Slime Mold Studio

Drop food at all three stations to grow a subway map.

Reinforced Possible Blocked

Grow a subway map

0 of 3 stations fed.

How the model works

This is bio-inspired AI. The model passes simulated flow between nodes. Tubes carrying more flow thicken, and lightly used tubes fade. The logic bench adds switches to the same network. If a reinforced path reaches the output pad, the circuit reads 1. The simulation runs in your browser.

Read the Physarum transport model paper

Life Sciences · Lesson 05

Slime Molds: Swarm Intelligence Without a Brain

Physarum polycephalum — the "many-headed slime" — is a single cell you can see with the naked eye. It has no brain, no neurons, and no central control center, yet it can find the shortest path through a maze, connect food sources with efficient networks, and even show signs of memory. Slime molds are living proof that intelligent behavior can emerge from simple local rules, repeated everywhere at once.

Quick ID

  • Not a fungus, plant, or animal
  • Kingdom: Amoebozoa, order Physarales
  • ~900 species known worldwide
  • Best-studied: Physarum polycephalum
  • One giant cell, thousands of nuclei, zero neurons

Biology & Life Cycle

A slime mold spends most of its life as a single wandering cell, but it gets there through a strange five-stage cycle that alternates between many tiny cells and one enormous one.

1 Spore A tough, dormant capsule that can survive dry conditions for months.
2 Swarm Cell In moisture, spores hatch into amoeba-like or flagellated single cells.
3 Fusion Two compatible swarm cells fuse into a single zygote.
4 Plasmodium The zygote's nucleus divides again and again without the cell splitting — one cell, many nuclei.
5 Sporangium When food or water runs out, the plasmodium grows fruiting bodies that release new spores.

Stage 4, the plasmodium, is the yellow, vein-like network most people picture when they hear "slime mold." Biologically it is a syncytium: a single cell membrane wrapped around thousands to millions of nuclei, with no internal walls dividing them. A plasmodium can grow to the size of a dinner plate — or, in record cases, a doormat — and crawl across a forest floor at roughly a centimeter per hour. Inside, the cytoplasm doesn't just sit still: it sloshes back and forth in rhythmic pulses called shuttle streaming, reversing direction every one to two minutes. This internal pumping distributes nutrients and chemical signals through the whole organism in seconds — and, as you'll see below, it also doubles as the organism's simplest possible "nervous system."

Bright yellow syncytial plasmodium of Physarum polycephalum spreading in network veins
Macroscopic Field Specimen Physarum polycephalum: The Active Plasmodium The vegetative feeding stage of Physarum polycephalum is a single giant multinucleate cell (syncytium). Rhythmic cytoplasmic shuttle streaming pulses back and forth through the branching tubular veins, allowing rapid distribution of nutrients and information across the entire organism without any central brain or nervous system. Photo: Bill Sheehan via Wikimedia Commons (CC BY-SA 3.0).
Delicate stalked sporangia fruiting bodies of a myxomycete slime mold
Reproductive Stage Myxomycete Sporangia (Fruiting Bodies) When food or moisture is exhausted, the plasmodium halts its streaming and metamorphoses into delicate, stalked sporangia. These structures produce dry, microscopic airborne spores capable of surviving environmental drought for months or years until moisture triggers germination into new swarm cells. Photo: Bernard Dupont via Wikimedia Commons (CC BY-SA 2.0).

Their Place in the World

Slime molds live almost everywhere there is shade and moisture — they've been found on every continent except Antarctica.

🌲 Habitat

Rotting logs, leaf litter, tree bark, and damp forest soil. They avoid direct light and dry heat, which shrivels the plasmodium.

🍽️ Diet

Bacteria, yeast, and fungal spores, engulfed whole through phagocytosis as the plasmodium flows over decaying wood and leaves.

♻️ Ecological Role

Microbial recyclers: by grazing bacteria and fungi, they help cycle nutrients back into forest soil, much like earthworms do at a smaller scale.

🏷️ A Taxonomic Wanderer

Because they form fungus-like spore-bearing fruiting bodies, slime molds were classified as fungi for over a century. Genetics eventually placed them in Amoebozoa — closer kin to amoebae than mushrooms.

🟡 Familiar Faces

Fuligo septica, the "dog vomit slime mold," is a bright yellow species that can appear on garden mulch overnight after rain.

🧬 Genome

Physarum polycephalum carries far more DNA per cell than a human cell — a reminder that genome size has little to do with complexity.

Swarm Intelligence: Solving Problems With No Brain

A single plasmodium behaves like a decentralized swarm packed into one cell. As it forages, it sends out thousands of finger-like tubes in every direction at once. Tubes that reach food thicken and speed up their internal flow; tubes that lead nowhere slow down, thin out, and are reabsorbed. That simple push-and-pull — reinforce what works, let go of what doesn't — is the same principle called stigmergy that lets ant colonies build efficient trails using only pheromones, with no ant ever seeing the whole picture. In Physarum, it happens without a single neuron.

This process leaves a trace: unlike ants, a plasmodium avoids crossing ground it has already explored because it leaves behind a non-living, slippery extracellular slime trail. That trail acts as a simple form of external spatial memory — a way to "remember" where it's already been, recorded outside the cell entirely. Researchers have also shown a subtler kind of memory: when a plasmodium repeatedly crosses a harmless but irritating substance on its way to food, it grows more tolerant over time — a basic form of habituation, and that learned tolerance can even be passed to a naive plasmodium when the two physically fuse together.

How Scientists Study Slime Molds

2000 · Nature

The Maze Experiment

Toshiyuki Nakagaki grew a plasmodium through a maze with food placed at two exits. Within about a day, the slime mold withdrew from every dead end, leaving behind the single shortest path connecting the food — an optimization problem solved with no eyes, no map, and no brain.

2010 · Science

Rebuilding the Tokyo Rail Map

Atsushi Tero, Nakagaki, and colleagues placed oat flakes on a map in the positions of Tokyo and its surrounding cities, with light (which Physarum avoids) marking mountains and water. The network the slime mold grew rivaled the real Tokyo rail system in efficiency, cost, and fault tolerance — a system engineers had spent decades refining.

Ig Nobel Prizes

Serious Science, Silly-Sounding Prizes

Both discoveries were honored with Ig Nobel Prizes — the maze study in Cognitive Science, the rail study in Transportation Planning — awards that celebrate research that "first makes people laugh, then makes them think."

Unconventional Computing

Living Computers

Researchers like Andrew Adamatzky have wired Physarum into logic gates, sensors, and even slime-mold-steered robots, using its growth and light-avoidance as a literal biological computing substrate.

Bio-Inspired Engineering

Borrowing the Algorithm

Engineers now write "Physarum-inspired" algorithms — reinforce, decay, avoid obstacles — to design resilient computer networks, transit systems, and disaster-logistics routes without any central controller.

Open Questions

What Counts as Cognition?

Because Physarum shows memory and habituation with zero neurons, it has become a key organism in debates over whether "learning" and "decision-making" require a nervous system at all.

Slime mold Physarum polycephalum navigating an agar maze to connect food sources
Laboratory Experiment Physarum Maze Navigation: Shortest-Path Computation In this replication of Nakagaki's pioneering experiment, Physarum polycephalum spreads through an agar maze, then selectively strengthens the tubes carrying nutrient flux while retracting from dead-end corridors. The result is the single shortest route connecting the food sources, calculated purely by local fluid-mechanical feedback. Laboratory experiment photo: Tim Tim via Wikimedia Commons (CC BY-SA 4.0).
Free-grow particle sandbox

This simulation models thousands of tiny "agents" using a local rule inspired by Physarum: sense three points ahead, steer toward whichever carries the strongest trail, then lay down more trail behind you. No agent knows the big picture — the network below emerges purely from that local rule, repeated thousands of times a second.

Click the canvas to drop a food source
Reading the map: dark = unexplored, amber = a reinforced tube, bright gold = a heavily-used route. Gold dots are food sources — the station preset places a cluster of food sources for the agents to explore.

Discussion Prompts

  • Does intelligent behavior require a brain? What would you need to see to convince you a system is "solving a problem" versus just following a rule?
  • Where else do you see stigmergy — indirect communication through a shared environment — in nature or human systems (ants, traffic, wikis, city streets worn into "desire paths")?
  • The Tokyo rail experiment compared a slime mold's network to one designed by professional engineers. What does it mean that both converged on similar solutions?
  • If a single cell with no neurons can show habituation and memory, how should that change the way we define "learning"?