Intelligence Without a Brain (Slime Molds & Single Cells)

The one idea

Problem-solving is not something a brain has. It’s something matter does when it’s arranged to store and act on information. A slime mold (Physarum polycephalum) is one enormous single cell with no neurons, yet it can find the shortest route through a maze to food. When researchers laid out oat flakes in the pattern of cities around Tokyo, it grew a feeding network that closely matched the real Tokyo rail system. Memory and intelligence, it turns out, are substrate-independent. They can run on cytoplasm just as they run on nerve cells.

The science

Physarum is a yellow, fan-shaped organism that, in its plasmodium stage, is a single cell with many nuclei and no brain or nervous system at all. In the now-famous 2000 maze experiment (Nakagawa & Toshiyuki Nakagaki), the mold first filled an entire maze. Once food was placed at two exits, it withdrew from the dead ends and concentrated its body along the shortest connecting path. In the 2010 Tokyo experiment (Tero, Nakagaki and colleagues, published in Science), oat flakes marked the positions of cities in the Tokyo region. The mold grew tubes between them that rivaled the actual rail network for efficiency, fault-tolerance, and total length, a result engineers normally reach with deliberate optimization math.

How does a brainless thing “decide”? Mechanism, not magic. The mold’s tubes thicken where nutrient flow is strong and wither where it’s weak. That’s a simple feedback rule, repeated everywhere at once, and it converges on efficient networks. The mold also leaves a slime trail behind it, an external memory marking where it has already been so it doesn’t waste effort re-exploring (Reid et al., 2012). And single cells can genuinely learn. The slime mold can be trained to ignore a harmless but off-putting substance (habituation), and ciliates like Stentor show graded, escalating responses rather than fixed reflexes. None of this requires a neuron. The book frames it well. “Memory doesn’t require neurons. It requires any system that stores information about past states to influence future behavior” (v10 l3584).

One honest caution for the room. It’s tempting to leap from “solves mazes” to “thinks like us.” Mainstream researchers are careful here. This is basal cognition, meaning sensing, remembering, and deciding. It is not consciousness, not understanding, and not evidence of a tiny mind. The Tokyo result shows that a cheap local rule can produce a near-optimal global design, which is remarkable on its own terms without needing to be more than it is.

What this changes about how you picture reality

We tend to file “intelligence” and “memory” under brain, as if a nervous system were the price of admission to problem-solving. The slime mold quietly demolishes that. Capabilities we thought were special to animals with neurons turn out to be ancient and general. They show up in a single cell, in bacterial colonies signaling electrically, in fungal networks, in plant roots negotiating which way to grow. Your own neurons aren’t the inventors of memory. They’re a refined, high-speed elaboration of tricks life was already using billions of years before brains existed (v10 l3660).

The earned awe here is not “everything is conscious.” It’s subtler and sturdier. The same logic that lets you find your way home runs, in slower and simpler form, in a puddle of yellow goo on a petri dish. Intelligence isn’t a rare spark added from outside. It’s what information-processing matter does at every scale, and you are the part of it that happens to be able to notice.

Two ways to see it

Put both of these in front of the room so people can react to the same phenomenon framed two ways.

• Clip / visual A, “Look what it can DO” (the awe framing): the time-lapse of the slime mold filling a maze and then pruning itself down to the shortest path, paired with the Tokyo-rail side-by-side (mold network beside the actual rail map). This is the wow. A brainless cell out-engineering a human transit system. Name it for the room as capability without a brain.
• Clip / framing B, “Don’t over-read it” (the skeptic framing): a researcher or science communicator explaining the mechanism. Tubes thicken with flow and thin without it, plus a slime-trail memory. No goals, no understanding, just a local rule run everywhere at once producing a global result. Name it as it’s physics and feedback, not a tiny mind. The tension between A and B is the discussion engine. Both are true at once, and sitting in that tension is the point.

(Optional third voice if you want to widen it: a clip on Michael Levin’s work, or on bacterial “microbial brain” biofilms, to show this isn’t a one-off oddity but a pattern across life.)

Discussion questions

1. When you watched the mold solve the maze, what word came to mind? “Solving,” “deciding,” “intelligence,” or something else? Why that word, and would the mechanism explanation change it?
2. If memory just means “storing information about the past to change future behavior,” what isn’t memory? A river carving a canyon? A scar? Where would you draw the line, and does it matter where you draw it?
3. The Tokyo result came from a simple local rule. Strong flow strengthens a tube, weak flow lets it die. Where else in life or society do you see complex, smart-looking outcomes emerge from dumb local rules?
4. Does learning that the mold is “just feedback and physics” make it less impressive to you, or more? What does your reaction tell you about what you actually find awe-inspiring?
5. We’re quick to call this “intelligence without a brain” and equally quick to insist “it doesn’t really think.” Are we protecting something by drawing that line, and if so, what?
6. If problem-solving doesn’t require neurons, what does that suggest about where else something brain-like could show up, in the universe or in machines?

Closing question

How do you feel about this science and its understanding of reality?

Take it further

• The Book, v10, the primary grounding for this session: the “Intelligence Without Brains” passage (l3576–3592), the Tokyo-rail and substrate-independence summary (l3650–3666, including “Intelligence is substrate-independent… What matters is how information is processed,” l3662), and the broader “Basal Cognition: Intelligence at All Scales” framing (l3638–3666). Bacterial “microbial brain” and electrical-signaling context sits at l3594–3611, and xenobots as a next step at l3668–3692.
• Nakagaki, Yamada & Tóth (2000), Nature, “Maze-solving by an amoeboid organism.” The original maze result. (book ref [S-258])
• Tero, Takagi, … Nakagaki (2010), Science 327:439, “Rules for biologically inspired adaptive network design.” The Tokyo rail experiment. Widely covered, and a strong, vetted external anchor for the session.
• Reid, Latty, Dussutour & Beekman (2012), PNAS, slime mold uses an externalized spatial memory (slime trail) to navigate. Good support for the “memory outside the body” point.
• Uncertainty flag: the book’s line that bacteria pass memories to offspring “for several generations without DNA changes” (l3656) is real but sits in an active research area. Present it as “observed in some systems,” matching the book’s own hedge at l3590 (“scope and generality vary”). Don’t overstate it to the room.

Visual notes

Lead the room with the slime-mold-solves-maze anchor visual. The time-lapse of the maze fill-and-prune is the single most legible “wow” we have, and it needs no narration to land. Hold the Tokyo-rail side-by-side (mold network beside the real transit map) as the second beat. That comparison is what converts “neat trick” into “wait, it matched human engineers.” Keep both on screen long enough for people to actually trace the paths with their eyes. If you use the skeptic framing clip (framing B), cut to a simple animation or diagram of the tube-thickening feedback rule so the mechanism is seen, not just told. The contrast between the gorgeous result and the dumb little rule is the whole emotional arc of the session. The room looks at the maze, then at the two networks side by side, then at the rule. Awe, then “how,” then back to awe.