced famine itself can behave differently if its grandmother underwent that stress. The inherited molecules allow descendants to retain information about environments they have not directly experienced.


The implications for human health are considerable. Until now, it was assumed that bacteria's response to antibiotics depended only on the type and concentration of the drug. But if bacteria retain the memory of past stresses — such as exposure to high temperatures or doses of antibiotics — their reaction to a treatment could be different. Clinicians may need to consider the environmental history of microbes to optimize therapies.

Another surprising aspect of this research is the connection with artificial intelligence. By developing a mathematical model of cellular processes, researchers discovered that the way the bacterium processes information corresponds to an architecture used in machine learning. Josiah Kratz indicates that biology and ai seem to have converged on a similar strategy. This suggests that learning can emerge from simple chemical reactions inside a single cell, without a nervous system.

Future work should explore whether this behavior extends to other stresses, such as antibiotics, and to other bacterial species. Researchers believe this phenomenon is likely widespread in the microbial world. Understanding how bacteria adapt to constant fluctuations in their environment — whether in the human gut, soil, or plants — is important for unraveling the mechanisms of life at the microscopic scale.