The 120-Year-Old Debate: Can Single-Celled Organisms Learn?
Scientists come to the Marine Biological Laboratory (MBL) for plenty of reasons, from the cutting-edge facilities and technology to a chance to collaborate with their peers. Daniel Cortes came to the MBL for its ponds.
The pools of water around Woods Hole are home to a genus of tiny organisms called Stentor. They are unicellular, trumpet-shaped protists, and at up to two millimeters long they are huge among single-celled organisms. They are characterized by the lines of flapping cilia running down their sides and the rings of cilia at the “opening” of their trumpet that they use to feed.
Cortes, a Grass Fellow at the MBL this summer, is collecting and culturing a whole variety of these strange organisms to study their behavior. In doing so, he is helping resurrect a 120-year-old debate between two of the most famous biologists of the early 20th century.
In the first decade of the 1900s, the MBL was home to a debate between Johns Hopkins professor Herbert Jennings and University of Chicago professor Jacques Loeb over whether simple organisms could express behaviors. Loeb believed that organisms like bacteria respond in consistent ways to a stimulus — a chemical, for instance, might always attract them, while a light might always make them swim away — and that these responses could be neatly categorized and predicted. Jennings, meanwhile, argued that single-celled organisms are far less predictable and believed that they engaged in rudimentary decision-making. In essence, Jennings gave single-celled organisms agency, and Loeb did not.
This debate raged across treatises, journals, and public talks. Jennings would publish examples of simple organisms that he believed were making decisions, and Loeb would try to dismantle his studies and expose their supposedly flawed results.
By 1910, the debate had wound down. Most scientists agreed Jennings's argument was more convincing; single-celled organisms had behavior that was simply too complex to be rigidly categorized. But, as historian Philip Pauly wrote, it was a “Pyrrhic victory” that “led American biologists to lose interest in invertebrate behavior.”
Recently, however, there has been renewed interest in one of Jennings’s more progressive beliefs: that single-celled organisms are capable of changing their behavior. This question fascinates Cortes, a professor of biological sciences at Virginia Tech, who is using Stentor to examine how unicellular organisms are capable of making choices. Ultimately, Cortes is trying to answer a more radical question: If Stentor have the ability to make decisions, do they have the ability to learn?
How a cell makes decisions
In 1902, in the midst of his feud with Loeb, Jennings performed an experiment on Stentor roeseli, a common species that anchors itself to a substrate (such as plant or debris) with its holdfast to survive. When Jennings exposed S. roeseli to a noxious chemical, it detached and swam away. But, since finding a new place to attach is risky and burns a lot of energy, it didn’t do so right away. Instead, Jennings noted this Stentor seemed to avoid swimming away if it could. First, it bent away from the source of the chemical. If it still detected the noxious substance, it reversed its ciliary beating to push the chemical away. If that didn’t work either, it would contract away from the chemical. Only then, if none of these techniques worked, would it detach and find a new place to grow.
To Jennings, this finding was clear evidence of something called hierarchical decision making. The Stentor employed a series of options, and only if one didn’t succeed would it try the next. The finding was controversial and the scientific community broadly dismissed it — It wasn’t until 2019 that another researcher managed to replicate Jennings’s results in S. roeseli. Even then, according to Cortes, it remains unclear if this study has broader implications, as it may not be indicative of wider Stentor behavior.
Cortes is testing how widespread this behavior is. He is taking Jennings’s experiment and applying it to a range of Stentor, from the bluish S. coeruleus to the bright green S. pyriformis. For each species, he is testing what happens when it is exposed to a noxious stimulus, in this case a magnetic bead. So far, his results seem to suggest that decision-making like this is common.
“All Stentors don’t do all four behaviors,” he clarified, “but generally, there seems to be a hierarchical decision tree.”
He is also investigating a different strange phenomenon in S. pyriformis, the species of Stentor found in Falmouth’s ponds. When Cortes took the Physiology course at MBL in 2018, he worked under the guidance of Stentor biologist Wallace Marshall. In the course, it was discovered that S. pyriformis moves towards light sources and has “the potential for a basic circadian rhythm.” It also has upwards of several hundred algae living within it. Physiology students discovered that this Stentor responded more actively to light at specific times of the day, with the behavior mostly disappearing by late evening. It was also discovered that if you remove the algae from this Stentor’s body, their response to sunlight “pretty much vanishes.”
Cortes is teaming up with Grass Fellow Kei Jokura to investigate this phenomenon. He believes the algae send signals to Stentor based on the intensity of light they are experiencing, so when this signal is asymmetrical, Stentor is drawn towards a light source. They plan to study exactly how the cilia of S. pyriformis generates movement towards light by exposing the Stentor to directional light and watching how their ciliary flows change. Hopefully, this research will unveil the mechanics behind another fascinating aspect of Stentor biology.
Can a cell learn?
After the Jennings-Loeb debate died out around 1910, the subject of unicellular intelligence went through a relatively dormant period thanks to Jennings’s “pyrrhic victory.” Then, in 1952, a researcher at the University of Chicago reignited the conversation. She published a paper asserting something even more radical than Jennings ever found: that the single-celled Paramecium can be trained. Her name was Beatrice Gelber. She spent the next eight years trying — and failing — to convince the scientific world of her findings.
Gelber’s experiment was simple. She took a platinum wire, coated it with bacterial food, and stuck it in a culture of Paramecium. The Paramecium, upon sensing the new food source, traveled toward the wire to feed. Gelber then repeated this test, exposing the Paramecium to the food-covered wire over and over. Then, she wiped the wire clean, sterilized it, and stuck it back in. Incredibly, the Paramecium still gravitated towards the wire. It seemed they had learned to associate it with a meal, like a dog trained to do a trick with the promise of a treat.
These findings were controversial and critics abounded. Yale psychologist Donald Jensen emerged as a major detractor who, as Cortes said, “made it his mission to continually discredit anything that [Gelber] suggested.” Slowly, through battles in journals like Science, Jensen wore down any popular support for Gelber’s findings. By the mid-1960s, she stopped publishing research altogether and was forgotten by scientific history for decades.
Now, Cortes is part of a group of scientists trying to revive Gelber’s work. He is using newer, better technology to test whether unicellular organisms can learn. “The evidence is contentious at best right now,” he said, “but it's definitely worth looking at again.”
So, Cortes is performing a “crazy, Woods Hole experiment,” he said. In his own version of Gelber’s experiment, he aims to see if Stentor can really learn.
Cortes plans to zap a group of Stentor with electricity to make them contract while simultaneously flashing them with a non-stimulating light. He'll do this over and over in order to, potentially, make Stentor associate the two stimuli. Then, he'll expose them to the light without the electrical stimulus to see if they still contract. If they do, it would strongly suggest that these single cells are capable of associative learning.
Any experiment demonstrating associative learning, however, needs to prove that no other confounding variable could be driving the behavior. “You need to be able to show that the environmental factors are stable,” Cortes said. This is especially challenging for single-celled organisms since “any stimulus you give them is going to affect their environment in a measurable way,” he said. The electrical shocks in the experiment can change the acidity of water, he said, which may impact Stentor behavior as well. To account for this, he is figuring out a way to slowly flow the exact right amount of freshwater into the Stentor’s environment to maintain more stable experimental conditions.
When asked if he thought his Stentor would ultimately demonstrate associative learning, Cortes was skeptical. “It’s probably not going to work,” he said. “But if it does, It'll be super cool.”
Cortes credits the MBL with providing an environment where he could perform such an audacious experiment. It was also his inspiration for these experiments — his introduction to Stentor came from 2018 Physiology course faculty member Wallace Marshall. As someone with a background in cell division, entering the world of Stentor cognition was a challenge. He found that the MBL and the Grass Fellowship made that challenge easier to overcome. Meeting and working with peers who had experience with Stentor and behavioral research, he said, allowed him to branch into a new field.