What is your first memory? A birthday party, a fall in the playground, a family trip? For most people, it dates to around age 3 or 4 at the earliest. Infant and toddler experiences are, seemingly, absent from our autobiography. Sigmund Freud named this phenomenon infantile amnesia and argued that early-life memories are so emotionally charged that they must be actively repressed. In this view, infants are capable of forming memories for important life events, but these memories become inaccessible over time (although he argued that psychotherapy might, in principle, recover them).
More contemporary psychological accounts have linked infantile amnesia to the immaturity of the hippocampus, a brain structure critical for forming episodic memories in older children and adults. According to these accounts, hippocampal circuits are not sufficiently developed in infancy to form enduring traces of life events. That is, infantile amnesia reflects a failure to encode experiences.
Here, we argue for a different view. Early-life experiences are encoded by the hippocampus and may persist over time. What changes is how easily they can be accessed. Infantile amnesia, therefore, reflects not a failure to encode memories but a failure to naturally retrieve them.
nfantile amnesia is not uniquely human. Nearly all mammals that have been studied exhibit accelerated forgetting in infancy. For example, similar to young children, mice and rats can form memories of specific events, but these memories do not endure. This phenomenon has been demonstrated across a range of infant-acquired experiences, including fear conditioning, object recognition and spatial learning. The exception is highly precocial animals, such as guinea pigs and degus, whose brains are much more mature at birth.
The advent of activity-dependent engram labeling and optogenetic tools has enabled researchers to directly address the encoding versus retrieval debate. In mice, researchers can tag engrams in infant pups, track their persistence across development and optogenetically reactivate them later in life to recover seemingly lost memories. Animals that undergo contextual fear conditioning in infancy show no behavioral evidence of remembering when tested in adulthood (that is, they do not freeze when placed back in a context where they were shocked as pups). However, optogenetic stimulation of the neurons that were active when the pup was first conditioned triggers expression of the memory; the mature animal now freezes.
Importantly, it is not just optogenetically activating any population of cells that produces the freezing response, only those that were active during infant learning. Together, these findings indicate that early-life memories are successfully encoded and may persist over time but cannot be accessed through natural sensory retrieval cues—pointing to a failure of retrieval rather than encoding.
These mouse results also suggest that infantile amnesia is unlikely to arise entirely from human-centric factors, such as the emergence of language, a coherent sense of self, or the repression of early emotional experiences. Instead, they point to a more fundamental process in which early memories are formed and stored but become increasingly inaccessible as the brain and its modes of retrieval change across development.
Of course, mouse studies cannot access the autobiographical, linguistic and subjective dimensions of remembering that are central to the human experience. A growing body of research is digging into the development of episodic memory in human infants. Researchers can measure behavior, physiology and, increasingly, hippocampal function itself. These studies can capture the richness of human development: changing perception, emerging language, growing motor abilities and engagement with the social world in which memories are embedded.
Research on human infants reveals that memory systems are active far earlier than once assumed; behavioral studies show that infants learn and remember from birth. But the tasks these studies employ involve motor actions, repeated training and reward or feedback. These requirements contribute to forms of memory, such as operant conditioning and priming, that can be dissociated from episodic memory and that engage brain regions other than the hippocampus, including the striatum, cerebellum and sensory systems.
Studying hippocampal-based memory in human infants has been challenging because of technical limitations; the neuroimaging methods typically used in this population, including electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS), measure scalp signals and cannot reach deep-brain structures. But this has changed over the past decade with the advent of functional MRI (fMRI) in awake infants. Greater hippocampal activity while infants view photographs of objects, faces and scenes—the “building blocks” of episodic memories—predicts whether they later remember those photographs. Remembering, in this case, is inferred from the fact that infants spend more time looking at photographs they have seen before than at new ones. This type of hippocampal encoding starts around 9 to 12 months of age.
W
During development, the brain changes how it interprets the world. We learn language, become experts at recognizing objects and faces, organize experiences into concepts and begin to understand other people’s intentions. As a result, the same sights and sounds that once triggered a memory in infancy no longer produce the same pattern of brain activity. The hippocampal engram is still there, but the cues needed to reactivate it may no longer match, preventing the memory from being naturally expressed. In this view, cognitive and brain development gradually reduce access to early memories rather than erase them.
In unpacking infantile amnesia, human and mouse studies have each made profound, yet still incomplete, progress. Human work helps define what changes across development; mouse work helps reveal how those changes are implemented in circuits and cells. Together, they offer a bridge between mechanism and behavior, between hippocampal development and lived experience. The future of infantile amnesia research will depend not on choosing between species or levels of analysis but on integrating them—building theories that can move from infant behavior to engram biology and back again. Infantile amnesia is not simply a story about forgetting the past. It is an opportunity to understand how memory itself is built.
