Memory in the fleshA radical 1950s scientist suggested memories could survive outside the brain — and he may have been right
The bomb was the size of a shoebox. The brown paper packaging that enveloped it was standard for US mail at the time, and its contents provided no indication — no ticking clock, no lit fuse, no clicking sound — of the explosion to come. Resting on the kitchen counter that separated Dr. James V. McConnell from his graduate student assistant, Nicklaus Suino, it awakened no suspicions. Even the bomb’s trigger was subtle: all it took was for Suino to reach across the counter and touch the clear fiber tape that held it together.
Despite their proximity to the explosion, both McConnell and Suino survived. The young graduate student suffered minor injuries and was kept at the local hospital overnight; McConnell experienced mild hearing loss. "It was quite traumatic at the time," Suino recalls. "Dr. McConnell and I both developed shock reactions to loud noises afterward."
McConnell was the target of Theodore "the Unabomber" Kaczynski’s 10th bomb. Why would Kaczynski, a man obsessed by a quest to rid the world of threats from modern technology and science, go after a psychology professor at a Michigan university?
Some thought it had to do with a prediction McConnell had made in the ‘70s, when he suggested that prisoners would one day be conditioned out of anti-social behavior. But McConnell was also famous for something else: in the ‘50s and ‘60s, McConnell had performed a series of memory experiments at the University of Michigan while Kaczynski was a student there — experiments that earned him a reputation as a kooky, arrogant, and deeply misguided academic who challenged everything we think we know about memory. McConnell’s radical research suggested that memories could exist outside the brain — and even be transferred between organisms. The conclusions were so outlandish and dystopian that some speculated they attracted Kaczynski’s ire.
But James McConnell may have been right.
McConnell’s experiments have largely faded from scientific memory — those that remember him tend to use him as a cautionary tale. But at the end of the ‘50s, McConnell was a big deal: his charismatic personality, combined with his controversial scientific accomplishments, landed him on TV more than once, notably on The Steve Allen Show. Instead of dazzling audiences with complicated science, McConnell captivated them with awe-inspiring concepts. In fact, the experiment that brought him his first taste of national fame was remarkably simple.
In a 1959 study, McConnell sought to demonstrate that memories could be stored in cells outside the brain. To do so, he enlisted the help of the common freshwater flatworm, Dugesia dorotocephala.
Flatworms were ideal because, like mammals and unlike, say, jellyfish, they have a centralized brain. They also have an exceptional ability to regenerate themselves from tiny morsels of flesh: sever the tip of a flatworm’s tail, and within 14 days you’ll have an entirely new specimen, fully equipped with a brand new brain. Moreover, flatworms can be taught to perform certain behaviors. Researchers can use electrical shocks to teach them to respond to lighting cues by moving to a particular part of a petri dish. That is, flatworms can be trained to remember a behavior and perform it on cue.
If McConnell demonstrated that the lowly flatworm could recall its training after its head had been cut off, and its brain had grown anew, the experiment would show that memories could live outside the brain.
"We didn’t know what the tail would remember, if anything, for it had to grow an entirely new brain, new eyes, and almost an entire new nervous system," McConnell wrote in a 1965 collection of science-related articles entitled The Worm Re-Turns. But, according to McConnell’s study, they did remember. The regenerated worms were able to recall the behaviors they had learned to execute under certain lighting conditions. "The tail regenerates," McConnell wrote, and "showed as good a memory of the original task as did the heads."
Published in the Journal of Comparative and Physiological Psychology, a highly regarded psychology journal, the study earned McConnell a lot of press, with mentions in Time, Medical World News, Newsweek, and Fortune. "Almost at once the mail began to come in asking for details," McConnell wrote.
"We didn’t know what the tail would remember, if anything"
But the fantastic takeaway of McConnell’s study — that cells other than neurons could store information — caused many to question the study’s methodology and conclusions. In 1960, he published a second experiment, one that pushed his theories far beyond what anyone would have imagined. This one left the scientific community completely stumped.
Widely known as "the cannibalism experiment," the study tested another McConnell theory: that memory could be transferred chemically from one flatworm to another through something called "memory-RNA." Memory-RNA, McConnell suggested, was a special form of RNA — the intermediary form of genetic information that fills the gap between DNA and proteins — that could store long-term memories outside the brain. His method was unorthodox, to say the least: McConnell fed bits of trained flatworms to their untrained brethren. As a result, McConnell claimed, the untrained flatworms performed behaviors that the trained flatworms had previously learned. In short, the dead flatworms’ memories had found a new home.
"Biologists and chemists said ‘no way,’" recalls Reeva Kimble, who did undergraduate research for McConnell in 1959 and 1960. Reeva later married Daniel Kimble, the student who gathered data for McConnell’s first regeneration experiment. McConnell’s opponents couldn’t make sense of his findings. For them, "there was no mechanism to understand his result, so it had to be hogwash," Reeva says.
McConnell’s cannibalism experiment was greeted with intense skepticism. Groups of researchers at competing universities did their best to replicate his study in order to disprove his work; if the study couldn’t be replicated, it would be enough to expunge his ideas from scientific memory.
Some researchers reported obtaining similar results when recreating McConnell’s study, but many didn’t. Others (rightly) faulted him for his small sample size and because the effect he reported was, albeit significant, relatively weak. Still, replicating work in which a human observes animal behavior precisely and manually is extremely difficult, and McConnell’s conclusion was never fully debunked.
For his part, McConnell insisted that other scientists failed to reproduce his findings because no one was able to fully recreate the conditions of his experiment. Eventually, however, the psychologist’s work was cast aside, or referred to as "a failure."
McConnell’s hobbies didn’t do much to strengthen his scientific credibility. He liked writing, science and humor, so he combined them in The Worm Runner’s Digest — a "scientific journal" that he used to reach the masses following the publication of his 1959 flatworm study. The journal contained real studies, but they were featured alongside poems, cartoons, and humorous takes on scientific research penned by colleagues and students. As a result, outsiders complained that they couldn’t tell the difference between the scientific subject matter and its more humorous content.
"We’ve insisted that the Digest be a mixture of fact and fun, for it seems to me that anyone who takes himself, or his work, too seriously is in a perilous state of mental health," McConnell wrote in 1965, in a compilation of the Digest’s articles. This attitude made him popular among his students, but failed to inspire confidence among his peers.
"He was very frustrated with people not taking him seriously," Suino remembered during a conversation in 2013. McConnell was disappointed with the dismantling of his memory work, he said. "I think that he felt that a lot of people who went into his research with a bias against it."
Why were McConnell’s ideas so unsettling? What he was proposing — the existence of a chemical memory trace that encodes training information in flatworms — is nothing short of astounding. Supposing that McConnell was correct, it would mean that the flatworm brain is capable of storing information in chemical structures which can be transferred to other parts of the body. Moreover, these chemical structures could form a "language" for memories that can cater to an infinite number of situations — and that other organisms can "read."
The chemical transfer theory goes against current and past conceptions of memory. Aristotle, for example, likened memory to a wax tablet in the mind upon which a person imprints knowledge and recollection as time goes by. The impact of that explanation has been profound and long-lasting; many point to it as the source of the "young, impressionable mind" trope.
Science has since moved on from the wax tablet analogy, of course. Currently, biologists believe that information is stored in neural networks in the brain, in the connections that allow information to be transmitted from one neuron to the next. That’s as far as most researchers are willing to go when discussing the underpinnings of memory, however. Press a scientist to tell you how memories are encoded and decoded in the brain, and you’ll soon find that the scientific community doesn’t have an answer. How do our brains represent memories like seeing X-rays for the first time? The question is so difficult to tackle that many researchers have chosen to focus on determining what sorts of modifications take place in the brain when memory is stored instead — changes in neuron structure, for example — in the hopes that they might reverse-engineer memory formation.
But looking at modifications in the brain isn’t the same as figuring out how memory is encoded, or where it's stored. Whoever finds the vault where we keep our memories is bound to go down in history.
Michael Levin doesn’t recall the first time he encountered McConnell’s 1959 memory study, but he remembers his reaction to it. "I came across a primary paper of McConnell showing that memory can survive regeneration, and I thought, ‘Well this is extremely profound because it really speaks to the ability of all tissues — those outside the head anyway — to store memories.’"
Levin, a developmental biologist at Tufts University, is well known for his work on limb regeneration; his studies have been featured in the Journal of Neuroscience and on the cover of Cell. One of his most well-known experiments demonstrated that it’s possible to trigger limb growth in young frogs by giving them a drug cocktail. He’s also recognized for figuring out how to grow a working tadpole eye in a location other than the head — namely on the tail and back. "He’s really well-known in the field of amphibian regeneration," says Alejandro Sanchez Alvarado, a developmental biologist at Stowers Institute for Medical Research. "I don’t know of anybody doing work remotely similar to what Mike is doing."
But a few years ago, Levin took on a side project. After stumbling upon McConnell’s work, he decided to take the basic principles behind the psychologist’s first memory experiment and try it out for himself. After all, there were already vast amounts of literature on aneural organisms — organisms that don’t have a brain to begin with — that can learn, so he thought McConnell might have been at least partially right. "Everything from plants, to slime molds, to single-cell organisms can learn, even sperm can learn to run mazes," Levin says. But unlike those creatures, flatworms possess a "true brain" — a centralized nervous system located in their triangle-shaped heads. This makes them a lot more like humans than most of us are willing to admit.
It took four years and over a million dollars to recreate and improve upon McConnell’s experiment. In an effort to sidestep the criticism his predecessor faced, Levin developed a rigorously documented, replicable experimentation method. Part of that was done by the invention of a machine, called the "Automatic Training Apparatus," that could train and track the movements of multiple flatworms without human intervention.
Late last year, I traveled to Medford, Massachusetts — just outside Boston — to visit Levin in his lab at Tufts University. Simpsons figurines, mugs, and blankets dot his office; a three-eyed plastic fish watches over him from the bookshelf. Word got around one day that the researcher liked the show. Now the toys just keep coming in; the students think it’s funny.
Eventually, Levin took me to a small room a few doors from his office to see the "Automatic Training Apparatus." There, I watched Moran Neuhof, a visiting student from Israel, put 12 petri dishes — each filled with a flatworm — into Levin’s machine. Pointing to a computer screen, Neuhof told me to observe an image of a worm as it moved inside a petri dish. Red targets tracked the worms as they swam around; an invisible grid helped translate their movement into time-stamped coordinates along the way. This is how Levin was able to objectively train the worms and record their behaviors, without ever having to be in the same room.
Designing the Automatic Training Apparatus "was a nightmare," Levin says. Flatworms are tiny and quick, and they love to hang off the sides of a petri dish instead of swimming at the bottom — all of which makes them tough to track with a camera. The researcher collaborated with several engineering firms to make an apparatus capable of simultaneously, and autonomously, tracking and recording the movements of 12 worms. Minimizing human participation was critical — a fully automated process was meant to protect Levin’s results from the scrutiny McConnell faced. "If you don’t believe me, you can take the tracking data and the Quicktime movies," Levin says. "Analyze it yourself, and see what you think of whether they behaved correctly or not."
The training protocol Levin devised is simple, but elegant:
Flatworms are cautious and like to explore their environment, so when they’re introduced to food in a new location, they tend to circle it for a long time before feeding, Levin says. This means that scientists can use "latency of feeding" — the time the worms take to approach a food item — as an indication of how well they remember a specific environment. If they remember that the environment is secure, or that it only contains one source of food, then they don’t need to explore it; they can go straight for the food. So, Levin trained them to recognise the textured petri dishes. Once the worms learned to recognize the rough floors of the dish, they started going for the food more quickly. Then, Levin and his team cut off their heads.
Two weeks later, the researchers re-introduced the worms, now armed with new heads, into the dish, but only for a short period of of time. This shortened re-introduction is meant to signal to the brain that the memory stored in the flatworm’s trunk is still relevant, Levin says. The next day, the worms were finally re-tested. "Worms that had never before associated the roughness with the liver, took a long time to approach the food, and worms that had been trained went straight for the liver — that’s the difference," Levin says.
The Tufts University researcher has yet to establish a mechanism to explain his findings, which he published in the Journal of Experimental Biology in 2013. He hypothesizes that memories could spread beyond the brain thanks to electrical charges generated by cells in the rest of the body. But until more research is conducted, it’s hard to know what to make of his results.
Other mechanisms have been suggested. Eva Jablonka, a developmental biologist who studies the evolution of nervous systems at Tel Aviv University, thinks that small RNAs might be involved. Small RNAs are short copies of DNA that aren’t translated into proteins. When the flatworm learns a behavior, the brain chemistry changes, and it’s possible these changes alter small RNAs, Jablonka says. Because these molecules can migrate between cells, altered small RNAs could end up in stems cells that remain in the decapitated worm. When the worm’s head grows back, the small RNAs migrate back to the head, changing the brain’s chemistry and allowing it to learn certain behaviors more quickly. If true, the memory that Levin thinks is stored outside the brain wouldn’t be memory at all. Rather, the small RNAs would allow the flatworms to recover a brain "environment" that helps them learn a specific behavior more quickly. But this scenario, Jablonka says, "is still imaginary."
It’s still possible that, like McConnell’s work, Levin’s study is flawed in some unidentified way
At this point, any attempt to explain Levin’s results is problematic. The truth is that we don’t know if Levin’s conclusions are even remotely correct. It’s still possible that, like McConnell’s work, Levin’s study is flawed in some unidentified way. It’s also possible that the flatworm’s unique regenerative abilities lie behind its ability to recall memories after growing a new brain. In that case, flatworms may be the only species with a central nervous system that can store memories outside of the brain. "I find that highly unlikely," Levin says, "but it’s a possibility."
Even if storing memory outside of the brain is universal among animals, that storage method might only work for simple pieces of information. Complex memories like the significance of the word "truth" or "caring" might not have a place beyond the brain. But if there is a small chance that the experiment is reproducible, and that this isn’t a trait reserved to some tiny insignificant worm, the impact could be revolutionary.
"I think [this work] will start a renaissance in human memory," says Oné Pagán, a planarian expert and pharmacologist at West Chester University. "Levin is one of the top researchers in developmental biology, so if anybody can pull it off, he can."
Levin’s work could completely alter the way we think of memory, with real-world applications, Pagán says. The new flatworm studies could one day lead to novel Alzheimer’s disease treatments. In neurodegenerative diseases, "there’s a certain destruction of cellular structures that deal with memory," he says. So, "if memory is stored outside the actual brain, that opens the door for people to potentially fully or partially recover those memories" following stem cell therapies, for instance.
Studies of flatworm memory recall could also give rise to new architectures for storing data and biological memories in computers. If information can be stored in many different types of cells, for example, then creating a robotic prosthetic with parts that can perform as computational devices — independently from others — might be within reach. Machines would be more robust, because a glitch in their central processors wouldn’t necessarily lead to a complete loss of function; other "smart" parts could pick up the slack, and maybe even help the robot fix itself.
Finally, Levin’s findings could one day help researchers activate limb regeneration and wound-healing in humans. "Let’s say that you lost a big part of your liver," Jablonka says — "if we were able to introduce the kinds of molecules that will help your body regenerate this liver, that would be a great thing!"
As with McConnell, the mass media has demonstrated intense interest in Levin’s study, with articles appearing in Scientific American, National Geographic, NPR, Wired UK, and here at The Verge, to name a few. "The worms’ memories were just as accurate as those worms who had never lost their heads," explained Carrie Arnold at National Geographic. "So these worms grew new heads with old memories, a remarkable finding," wrote Robert Krulwich at NPR. Levin’s study showed that "it’s clear the action is happening somewhere outside of the brain," exclaimed Wired UK’s Liat Clark.
But the scientific community’s reaction to a revival of McConnell’s work has been mixed.
"I think it’s a fine piece of work," Alvarado says. "The data that [Levin] obtained seems like a reasonable consequence of his manipulation." Jablonka also thinks Levin’s methodology is sound. "He automatized the experiment so they can follow a lot of animals at once," she says. "Michael is a very original scientist." "Levin’s [study] was done in such an exquisite way," Pagán says. "With appropriate controls and no human bias." And Scott Rawls, a pharmacologist who works with flatworms at Temple University in Philadelphia, agrees. "From a mechanistic perspective, I think he covered the bases and did the experiment pretty well."
Others are less enthusiastic. "The paper from Tufts might be right, but they still haven’t quite done the experiment in a way that really convinces you," says Robert Kentridge, a psychologist at Durham University in the UK. What Levin has yet to do is show that the transfer is memory-specific, he says. "The effect of stress could have the same consequences as learning." What Levin might have picked up on is behavior induced by a stress hormone, itself triggered by the texturized petri dishes Levin used. A superior experiment would have eliminated that possibility by training flatworms to also recognize an un-textured petri dish, he says. "That would have shown that what’s being transferred is quite specific and has nothing to do with a dish that might be ‘stressful.’"
Kentridge isn’t alone in thinking that the study could have been better designed. Noelle L’etoile, a cellular biologist and nematode researcher at UC-San Francisco, says that Levin should have studied more variables and other forms of stimuli. The fact that the training protocol relied on a textured petri dish might have contributed to the results in ways that we don’t yet understand, she says. The rough surface could have altered the neuron circuits that allow the worm to move, which would in turn help the worm get around more easily inside a petri dish. Those altered circuits might also become more efficient after training, she ventures.
It’s crucial that other groups of researchers tackle this study, L’etoile says. Levin should give his machine, or the design for his machine, to other groups to see if they can reproduce the findings. If they can show that Levin’s results don’t stem from a series of flukes, it would "go a long way to support his contention that he has found something," she says.
L’etoile readily admits that McConnell’s failures continue to loom large over Levin’s study. It’s a "cautionary tale in the field of experimental biology" — one that comes with "a lot of baggage," she says. The fact that so many research groups had trouble replicating flatworm training protocols in the ‘60s and ‘70s means that the bar to convince her has been set very high, she says. "I think I have a great deal of skepticism about [flatworm] training."
Five years after the Unabomber attack, at the age of 64, James McConnell suffered a heart attack, and passed away. An obituary printed in The New York Times mentioned both his controversial experiments and the Worm Runner’s Digest, which the author said "often needled his colleagues."
Neither Suino nor the Kimbles attended James McConnell’s funeral, but his friends and colleagues remember him still. "He was the person who convinced me to go to graduate school," Reeva Kimble says. Despite the criticism he faced, McConnell’s showmanship and larger-than-life ideas resonated with his students. He was the researcher who showed them that science could be light-hearted and fun. He was also the researcher who didn’t back down when he was challenged for making outlandish claims.
"I think what [McConnell] was right about is that one should pay attention to these unusual things on the fringes of science that can disrupt the current way of thinking," Levin says.
McConnell’s ideas were set aside, seemingly forgotten, for half a century. That’s not unique; scientific ideas are regularly abandoned, in favor of more sensible, simple proposals. But science has demonstrated more than once that what humans perceive as being "simple" isn’t always biologically simple. Our ideas are partly informed by culture — a culture that, in this case, believes firmly in the brain as the center of our physiological and psychological universe. This is undoubtedly correct. But it might not be the whole story.
Like dormant memories, James McConnell’s ideas have resurfaced. What science does with them now will depend on Levin and his peers, as they try to prove that the results are wrong — in a counter-intuitive effort to show that maybe McConnell was right. Should they succeed in poking holes in Levin’s study, McConnell and his flatworms will likely remain a cautionary tale. But if they fail, James McConnell will be remembered, forever, as a pioneer.