How a bacterium steers toward food
The clearest example of a single cell making a decision is a swimming bacterium looking for food. Escherichia coli propels itself with rotating flagella in a pattern biologists call 'run and tumble': it swims in a fairly straight line (a run), then briefly tumbles to pick a new random direction, then runs again. The decision the cell makes, thousands of times a minute, is simply how long to keep running before it tumbles.
It makes that choice by remembering. A bacterium is far too small to sense a chemical gradient across its own body, so instead it compares the concentration it detects now with the concentration a moment ago. If food is getting more plentiful, it suppresses tumbling and keeps going; if not, it tumbles and tries a new heading. This short-term chemical memory is built from a handful of proteins (receptors feeding a kinase called CheA, which sets the state of a switch protein called CheY), and the result is a cell that reliably climbs toward what it needs and away from what harms it. Bacteria also act together: through quorum sensing, described by Bonnie Bassler and colleagues, they release and count signaling molecules and switch on group behaviors only once enough neighbors are present.
The brainless problem-solver
Move up in size and the decisions get startling. Physarum polycephalum, a slime mold, is essentially one giant cell with many nuclei and no nervous system at all. In 2000, Toshiyuki Nakagaki, Hiroyasu Yamada and Agota Toth reported in Nature that when they placed this organism in a maze with food at two ends, it withdrew from the dead ends and settled on the single shortest path connecting the food. A decade later the same group let it grow on a map of Tokyo with oat flakes marking cities, and it produced a transport network strikingly similar to the real rail system.
The slime mold has no plan and no memory in any familiar sense. It solves the maze by a physical process: tubes carrying more nutrient flow grow thicker while unused tubes wither, so the network as a whole converges on an efficient solution. That is exactly the point. A 'decision' here is not a thought; it is a system settling on one option among many because of how its parts push on each other.
Can a single cell learn? An honest look
If cells decide, can they learn? Here the science is genuinely unsettled, and it is worth being careful. In 2016 Romain Boisseau and Audrey Dussutour reported that Physarum shows habituation: forced to cross a bridge laced with harmless but bitter quinine or caffeine, it hesitated at first but within days ignored the bitterness and crossed at full speed, then grew cautious again after a period of rest. They ruled out simple fatigue, which makes this one of the best cases for learning in a brainless organism.
Other claims are shakier. The ciliate Stentor roeselii, famous from H. S. Jennings' century-old experiments, was restudied in 2019 by Joseph Dexter and colleagues, who confirmed it runs through a graded sequence of escape moves but found it did not fit any tidy model of learning. So the honest summary is this: single cells clearly sense, integrate competing signals, and act toward goals, and some show simple memory. They almost certainly do not think or feel. 'Decision' is the right word only if we strip it of any suggestion of an inner life.