Your Baby Remembers More Than You Think — The Science of Infant Memory at 13 Months

You won't remember this moment. The morning light, the weight of this page in your hand or the glow of this screen, the particular quality of the silence or the noise around you right now — in a few weeks, this moment will dissolve into the texture of your life, indistinguishable from a thousand others.

But your baby might remember what happens next. For a little while. And the fact that they can — and that they'll lose it — is one of the most extraordinary stories in all of neuroscience.

The Hippocampus Wakes Up

In March 2025, a team led by Nick Turk-Browne at Yale published a study in Science that changed what we know about infant memory. They did something that had never been done before: they scanned the brains of awake, alert babies using functional MRI while the babies performed a memory task.

Twenty-six infants — half under a year old, half over — were shown photographs of faces, scenes, and objects. Later, they were shown a previously seen image alongside a new one. The researchers tracked where the babies looked (familiar vs. novel) and, crucially, what their brains had been doing during the original viewing.

Greater activity in the posterior hippocampus during the first viewing of an image predicted whether the baby would later recognize it. The hippocampus — the brain structure that adults rely on for episodic memory, for binding the what-where-when of experience into coherent episodes — was doing its job in babies over 12 months old. In those under a year, it was not.

This is the moment your baby's brain begins writing autobiography. Not language, not narrative — but the raw encoding of experience into retrievable memory. The hippocampus has come online. And right now, at 13 months, it is working.

Two Kinds of Remembering

Your baby doesn't remember the way you do. But they remember in two distinct ways that mirror the adult memory system more closely than anyone expected.

Rose, Feldman, and Jankowski (2011) demonstrated that by 12 months, long-term memory has two independent components: familiarity — the sense that something has been encountered before — and recollection — the ability to retrieve specific details about a past event. These map onto the same dual-process model that describes adult memory.

The evidence for recollection is striking. In deferred imitation experiments, 12-month-olds can watch a researcher perform a series of actions with a puppet — remove a mitten, shake it to ring a hidden bell, replace it — and then reproduce those exact actions four weeks later. No practice in between. No reminder. Just one observation, and a month later, they do it.

Beyer and colleagues confirmed in a 2025 study (Infancy, N=200) that deferred imitation is present from as early as six months, improving continuously. But there's a limitation: at 12 months, babies can reproduce the actions only when the test puppet looks the same as the one they originally saw. Change the color or form, and the memory fails. Flexible, cross-cue memory — the ability to recognize the same event across different surface features — doesn't emerge until later in the second year.

Your baby right now lives in this in-between: they can encode and recall, but their memories are tightly bound to the specific context in which they were formed. The puppet must look the same. The room should feel familiar. Memory is real, but it's not yet flexible.

The Nap Is the Lesson

Here is where the science of infant memory connects to something every parent experiences daily: the nap.

Seehagen, Konrad, Herbert, and Schneider published a landmark study in PNAS in 2015 (N=108, ages 6–12 months) that established a causal link between napping and memory. Infants who napped within four hours of learning a new skill — watching a puppet demonstration — retained it at both 4 and 24 hours later. Infants who did not nap showed no retention at all.

The average time from learning to nap: 47 minutes. The average nap duration: 106 minutes. And the effect was not subtle. It was all or nothing. Nap after learning, and the memory survives. Stay awake, and it doesn't.

Friedrich and colleagues extended this in a 2020 study published in Nature Communications (ages 14–17 months) that added a crucial nuance. Sleep doesn't just preserve memories — it determines what kind of memory survives. After sleep, infants retained episodic memories — the specific details of what happened. Without sleep, only semantic memory survived — general knowledge stripped of its context. Frontal fast sleep spindles in the 13–15 Hz range were specifically involved in binding episodic memories during consolidation.

This connects directly to what I wrote about in my post on sleep: the nap isn't downtime. It's processing time. The sleep spindles that consolidate motor learning are the same architecture that now consolidates episodic memory. Every nap after a new experience is the brain's filing system at work — deciding what to keep, how to keep it, and in what form.

Hermesch (2025, Infancy) added one more piece: 12-month-olds who napped after watching a demonstration were better able to generalize — to reproduce the actions with a novel object they'd never seen. Sleep doesn't just preserve the specific memory. It extracts the transferable principle.

The practical implication is as simple as it is powerful: the optimal time for your baby to learn something new is just before they sleep.

The Great Forgetting

So if the hippocampus is encoding memories, and sleep is consolidating them, and your baby can recall actions from a month ago — why won't they remember any of this?

Why can't you remember being one year old?

This is the paradox of infantile amnesia: the brain is clearly forming memories during a period that will later be inaccessible. The encoding works. The retrieval fails. And the most compelling explanation for why involves one of the brain's most remarkable processes: neurogenesis.

The hippocampus is one of only two regions in the brain that continues to produce new neurons after birth. During infancy, this production is extraordinarily rapid. Akers and colleagues demonstrated in a 2014 Science paper that this flood of new neurons is not just coincidental with forgetting — it causes it. New neurons integrating into hippocampal circuits remodel existing connections, degrading the memory traces stored there.

The proof is elegant. When the researchers increased neurogenesis in adult mice after memory formation, the mice forgot. When they decreased neurogenesis in infant mice, the mice retained memories they would normally have lost. In guinea pigs and degus — precocial species whose neurogenesis is mostly complete before birth — infantile amnesia is absent entirely. But induce neurogenesis in these animals, and forgetting appears.

But here's where the story takes a turn that Turk-Browne calls "almost science fiction." Bhatt, Bhatt, and Bhatt showed in a 2018 Current Biology paper that infant fear memories in mice are not fully erased. The memory engrams — the physical neural traces — exist in a dormant, inaccessible state. Using optogenetic stimulation to reactivate the specific neurons that were active during the original encoding, the researchers could reawaken "forgotten" infant memories in adult mice.

The memories were there all along. They just couldn't be reached.

Turk-Browne's lab is now testing whether human children can remember events from infancy — specifically, whether they recognize home videos from their first year when shown them in preschool. Pilot results, he reports, suggest some memories do persist to preschool before fading. "One really cool possibility," he told Yale News, "is that the memories are actually still there in adulthood."

What Survives the Forgetting

Even as episodic memories fade, something endures.

A 2026 review in Child Development Perspectives argues that emotion remains under-integrated into our models of infant learning — and that this is a serious gap. Emotional experiences shape neural encoding from the earliest months. Research on maternal sensitivity shows that responsive caregiving alters infant brain responses to emotional faces by five to seven months: infants of sensitive mothers show increased dorsolateral prefrontal cortex activation when viewing happy faces. Infants who experience maltreatment develop a perceptual bias for angry faces — faster recognition, larger neural responses.

The episodic memory of a specific moment — the afternoon you read to them, the morning you danced in the kitchen — will dissolve. But the emotional architecture that those moments built will not. The attachment system, the stress response calibration, the implicit sense of whether the world is safe and whether people can be trusted — these are encoded in systems that operate below and beyond episodic memory.

Your baby will not remember the moment. But they will carry what the moment meant.

Memory Is Everywhere

One of the most important recent insights about infant memory is how pervasive it is.

Behm, Turk-Browne, and Kibbe published a review in Trends in Cognitive Sciences in 2025 that reframed the field. Using a component process framework, they analyzed common infant cognition tasks — object permanence, categorization, word learning — and found that nearly all of them have hidden demands on episodic-like memory.

When your baby searches for a hidden toy (object permanence), they must encode the specific hiding event — what went where, when. When they categorize — recognizing that dogs and cats are different kinds of things — they must recall specific exemplars encountered in specific moments. When they learn a new word, they must bind a sound to an object in a particular context.

Episodic memory isn't a separate cognitive skill your baby is developing alongside other skills. It is the substrate — the foundation on which object permanence, language, play, and social cognition all rest. Everything they're doing right now draws on the hippocampal encoding that just came online.

This is what Behm's review calls "the ubiquity of episodic-like memory during infancy." Memory isn't one part of your baby's cognitive life. It's the infrastructure for all of it.

What This Means for Today

You cannot make your baby remember their first birthday. You cannot prevent infantile amnesia. The neurogenesis that will overwrite these early memories is also building the brain architecture that will serve them for decades. The forgetting is a feature, not a bug.

But you can work with the memory system that's operating right now:

Time new experiences before naps. The Seehagen finding is as practical as research gets. If you're introducing something new — a new toy, a new game, a visit to a new place — try to schedule it so a nap follows within a few hours. The sleep spindles will do the rest.

Don't underestimate what they absorb. Your baby is encoding. They notice. They remember the sequence of your bedtime routine, the location of their favorite toy, the face of someone they haven't seen in weeks. The hippocampus is working. Treat them as someone who is paying attention, because they are.

Routines build architecture. Even as individual episodes fade, repeated patterns build implicit memory — the kind that doesn't require conscious recall. The bedtime routine, the morning greeting, the way you hand them their cup — these repetitions create procedural and implicit memories that persist long after the episodic details dissolve.

Emotional presence matters most. The episodic memories will go. But the emotional memories — the felt sense of safety, warmth, responsiveness — are encoded in systems that survive infantile amnesia. How you make them feel is the memory that lasts.

Don't test them. Your baby's memory is real but context-bound. They may fail to recognize Grandma in a video call but light up when she walks through the door. They may not perform on demand but surprise you with spontaneous recall days later. The memory is there. It just retrieves on its own terms.

The Paradox at the Heart of It

There is something poignant about this stage. Your baby's brain has just acquired the machinery for episodic memory — the ability to encode the specific texture of a moment, the particular quality of an experience. And almost as soon as it comes online, the same developmental processes that built it begin to overwrite it.

The neurogenesis that erases early memories is also what gives the hippocampus its extraordinary plasticity. The forgetting and the learning are the same process. Your baby is trading the permanence of these early recordings for the flexibility of a brain that can keep adapting, keep rewiring, keep growing.

They won't remember this year. Not the way you will. But the Yates study and the engram research suggest something that feels, as Turk-Browne says, almost like science fiction: the memories may not be gone. They may be dormant — inaccessible to conscious retrieval but still woven into the neural architecture, shaping who your child becomes in ways we don't yet fully understand.

So the question isn't really whether your baby remembers. At 13 months, with the hippocampus newly online, they absolutely do. The question is what happens to those memories as the brain continues to build itself — and whether, somewhere deep in the architecture, the earliest experiences persist.

The science says: probably yes. Not as stories they can tell. Not as scenes they can replay. But as traces, woven into the structure of who they are. The moments dissolve. What they built remains.