In 2016, a paper published in the Proceedings of the Royal Society B challenged one of the most basic assumptions in neuroscience: that learning requires a nervous system. The paper showed that Physarum polycephalum, a single-celled organism with no neurons whatsoever, could learn through habituation and retain that learning for weeks. Even more astonishing, it could pass its memories to another individual by fusing with it. This article tells the full story of that discovery and explores what it means for our understanding of memory.
What Is Habituation?
Before diving into the experiments, it helps to understand habituation itself. Habituation is considered the simplest and most universal form of learning in the animal kingdom. It is the process by which an organism gradually stops responding to a repeated stimulus that turns out to be harmless.
You experience habituation every day. The ticking of a clock that you stop noticing after a few minutes. The feeling of clothes against your skin that fades from awareness. The sound of traffic outside your window that becomes background noise. In each case, your brain has learned that the stimulus carries no useful information and filters it out.
Habituation has been documented across the animal kingdom, from sea slugs to humans. It was considered a hallmark of organisms with nervous systems. The idea that a brainless single cell could habituate was not on anyone's radar.
Dussutour's Habituation Experiment
French biologist Audrey Dussutour at CNRS in Toulouse designed an experiment to test whether Physarum polycephalum could learn through habituation. The experimental design was simple and elegant.
The Setup
The researchers placed slime mold specimens at one end of a bridge that connected to a food source (oat flakes) at the other end. The bridge was the only path to the food. Here was the catch: the bridge was coated with a substance that slime mold finds repellent but that is actually harmless. Two substances were used in separate experiments: quinine (bitter, repellent) and caffeine (also repellent to slime mold).
Day by Day
| Day | Behavior of Naive Slime Mold | Interpretation |
|---|---|---|
| Day 1 | Strong avoidance. Slime mold barely touches the bridge. Crossing time: very long (6+ hours). | Normal aversion response to repellent substance. |
| Day 2 | Hesitant approach. Some extension onto the bridge, then partial retraction. Crossing time: long. | Beginning of exploration despite aversion. |
| Day 3 | More confident approach. Less retraction. Crossing time: noticeably shorter. | Aversion response weakening. |
| Day 4-5 | Crosses bridge with only slight hesitation. Crossing time: approaching normal. | Habituation in progress. |
| Day 6+ | Crosses bridge at normal speed, as if no repellent were present. | Habituation complete. Organism has "learned" the substance is harmless. |
Control groups confirmed that this was genuine learning, not mere fatigue or adaptation. Slime mold that had been habituated to quinine still showed normal avoidance of caffeine, and vice versa. The learning was substance-specific, exactly what you would expect from true habituation rather than a general decrease in responsiveness.
Why This Matters
Before this experiment, habituation had only been documented in organisms with nervous systems. Dussutour's work demonstrated that the capacity to learn from experience is not an exclusive property of neurons. It can emerge from the biochemistry of a single cell.
Where Is the Memory Stored?
If slime mold has no brain and no neurons, where does it store learned information? This question gets to the heart of what memory fundamentally is.
The answer appears to lie in the vein network itself. Research published after the initial habituation paper suggests that the memory of habituation is encoded in structural and biochemical changes within the veins of the plasmodium.
The Vein Structure Hypothesis
When slime mold encounters a repellent substance, certain veins contract and the organism changes its flow patterns. During habituation, these flow patterns gradually return to normal, and the vein architecture adjusts accordingly. The "trained" vein configuration is different from the "naive" one, and this difference persists even after the repellent substance is removed.
More recent work by Dussutour's group and collaborators (Kramar and Alim, 2021) has provided more detail on this mechanism:
- Exposure to a stimulus causes local changes in vein diameter throughout the network.
- These diameter changes alter flow patterns, which in turn affect the distribution of signaling molecules.
- The altered flow patterns persist because the vein walls physically remodel in response to sustained changes in flow.
- The result is a structural "imprint" of past experience encoded in the geometry of the network.
In a sense, the vein network acts as both the processing system and the storage medium. The memory is not stored in a separate location (like a brain stores memories in synaptic connections). Instead, the memory is the network. The shape and connectivity of the veins constitute the organism's record of past experience.
| Memory Feature | Animal Brain | Slime Mold |
|---|---|---|
| Storage medium | Synaptic connections between neurons | Vein network architecture |
| Encoding mechanism | Changes in synaptic strength (LTP/LTD) | Changes in vein diameter and flow patterns |
| Retrieval | Neural activation patterns | Cytoplasmic flow through modified veins |
| Duration | Seconds to lifetime | Days to weeks (habituation); longer unclear |
| Transferable? | No (not between individuals) | Yes, through cell fusion |
Memory Transfer Through Fusion
The most headline-grabbing finding from Dussutour's 2016 paper was the memory transfer experiment. The protocol was straightforward:
- Train one group of slime mold specimens to tolerate quinine (the "trained" group).
- Keep another group naive, with no exposure to quinine (the "naive" group).
- Fuse a trained specimen with a naive specimen by placing them in contact and allowing them to merge into a single organism.
- Test the fused organism's response to quinine.
The Results
Fused organisms containing one trained and one naive individual behaved as if they had been trained. They crossed quinine-coated bridges at speeds comparable to fully trained specimens, showing little or no avoidance response.
The critical control experiment involved fusing two naive specimens. These fused pairs showed normal avoidance of quinine, confirming that fusion itself did not cause habituation. The tolerance came specifically from the trained partner.
Further investigation revealed that the transfer required approximately three hours of contact between the two specimens. During this time, a vein connection had to form between them, allowing cytoplasm to flow from the trained organism to the naive one. If the specimens were separated before a vein connection was established, no transfer occurred.
Memory Without Neurons
In animal brains, memories cannot be transferred from one individual to another (despite decades of attempts). In slime mold, learned information moves freely between individuals through a physical vein connection. This suggests that the substrate of memory in slime mold is fundamentally different from neural memory, something chemical or structural that can be carried by the flowing cytoplasm.
Anticipatory Behavior
Memory in slime mold goes beyond simple habituation. In 2008, a team led by Toshiyuki Nakagaki (building on the legacy of his 2000 maze experiment) and Tetsu Saigusa demonstrated that Physarum polycephalum can anticipate periodic events.
The Experiment
Researchers exposed slime mold to cold, dry conditions at regular intervals (every 60 minutes). Cold and dryness cause the organism to slow down and retract. Between pulses, the organism resumed normal movement.
After several cycles, the researchers stopped the periodic stimulus. The slime mold continued to slow down at the expected times, even though no stimulus was applied. It had learned the rhythm and was anticipating the next event.
This anticipatory behavior persisted for several cycles after the stimulus was removed, then gradually faded. It could be reactivated with just one or two reminder pulses.
| Phase | Stimulus Present? | Slime Mold Behavior |
|---|---|---|
| Training (first 3-4 pulses) | Yes (cold every 60 min) | Slows down in response to cold |
| Testing (stimulus removed) | No | Still slows down every ~60 minutes |
| Extinction (extended absence) | No | Anticipatory response gradually fades |
| Reminder (1-2 pulses) | Yes (briefly) | Anticipatory response returns immediately |
This experiment demonstrated that slime mold can form temporal memories: it does not just remember what happened, but when it happened. This is a more complex form of memory than habituation and was previously thought to require at least a simple nervous system.
What Slime Mold Memory Teaches Us
The discoveries made by Nakagaki, Andrew Adamatzky (who explored the computational implications of slime mold's information processing), and Dussutour have collectively reshaped how scientists think about memory and learning.
Several key insights have emerged:
- Neurons are not required for learning. Habituation, the simplest form of learning, can occur in organisms with no nervous system at all.
- Memory can be stored in physical structures other than synapses. The vein network of slime mold acts as both a processing network and a memory substrate.
- Memory can be transferred between individuals. In slime mold, learned information flows through cytoplasmic connections, something impossible in neural systems.
- Temporal patterns can be encoded without a clock. Slime mold anticipates periodic events using oscillatory dynamics in its cytoplasmic flow, not a dedicated timekeeping organ.
- Intelligence may be more widespread than we assumed. If a single cell can learn, remember, and anticipate, then the boundary between "intelligent" and "unintelligent" life may need to be redrawn.
Open Questions
Despite the progress of the last decade, many questions remain unanswered:
- What is the exact molecular mechanism of memory storage in the vein network? Is it purely structural (vein diameters), or are specific signaling molecules involved?
- How long can habituation memory last? The longest tested duration is a few weeks, but the theoretical limit is unknown.
- Can slime mold learn more complex associations beyond habituation and temporal patterns?
- Does the sclerotium (dormant form) retain memories formed during the active plasmodium phase?
- Could the principles of slime mold memory inspire new forms of data storage or computing architectures?
These questions are the subject of active research in labs across France, Japan, Germany, and Australia.
Explore Further
To understand the biological infrastructure that makes memory possible, read Slime Mold Biology: The Giant Single Cell. For the broader context of slime mold research, see The History of Slime Mold Research.