Tokyo, Japan – Remembering the past can help living organisms make better decisions, but memory comes at a cost. Whether it is an animal searching for food or a single cell responding to its surroundings, storing and using information requires energy and resources. So, when is memory worth the effort?
In an article recently published in Physical Review Letters , researchers from the Institute of Industrial Science, The University of Tokyo and RIKEN have developed a new mathematical theory that addresses this question. The results have highlighted that the amount of available resources determined whether an organism relied on its memory, or made decisions solely based on current sensory input.
The researchers created a simplified model to simulate the estimation strategies made when an organism processes events in a changing environment. To do this, the organism was permitted to use both current sensory information and memories of past observations. However, as maintaining memory carries a cost, a tradeoff between accuracy and resource use was created.
“This tradeoff can produce surprisingly dramatic behavior,” says lead author Takehiro Tottori. “When resources are scarce, the best strategy is to ignore memory and react only to current information. But once enough resources become available, remembering suddenly becomes worthwhile, causing an abrupt shift to a memory-based strategy.”
The study also revealed that memory is most useful when sensory information is moderately uncertain. If information from the environment is very clear, memory provides little extra benefit. Similarly, if the information is too noisy and unreliable, storing it is not much help. However, between these extremes, remembering the past can significantly improve performance.
“These results help explain why organisms do not always use memory, even when it could in principle improve their decisions,” explains senior author Tetsuya J. Kobayashi. “Whether memory is useful depends not only on the resources available, but also on environmental uncertainty.”
The team noted that their conclusions were consistent with recent behavioral experiments, which also suggested that humans adjust how much they rely on memory depending on resource availability and sensory uncertainty. In doing so, the framework provides a theoretical explanation for why these shifts occur.
A Variety of biological information processing systems exist across scales, with each demanding varying levels of memory capacity. The findings therefore provide a theoretical basis for explaining how memory-consuming, yet sophisticated, biological computation systems like the brain emerged through evolution. To remember, or not to remember: that may be evolution’s question behind the diversity of biological computation.
###
The article, “Theoretical Analysis of Resource-Induced Phase Transitions in Estimation Strategies,” was published in Physical Review Letters at DOI: 10.1103/5ynb-7k4v .
The Institute of Industrial Science, The University of Tokyo (UTokyo-IIS) is one of the largest university-attached research institutes in Japan. UTokyo-IIS is comprised of over 120 research laboratories—each headed by a faculty member—and has over 1,200 members (approximately 400 staff and 800 students) actively engaged in education and research. Its activities cover almost all areas of engineering. Since its foundation in 1949, UTokyo-IIS has worked to bridge the huge gaps that exist between academic disciplines and real-world applications.
Physical Review Letters
Theoretical analysis of resource-induced phase transitions in estimation strategies
28-Jul-2026