Don't remember your baby days? Blame new neurons!

This post was originally published with slight edits at Knowing Neurons.

Think back to when you were two years old. Think HARD. Anything?

If you’re like most people, you’re probably drawing a blank. Across cultures, adults can’t seem to recall any events that occurred in the first 2-3 years of their lives; memories between 3-7 years of age are often filled with gaping gaps and a healthy mix of reality and fantasy.

This phenomenon, first dubbed by Sigmund Freud  in 1900 as “infantile amnesia,” has left scientists scratching their heads for decades.  An early hypothesis thought this is because young children don’t have a sense of self and thus can’t form autobiographical memories. However, go eavesdrop on a toddler’s “this was my day” crib-mumblings and you’ll quickly think otherwise. In a landmark experiment, Dr. Byron Campbell found that precocious guinea pigs –born with relatively mature brains– could retain a fear memory learned in their infancy for over 2 weeks. If you don’t find this impressive, look at similarly aged young rats: when trained and tested with the same protocol, they seemingly couldn’t remember a thing. Since rats go through extensive brain maturation after birth, this suggests that ongoing brain development in infancy is to blame for infantile amnesia.

But what’s the specific culprit? Accumulating evidence is pointing to a process called adult neurogenesis (birth-of-neurons).

Two areas of our brain continually generate new neurons after birth, rapidly at first then gradually declining as we age. One area with abundant neurogenesis is the dentate gyrus in the hippocampus, a brain region important for learning and memory. These youngster neurons are high excitable and eager to encode memories – and they indeed do. In adult rodents, artificially increasing neurogenesis before learning helps the brain retain and differentiate between two very similar experiences; in other words, it helps them remember better.

But more computational power is not always a good thing.  When new neurons integrate into an existing circuit, they may profoundly “shake up” the wiring and excitability of hippocampal networks, potentially disrupting old memories stored previously in synaptic connections. Is this what leads to infantile amnesia?

Researchers from the University of Toronto explored this idea with an elegant set of experiments.  First, they trained a group of adult mice to find a hidden platform in a giant tube of murky water. Once the mice had learned the task, researchers jacked up their neurogenesis by 2-3 fold, essentially turning the rate of neurogenesis “infantile.”  Many methods exist, ranging from au natural voluntary running (used here) to highly technical chemical-genetic manipulations that prevent new neurons from spontaneously dying*. When tested later, mice with increased neurogenesis performed way worse than their non-manipulated peers. Thus, endorsing adults with a “baby”-like rate of neurogenesis deteriorated an otherwise stable memory. (*For example, Sahay and coworkers targeted Bax, a death-promoting factor required for apoptosis. In one type of transgenic mice, the Bax gene in their neural progenitor cells is flanked by something called LoxP. Think of LoxP like a cutting site. When researchers give the mice a drug called tamoxifen, it activates an enzyme called CreER, which goes and cuts at the LoxP sites like scissors, thus essentially cutting out the Bax gene – stuck in the middle of two LoxPs – from the genome. Without Bax promoting spontaneous cell death, more young neurons survive.)

Researchers then tackled the hypothesis from the other end: they slowed down neurogenesis in baby mice, thus somewhat mimicking an adult state. Here’s how they did it: there is a type of mice that have a protein called thymidine kinase expressed exclusively in their neural progenitor cells (cells that make new neurons). Tymidine kinase has a useful quirk: when it binds to a drug called ganciclovir, it kills off the cell. Thus, researchers gave ganciclovir to young mice and selectively killed many of their newborn neurons. In a sense, this “slows down” their rate of adult neurogenesis. Researchers then trained these mice, along with normal baby controls, to associate a peculiar-looking box with an electrical shock.

28 days later, after sufficient time for new neurons in the control mice to “infiltrate” into existing networks, researchers put the mice back into the box. Normal baby mice skittered around undisturbed; they had forgotten that the box is a dangerous place! However, baby mice that were manipulated to have less new neurons froze in fear. Hence a lower rate of neurogenesis in infancy confers immunity to infantile amnesia, at least in baby mice. Of course, it’s possible that other unknown mechanisms are also at play.

These are amongst the first data to support adult neurogenesis as a factor driving babies to forget. Many questions remain. Does the magnitude of increased neurogenesis determine how much and how fast a baby forgets? How are existing synapses selected for elimination (a process called “pruning”)? Some scientists propose that memories initially dependent on the hippocampus eventually “transfer” to the cortex for permanent storage – are these memories spared from forgetting? And, on the grand scheme of things, are we unfairly robbed of our earliest memories? Or is adult neurogenesis doing us a favour by decluttering the baby brain, thus allowing new and more important memories to be stored?


Josselyn SA, & Frankland PW (2012). Infantile amnesia: a neurogenic hypothesis. Learning & memory (Cold Spring Harbor, N.Y.), 19 (9), 423-33 PMID: 22904373