The memory eaters

Dr. Andy Lattal is an incredible behavior analyst in his own right, but has also written several short pieces on the history of behavior analysis for the Cambridge Center. All of them are worth reading, but one has an incredible hook: the Unabomber once targeted, but failed to kill, a “behaviorist.”

It’s an incredible story, but it almost certainly boils down to some type of mistaken identity. Kaczynski wanted to destroy technology, and to destroy people who he viewed as manipulators of human behavior. Kaczynski was undoubtedly a mathematical prodigy and genius, and his manifesto (published in newspapers after one of his threats) has gained a cult popularity in recent years as people increasingly see that not every technology benefits humanity.

And yet…

The manifesto is pretty bad. It has the crank energy of fascism, wherein Kaczynski defines his enemy as “leftists” who are mysteriously indefinable, but also powerful, but also weak. Kaczynski also missed the mark when he tried to bomb James V. McConnell, whose research mainly involved the behavior of planaria, a species of flatworm. Later in life McConnell “oversold” behavior analysis in the popular press, which is likely what attracted Kaczynski’s ire.

McConnell was an interesting figure. His research was somewhat impactful, but controversial. He published a semi-satirical journal, The Worm Runner’s Digest, in which B.F. Skinner published (twice!) pseudonymously. His journal was somewhat impactful, but controversial. 

A brief aside in the Lattal article illustrates why:

During the years of the Digest McConnell published several articles not only claiming to have shown classical conditioning in planaria, but also gains in acquisition of a conditioned response when a naïve planarian was fed another planarian that had been previously conditioned. McConnell’s various claims proved to be controversial, resulting in a series of back and forth experiments attempting to sort out fact from fiction about planarian conditioning.  [Emphasis added]

So…worms that gained the memories of other worms, from eating them?

---

Let’s begin with a visit from our good friend, George Razran. A Russian emigré and behaviorist, he reported in the 1930s on the overwhelmingly Pavlovian-influenced Russian state laboratories. In 1933, he writes, in Conditioned responses in animals other than dogs, that colonies of single-celled organisms (Carchesii lachmani) acquired a conditioned response of contraction when a neutral stimulus of colored light was paired with an unconditioned stimulus of touch with a glass filament. 

To all of us, nearly 100 years later, this is expected; at the time the study was conducted, this was a relatively new finding. Even two decades later, esteemed researchers argued that learning was a function of synapses, which of course are not present in unicellular organisms. 

But what might surprise you, dear reader, is that when portions of the experimental colonies were transplanted into novel colonies, the mixed colonies displayed the conditioned response in 1-3 trials – while the starting colonies acquired the response in an average of 138.5 trials. Razran notes that this is:

“[...] a rather peculiar and interesting case of transfer or generalization of C-Rs.”

---

By McConnell’s own account, a colleague named Bob Thompson suggested in 1953 that they attempt conditioning in planaria, because it was the simplest lifeform to possess a synapse. If learning depends on a synapse, planaria ought to be able to learn. By 1955, they had paired a light with a delayed shock, increasing the response (curling up) to light alone from approximately 25% percent of opportunities to 50%. They also demonstrated habituation (or respondent extinction) by presenting the light only, which eventually elicited fewer responses.

Thompson and McConnell parted ways, with Thompson becoming a professor at Louisiana State University and working with rats. McConnell became an adjunct at Michigan, and continued working with planaria.

A unique feature of planaria is that, when bisected in either direction, both sections can regenerate into two complete organisms. By 1958 McConnell trained his planaria over a longer period, and eventually they responded to the light preceding the shock in 92% of opportunities. He then bisected them horizontally, with a head and tail segment each regenerating into a separate organism.

Which section contained the “memories” of the light and shock? As it turns out, perhaps both. Each regenerated organism learned the pairing more quickly during a second period of training.

The head of the planaria contains the brain. The tail regenerated a new brain, but contained “memories” of prior training. How could this be?

---

Two hypotheses of memory were in the zeitgeist: the previously mentioned synapse hypothesis, and the RNA hypothesis. The synapse change hypothesis came out of the logic of a scientific understanding of learning being localized in the brains of higher organisms; brain damage induced in lab animals caused poor performance on memory-based tasks, therefore memory was localized in the brain. The RNA hypothesis came from a scientist named Holger Hydén, who discovered that animals display changes in the nucleus of their RNA after a learning task.

Indeed, other scientists beat McConnell to the punch: they demonstrated that RNA injections, from test rats into novel rats, could induce learning. McConnell was still trying with planaria, which resulted in a bizarre scenario: extracting RNA from hundreds of planaria, dropping the test tube, and scraping planaria RNA off the floor – only to end up using it anyway.

But the needles were too big to inject the planaria reliably, with one author comparing it to “stabbing a prune with a javelin.” After a tip from a fellow researcher, they decided to feed them to each other.

According to McConnell’s results, this too transferred learning.

---

Could it be that RNA is the memory molecule – that it “contains” memories? McConnell, never failing to take advantage of an opportunity to exaggerate, wrote:

You may have noticed, dear reader, that this state of affairs has not come to pass.

---

The big question remains: could you feed worms to each other to “transfer memories”?

Of course not! Or…maybe?

(One aside: a research group tried to replicate McConnell’s “cannibal” study and found null results. McConnell argued that the replication was bound to fail because they ground up the planaria, which may have destroyed the RNA. His planaria were simply chopped up into pieces. This may or may not read like a convincing argument to you.)

After initially facing a less-than-enthusiastic reception – here's a study from 1963 examining how experimenters could have exaggerated learning in planaria – and eventual “discrediting” of the RNA hypothesis, there has been a re-evaluation of McConnell’s work. His initial study – training the planaria, cutting them in half, and re-training them – has been replicated more rigorously. 

And the RNA hypothesis, if you squint a little, sounds an awful lot like epigenetics. Briefly, if genetics determines some aspects of development on a molar level (e.g., in a species, over time), epigenetics determines some aspects of development on a molecular level (e.g., within a lifetime). For example, certain pharmaceutical drugs can alter gene expression.

Could it be that “memories” are not exactly what was stored in RNA, but rather an epigenetic experience critical to survival?

In searching for whether “memory transfer” has been attributed to epigenetics, we came across this article, Epigenetics and memory: causes, consequences and treatments for post-traumatic stress disorder and addiction, co-written by Andy Lattal.

---

McConnell hypothesized that we would reach a stage where we could transfer memories. It seems unlikely. But if you want to try, chop – don’t grind.