Why slime molds can solve mazes better than robotsS

Despite not having brains, slime molds are able to remember where they've been. They avoid oozing back over paths that didn't lead them to food by detecting "memories" in the trails of slime they leave behind. It's possible that these slimy recollections were what led to the evolution of more complex cognitive structures in animals.

Physarum polycephalum is a species of slime mold that moves around (at a pace of about 1mm/hour) by shifting the cellular fluid within itself, sending tendrils out in a variety of directions. It responds to various environmental stimuli, like light or the presence of food, and it creates links between food sources to shuttle nutrients around. These links famously mimic the efficiency of intercity rail systems.

Slime molds don't move around completely at random. In fact, they're able to solve mazes better than robots because they follow a simple pattern. If a slime mold moves to an area and doesn't find food or moisture there, it doesn't go that way again. It knows where it's been because it leaves a trail of slime behind. If it detects its own slime, it knows to head in a different direction. This eliminates backtracking and allows the slime mold to find an efficient route to a food source far more quickly than purely random movements would account for.

In a paper published in the Proceedings of the National Academy of Sciences (PNAS), researchers were able to quantify just how much a slime mold's memory improves its performance. They placed a slime mold in a U-shaped trap, a form of simple maze. On the other side of the U was a food source. Under normal conditions, the slime mold made its way to the food within the 120-hour time limit in 96 percent of trials. When the maze had been coated with slime prior to running the test (preventing the slime mold from detecting its own slime trails), the molds were successful about a third as often within the time limit. It was clear that the slime molds ended up wasting lots of time backtracking without their slime trail "memory" to guide them.

That's pretty cool by itself, but the researchers pointed out that such a memory system supports "the theory that an externalized spatial memory may be the functional precursor to the internal memory of higher organisms." A slime mold does pretty amazing things with a very simple stimulus/response system based entirely on chemical signals. A slime trail is a chemical pattern that tells the organism, "Don't go here." If that chemical pattern were somehow internalized into the organism's body, that could be considered a very primitive form of neural system.

This also gives hope to amoeboid sci-fi aliens. Slime molds can get pretty big. Big enough to gain sentience on some alien world in a far off galaxy? Why not? Hell, one of the researchers (Chris R. Reid) even broke out the old, "I, for one, welcome our new gelatinous overlords" line.

Source: "Slime mold uses an externalized spatial 'memory' to navigate in complex environments." Chris R. Reid, Tanya Lattya, Audrey Dussutour, and Madeleine Beekman. PNAS.
Photo: Chris R. Reid/PNAS.