In September 2000, a team of Japanese researchers did something unusual: they chopped up a slime mold, scattered the pieces throughout a plastic maze, and watched what happened. Within hours, the organism had reconnected itself, filling every corridor. Then they placed food at the entrance and exit. Four hours later, the slime mold had withdrawn from every dead end and grown exclusively along the shortest path between the two food sources. A single-celled organism with no brain had solved the maze optimally.
The Unlikely Problem Solver
Physarum polycephalum looks like extra cheesy macaroni smeared across a petri dish. Bright yellow and gelatinous, it moves at speeds up to 5 centimeters per hour, flowing in an amoeboid fashion across whatever surface it encounters. Despite being a single cell, it can grow as large as 900 square centimeters and contains millions of nuclei floating freely inside its protoplasm. Biologists classify it as a protist, a taxonomic group one researcher describes as "everything we don't really understand."
This organism has existed for 600 million to 1 billion years, evolving long before brains or nervous systems appeared. Which makes its problem-solving abilities all the more puzzling.
How Memory Works Without a Mind
The maze experiment, conducted by Toshiyuki Nakagaki at Hokkaido University, raised an obvious question: how does an organism without a brain remember where it's already been? The answer turns out to be elegantly simple.
Chris Reid at the University of Sydney discovered in 2012 that slime molds use "externalized spatial memory." As they move, they leave behind translucent slime trails. The organism avoids areas where it has already traveled by detecting its own slime, effectively reminding itself to explore new territory instead of retracing old paths.
Reid tested this by presenting slime molds with U-shaped barriers blocking their path to food. When the dish was clean, 23 of 24 slime molds found the food. But when researchers pre-coated dishes with slime before the experiment began, only 8 of 24 succeeded. The existing slime confused their navigation system, demonstrating that the trails weren't just waste products but functional memory storage.
The slime mold doesn't need to know it's solving a maze. It simply follows local rules: move forward, avoid your own trail, grow toward food. The solution emerges from these simple behaviors repeated across the organism's sprawling network.
Designing Tokyo's Railways
The maze experiments were impressive, but researchers wondered if slime molds could tackle real-world engineering problems. In January 2010, they published results that suggested the answer was yes.
The team arranged oat flakes—a slime mold delicacy—in patterns matching the locations of major cities around Tokyo. They placed a slime mold on the position of Tokyo itself and watched. Within hours, the organism began refining its pattern, strengthening tunnels between oat flakes while other connections disappeared. After about a day, the network looked almost identical to Tokyo's actual rail system, with stronger, more resilient tunnels connecting centrally located points.
The slime mold wasn't just finding any solution. It was finding efficient solutions that balanced competing demands: short total length, redundancy in case of damage, and strong connections between major nodes. The same approach successfully replicated highway systems in Canada, the U.K., and Spain.
Andrew Adamatzky at the University of the West of England Bristol has proposed using either slime molds or computer programs mimicking them to plan future roadway construction. Researchers can simulate real-world constraints by using deterrents like salt or bright light to represent obstacles such as mountains or water bodies. The slime mold routes around them naturally.
An Organism That Tells Time
Perhaps the strangest discovery came from experiments on temporal learning. Tetsu Saigusa and colleagues subjected slime molds to unfavorable dry conditions every 30 minutes. The organisms slowed down during these periods to conserve energy, which made sense as a survival response.
Then the researchers stopped changing the conditions. The slime mold's pace still slowed every 30 minutes spontaneously. It was anticipating an event that was no longer happening, based purely on the rhythm it had experienced before. The effect worked at intervals of 30, 60, and 90 minutes, though only about half of slime molds showed the spontaneous slowing.
The mechanism likely involves the rhythmic pulsating of cytoplasm throughout the organism. The membrane constricts and relaxes with a period of 1 to 5 minutes, creating an internal oscillator that can apparently track longer intervals. Without anything resembling a brain, the slime mold had developed something resembling memory of temporal patterns.
Nutritional Wisdom
The slime mold's intelligence extends to dietary choices. Audrey Dussutour at the University of Paul Sabatier in France demonstrated that these organisms survive best on a diet of two-thirds protein and one-third carbohydrates. When presented with 11 different food options with varying protein-carbohydrate ratios, slime molds consistently selected the piece with optimal nutrient balance.
This isn't trial and error over generations. Individual slime molds make these choices, somehow sensing the nutritional content of food sources and selecting appropriately. The organism has no central awareness of the overall problem it's solving, yet it produces solutions that would require considerable calculation if approached consciously.
What Counts as Thinking?
Chris Reid puts it plainly: "Slime molds are redefining what you need to have to qualify as intelligent." For centuries, we've assumed that solving complex problems requires something like a brain—centralized processing, memory storage, decision-making architecture. Physarum polycephalum suggests otherwise.
The slime mold operates through distributed processing. Each part of the organism responds to local conditions: nutrient gradients, chemical signals, the presence of its own slime trails, the rhythm of cytoplasmic oscillations. No single part coordinates the whole. Yet the whole solves problems that stump human engineers.
This has implications beyond biology. The algorithms that mimic slime mold behavior are now being applied to network design, routing problems, and optimization challenges. Sometimes the most efficient solution doesn't require understanding the problem at all—just the right rules applied consistently at a local level.
A single cell, flowing across a surface for a billion years, has been thinking all along. We just needed to recognize what thinking could look like.