Study reveals organisms strategically toggle memory use based on available resources

In a laboratory at the University of Tokyo, researchers have mapped out something that every living thing knows in its bones: remembering takes work. A brain storing yesterday's lesson, a bacterium tracking chemical gradients in its environment, a foraging animal retracing its steps—all of it burns energy. The question that drove a team from the Institute of Industrial Science, the University of Tokyo, and RIKEN was deceptively simple: when is that cost worth paying?

Their answer, published recently in Physical Review Letters, hinges on a principle as old as survival itself. The researchers built a mathematical model to watch how organisms make decisions when they can draw on two sources of information: what they sense right now, and what they remember from before. The catch is that memory maintenance drains resources. This creates a genuine tension. You can be accurate, or you can be efficient. Rarely both.

What emerged from the model was striking in its clarity. When resources run thin—when an organism is starving, or a cell is barely scraping by—the smartest move is to forget. Abandon the past. React only to what your senses tell you in this moment. But the moment resources become abundant enough, something shifts. Memory suddenly becomes worth the price. The organism flips strategies, not gradually, but abruptly. One threshold crossed and the entire calculus changes.

Lead researcher Takehiro Tottori describes the behavior as surprisingly dramatic. The model shows that organisms don't simply dial memory use up and down like a volume knob. Instead, they toggle it on and off based on what they can afford. This isn't recklessness or wisdom—it's optimization. It's the mathematics of survival.

But the story grows more textured when you add environmental noise into the equation. The researchers found that memory proves most valuable in a narrow band of conditions: when the sensory information coming in is moderately uncertain. If the world is sending you clear, unambiguous signals, memory adds little. Your senses are already telling you what you need to know. Conversely, if the information is too garbled, too unreliable, storing it doesn't help much either. You're just filing away garbage. It's in the middle ground—where the signal is real but imperfect—that remembering the past becomes genuinely powerful.

Senior author Tetsuya J. Kobayashi notes that this framework explains something biologists have long observed but struggled to account for theoretically: organisms don't always use memory, even when it could help. The decision to remember or forget depends on two variables working together—the resources available and how much uncertainty the environment is throwing at you. Neither alone tells the full story.

The theory aligns with recent behavioral studies showing that humans themselves adjust their reliance on memory based on how much mental energy they have to spare and how trustworthy their immediate perceptions are. We are not exempt from this calculus. We toggle memory on and off, consciously or not, based on what we can afford and what we need.

The implications ripple backward through evolutionary time. If organisms across every scale of life—from single cells to brains—face this same tradeoff between memory and resources, then the wild diversity of biological information-processing systems we see today makes sense. Some creatures evolved to be memory-rich and computationally sophisticated. Others stayed lean and reactive. Each solution was optimal for its own ecological niche, its own resource constraints, its own environmental texture. The question that haunted evolution was not whether to remember, but when.