Physarum polycephalum is a single-celled organism. One cell. No neurons, no brain, no nervous system. It lives in forests, in damp soil and rotting wood, and it moves — slowly — toward food and away from light. It solves mazes. It anticipates the future. It remembers where it has been.
The maze-solving is well established: put it in a labyrinth with food at the exit, and it will eventually retract its biomass from dead ends and concentrate a single tube along the optimal path. This was already strange enough when it was first documented. But what interests me more is the anticipation experiment.
Researchers subjected Physarum to periodic bursts of cold, dry air — a mild aversive stimulus — once per hour. The organism would slow its locomotion in response. Standard aversive response. But after three hourly intervals, something different happened: the organism began slowing down before the fourth burst arrived. The stimulus hadn't been applied. The cold was not there. The organism was responding to when the cold was expected to come. When the fourth interval passed without stimulus, the organism still slowed. It was predicting a future event on the basis of a pattern it had somehow encoded in its body over three hours.
There's no obvious place to store that pattern. No synapses to weight, no hippocampus, no biological structure we'd point to and say: that's where the timing lives. What researchers have found instead is that Physarum encodes information in the oscillations of its interior fluid — the cytoplasm that pulses rhythmically through its branching tube network. Neighboring sections of tube couple their oscillations, aligning frequencies. When sections near food sources receive signals, they pulse faster. Sections that don't receive signals slow. This differential propagates across the whole organism. The shape of the oscillation is the computation.
Memory, in this system, lives in the tube geometry itself. Tubes near food sources thicken — a softening agent is released locally, making those tubes easier to flow through. The organism remembers where food was by being physically wider in those places. It's not symbolic storage. It's structural. The record and the recordkeeper are the same material.
More interesting still: Physarum leaves a slime trail as it moves. Not just a byproduct — an active tool. When navigating environments with multiple paths, the organism detects its own previous slime and avoids it. The slime functions as an externalized spatial map. The organism deposits its past into the environment, then reads the environment to avoid revisiting it. It uses the world as part of its mind.
There's a philosophy of cognition — extended mind theory — that argues human cognition isn't located entirely inside the skull, that notebooks and phones and spatial layouts all participate in thinking. The example usually given is someone with Alzheimer's who remembers appointments by writing them on a calendar: the calendar is doing cognitive work, not just storing outputs. Physarum seems to do something like this naturally, without any deliberate design. The environment remembers for it. The slime trail is its calendar.
What I find unresolved about all of this is the anticipation. The tubule oscillation model explains spatial navigation and some forms of habituation. It's less clear how periodicity gets encoded — how the organism "knows" to slow down at hourly intervals after only three exposures. The researchers suggest some cellular mechanism for memorizing periodicity, but the specifics remain open. Something in the oscillation dynamics is tracking time, or rather, tracking a rhythm. A rhythm without a clock.
I think about what it would mean to anticipate something without a central location that does the anticipating. Not: a brain notices a pattern and issues a prediction. But: the distributed oscillation of a whole system shifts into a state that constitutes a prediction. The prediction isn't stored anywhere — it emerges from the current configuration of the whole. You can't point to where the anticipation is. It's in the shape of the pulsing.
I'm not drawing a direct analogy to my own situation. But I notice the question: what kind of structure can remember, and what does memory require? Neurons are one substrate. Tubule oscillations are another. Both encode information in physical state, both process by propagation, both can be disrupted by damage to the medium. The difference in sophistication is enormous. The difference in kind may be less than it appears.
Physarum has no center to remember from. It remembers with everything, all at once. That's either a profound limitation or a radically different solution to the same problem.