#SfN14: Can we bring back repressed childhood memories?

Image credits: http://www.creativespaceorganizing.com/

Note: this post was originally published Nov 16, 2014, but was subsequently taken down due to the request by the study lead author because it contained unpublished results. The study is now up at Nature Neuroscience, and I'm happy to finally be able to repost.

Poster: Travaglia A, Bisaz R, Alberini C. New York University. Memory traces and their underlying mechanisms during the infantile amnesia period.

If you're like me, you probably don't remember anything from your life before the age of three. This phenomenon, first dubbed by Sigmund Freud as "infantile amenesia", occurs in many different species, yet how it happens in the brain remains a mystery. Freud (being Freud), was the first to put forward a hypothesis: that early memories are actively repressed because of their highly charge psychosexual context. Later thinkers proposed that infantile amnesia was due to a lack of language –hence an inability to encode the world – or a missing sense of self in the very young. Yet these "black box" theories can't explain why young animals also exhibit infantile amnesia. They also can't tell us how it works.

Considering that psychologists have known about the phenomenon for over 100 years, it's pretty remarkable that the first neurobiological mechanism behind it was proposed only a few months ago. Scientists at the Hospital for Sick Children in Toronto, Canada, pointed to a surprising source of toddler forgetting: a rapid stream of newly born neurons in the hippocampus, a brain area important for encoding the memory of life events. The idea is this: memories are stored amongst the connections formed between networks of neurons; when newly born neurons try to integrate into existing networks, they disrupt the information stored within and thus erase the memory trace. As we age, the brain pumps out new neurons at a much slower speed, which allows steady storage of memories formed later in life.

Taking a step back from the specifics, however, all of the above hypotheses suggest that our toddler memories are erased. But what if they're simply repressed, waiting for the right cue to become active again? In a poster presented Saturday at the Society for Neuroscience annual conference, researchers from New York University offered the first tantalizing evidence that this might actually be the case.

The researchers began their experiments using two groups of rats: one 17 days after birth, the other 24 days old. They picked the ages with care: 17 day old rats are roughly equivalent to 4-year old human toddlers, who generally still display infantile amnesia. On the other hand, the 24er rats somewhat represent 7-8 year olds – as you can see from the graph below, they usually have the ability to form steady memories that can last into adulthood.

Source: Learn. Mem. 2012. 19: 423-433

To induce a new memory trace, the researchers put the rats into a distinctive box: one side was brightly lit, the other comfortingly dark. Rats are nocturnal animals, and given the chance would happily dart to the dark side with little hesitation. Once they reached the comfort zone, however, the researchers gave their paws a little zap. This type of learning, called inhibitory avoidance, forms extremely strong memories that can last a lifetime in adults. When researchers dropped the rats back into the lit side immediately after the first zap and gave them a second chance to escape into the dark, both groups of rats hesitated. This suggests that both age groups have learned that darkness isn't safe.

Researchers then waited 30min, a day or a week after the initial zap before re-testing the rats for the fear memory. This time, only the older group displayed any hesitancy; the toddler mice bounced into the dark side as if they'd never been shocked. Just 30min after the initial trauma, they seemed to have forgotten all about it.

If the culprit behind forgetting is adult neurogenesis, it would take the rats much longer – roughly a week – to erase the fear memory. This is because new neurons need to mature before integrating into existing neural networks, a process that takes time. That the 17 day olds forgot about the zap after only 30min after learning suggests that something else is at play; perhaps something that represses, rather than erases the fear memory trace for good.

If repression is the answer, then perhaps the right triggers may reactivate the memory. To test this idea, researchers trained toddler mice (17d old) on the same task, waited a week, and then placed them back into the original box (trigger 1) to test for fear. Again, no sign of remembrance. A day later, researchers gave the rats a new zap (trigger 2) in a completely different environment. They then waited either a day or a week, and introduced the rats back into the original box. Something changed: the rats hesitated on the bright side of the box refusing to enter the dark – a sign that they had recalled the unpleasant early life experience. These results suggests that the original fear memory trace is still encoded somewhere in the brain. However, without the right cues it was strongly repressed.

Further testing showed that the rats need to be exposed to both triggers (the original box and the shock) individually; omit either one and they showed no re-remembering of fear. The timing of the new shock was also crucial: when given 3 days after the original training (2 days after re-exposure to the box), the memory remained repressed; when delivered 5 days or 7 days later, the memory sprung back. Re-activating the memory trace depended on the hippocampus, as amping up inhibitory activity in that brain area also inhibited the memory from returning.

What's the heck is going on? To tease out some potential mechanisms, researchers decided to look at the neuronal activity in the brain after training. If neurons in toddler mice show deficient activity, maybe that could explain why their fear memory traces are initially repressed.

One marker for neuronal activation is the expression of immediate early genes. The protein products of these genes – c-fos, arc, zif– increase after neurons fire and rapidly decline thereafter, so generally they're pretty good surrogate markers of recent activity. This is in fact what the researchers saw in the 24 day olds: a rapid rise in gene expression immediately after training that returned to baseline within hours.

Results from the toddler mice, however, left researchers completely baffled. Instead of the transient increase in gene expression, the upregulation lasted at least 48 hours after the initial training (and thus can no longer be called "immediate early genes"). This suggests that training did trigger neuron activation; yet when tested within this time window for the fear memory, the rats showed no recollection of the shock.

So the mystery remains. The scientist are unsure what to make of all this. It seems that early childhood memories, when given all of the right cues (the context and the shock) can float back into consciousness, and revert into a retrievable state . This suggests that at least some toddler memories are repressed in the adult brain rather than forgotten. But why is there a crucial time window before the cues are able to successfully re-activate the memory? Does memory transfer from the hippocampus to the cortex during consolidation have anything to do with it? How is the memory being repressed in the brain? And what to make of the strange neuron activation results?

"Hopefully I'll have the answers by next year", the researcher said. The implications of this research are actually quite profound, he continued. Early life experiences – perhaps even the ones we can't remember – heavily shape our adult personalities, "for example, sometimes you're deathly afraid of something but don't know why". If the results hold true in humans as well, he said, perhaps early life memories can be retrieved and analyzed for therapeutic purposes.

Update: Here is a more general article about the study over at my new home on the web, Singularity Hub.