What is your earliest memory? A frightening fall down the stairs? Blowing out candles on your third birthday? Or perhaps it is a trip to the hospital to visit a newborn sibling? Whatever the content, it is probably short and rather hazy. Adult recollections of infancy and early childhood are typically fragmentary. We forget so much, in fact, that psychologists have coined the term “infantile amnesia” to describe the profound memory loss associated with the start of life. Indeed, infantile amnesia seems to create a paradox concerning the brain’s sensitivity to early experiences. From one viewpoint, there is considerable evidence that early experiences impact the development of the brain. Adult social behaviors, resistance to stress, and some language skills are clearly affected by what happens during the first stages of life. But if the brain is so strongly affected by what happens in these early days, why can’t we remember any of it?

Despite the importance of early learning, memories formed during infancy seem more fragile than those formed later in life. One 1962 study illustrated this point well through an investigation of long-term memory in rats. Researchers trained rats of different ages to fear one compartment of a double chamber, and then tested how long the rats remembered the experience. They found that the ability to remember increased dramatically with age. Infant rats avoided the fear chamber immediately after training, but didn’t remember the training three weeks later. In contrast, rats trained as adults remembered to avoid the fear chamber long after the trained occurred.

Research on infantile amnesia in humans relies on more anecdotal evidence. Some people can recall a few memories formed when they were as young as 2 or 3 years of age, but most of us can recall much more from when we were 5 or 6 years old. Studies suggest that we’re not simply forgetting what happened during our earliest years; far fewer autobiographical memories exist from early childhood than simple forgetting predicts. So the fate of early memories remains puzzling; solving the mystery of infantile amnesia may go a long way towards a more general theory about how we remember and why we forget.

What Happens To Early Memories?
Theories about infantile amnesia can be divided into two broad categories: those which hold that the memory loss is due to a storage difficulty (i.e., early experiences are not properly transformed into long-term memories) and those that claim the memory loss is a retrieval problem (i.e., the memories exist, but we can’t recollect them).

Possibly A Storage Failure
The idea that infantile amnesia may be caused by inadequate memory formation stems from studies which show that the neural circuitry of the brain is not fully functional in infants. For example, we know that much of the visual system is still developing after birth, and that myelination in many cortical areas isn’t completed for quite a while. In many animal species, the hippocampus, a brain structure that is critical for many types of memory formation, is not entirely developed at birth. Numerous studies have illustrated that rats improve markedly on memory tasks 18 to 23 days after birth, during the time that the hippocampus becomes mature. In humans, however, the hippocampus seems nearly mature at birth, so hippocampal development is probably not at the heart of infantile amnesia. Instead, research has shown that maturation of the infereotemporal cortex and the prefrontal cortex corresponds with the improvement on a number of memory tasks. The activity of these regions may be the key to the whereabouts of our earliest memories.

The Language Link
Perhaps the largest developmental change in humans is the acquisition of language, which generally coincides with when we start to remember things (age 3-4 years). This close association has led some researchers to suggest that language development allows internal and external rehearsal of experiences, and hence better storage in long-term memory. Though there is little doubt that memory for autobiographical experiences and language ability must be linked to some degree, the fact that pre-verbal babies can demonstrate functional memories suggests that language not is necessary for long term memory storage or retrieval.

So although there is definite evidence that brains of youngsters are still developing, there seems to be little persuasive data demonstrating that the immature nervous system is incapable of storing memories per se. Pre-school children as young as 2 years old are able to recall events after intervals of 6 months or more, so by this point long term memory storage is clearly functional. But by the time those children become adults, the memories formed during early years are difficult to recall. This failure of memory suggests a problem with retrieval rather than storage, and indeed there is increasing evidence to support this hypothesis.

Retrieval Failure
A multitude of clever experiments illustrate that young memory capabilities are more extensive than previously believed. For example, infants as young as 5 months prefer novel pictures over those already presented to them, even if the second viewing occurs several weeks after the initial presentation.

Among the most definitive studies on early memories were conducted by Carolyn Rovee-Collier. Rovee-Collier taught 2 to 6 month old infants that kicking a mobile strung above their cribs would activate the mobile’s motor. The infants were tested again days later to see if they could remember how to activate the mobile. Babies as young as 2 months old remembered the foot kick technique for 1-3 days, and 6 month old infants remembered the correct action for 15-16 days. Interestingly, Rovee-Collier also found that reminding the babies of the activity by moving the mobile herself prompted the youngest infants to remember the foot kick, suggesting that the memory was still present but the infants needed help in retrieving it.

If infantile amnesia is due to our inability to access old memories, one must ask what causes the retrieval problem. One strong possibility is that growing up alters our perception to such a degree that appropriate retrieval cues are never presented. For example, when you were 6 months old, everything in the world probably seemed huge. Your memories of that time are probably colored by towering furniture and lots of talk that you could not yet understand. Accessing these memories might be difficult in the adult world, where tables are no longer three times as big as you are. As we age, our view of the world changes so much that the cues that we associated with our earliest memories are no longer present, so we lose our connection to infant memories. Indeed, evidence suggests that infant memory retrieval might be highly dependent on contextual cues. If Rovee-Collier changed one of the objects hanging from the mobiles in her experiment, infants no longer remembered how to activate it with a kick. Given that infantile memory retrieval seems to occur only when infants encounter stimuli that are identical to those present during training, developmental changes in perception would make it significantly harder to remember what happened after a long period of time.

Multiple Memory Systems
One of the important insights to emerge over the last 20 years is that memory is not a single entity but that it is made up of several abilities controlled by distinct brain systems. The major distinction is between conscious recollection of experience, called declarative memory (like remembering your childhood best friend), and unconscious memories of skills and habits, called nondeclarative memory (like your effortless backhand at tennis). The unconscious memory system appears available from birth, while a conscious memory counterpart develops later.

A fascinating series of studies that illustrates the impressive memory abilities of infants comes from Andrew Meltzoff. He argues that Rovee-Collier’s work with the mobiles tests only the unconscious “skills and habits” memories. Thus, he designed a test to see if infants were capable of some form of conscious declarative memory. The experiment proceeded as follows: 14- to 16-month old infants were shown a demonstration of a series of novel toys, such as collapsible cups, each accompanied by a specific “target action.” One group of infants were allowed to play with the toys immediately afterward, the second group was only permitted to watch. The children were tested either 2 or 4 months after the initial visit. During the test, the children were allowed to play with the toys while an observer rated the number of times they performed the predefined “target actions.” After two months, both groups of infants demonstrated good memory about how to play with the toys, performing the target actions frequently. Results at the 4 month test were similar, though there was a substantial decline in the number of times the infants performed target actions, suggesting that the children might be starting to forget. Nevertheless, the results are quite surprising. With only one exposure to the toys lasting just one minute, infants were able to remember multiple target actions up to four months after the demonstration.

Emotional Memories May Be More Available for Recall
In recent years there has been increased awareness about the role of emotion in the modulation of memory, accompanied by the discovery that certain brain structures like the amygdala are specialized for emotional learning. Moreover, some researchers have found that high levels of stress may actually benefit recall. The links between emotion, stress, and memory have led scientists to wonder whether there might be less infantile amnesia associated with traumatic childhood events.

Ulric Neisser and collegues at Cornell University examined this possibility in a study investigating college students’ recall of four specific, life-altering events that occurred when the students were between 1 and 5 years old. The students answered questions pertaining to the birth of younger siblings, death of a family member, moving to a new home, and hospitalization, and their answers were verified with their parents. Sibling birth and hospitalization seemed especially memorable, since college students could recall these events even if they were only 2 years old when the experiences occurred. Moving and a family member’s death seemed to emerge from the haze of amnesia around the more traditionally accepted age of 3. There were no verifiable reports for recollection of events occurring before age 2.

Future Directions
The question of the storage failure versus retrieval failure remains unresolved, but the latest multiple-memory system approach actually incorporates both concepts. The relation between emotion and early memory is perhaps the least understood aspect of infantile amnesia. While there is some data to support the notion that emotions reinforce early experiences, the picture is far from complete. Research so far has relied on individual recollections of events that happened many years in the past; those memories have no doubt faded and have been influenced by experiences like family stories, i.e. “You were so happy when your baby brother was born!” The need for controlled studies where the subject’s emotion can be recorded as experienced is imperative.

The emerging picture of infantile amnesia depicts a complex process, perhaps mediated by several, differentially-developing memory systems. With future research using long-term studies, measures of both conscious and unconscious memory, and better controlled emotional memory studies, we may yet gain some insight into the mysteries of our earliest days.