Between the mid-teens and the mid-twenties, the human brain records an unusually rich collection of memories—not because it works better than before or after, but because life is packed with firsts that rarely happen again.

Ask most adults over forty to list the memories that come back to them unprompted, and a strange pattern shows up. They do not distribute evenly across a life. They cluster. Disproportionately, they come from somewhere between the ages of about ten and thirty, with the heaviest concentration falling in the mid-teens to mid-twenties. Childhood before that stretch is thin. Middle age after it, despite being longer and often busier, produces comparatively little that surfaces without effort.

Memory researchers have a name for this: the reminiscence bump. It shows up so consistently, across word-cued recall tasks, autobiographical interviews, and studies of which public events people remember best, that a 2018 systematic review in PLOS ONE by Khadeeja Munawar, Sara Kuhn and Shamsul Haque, which pooled 68 separate studies, called it one of the more robust findings in the whole field of autobiographical memory. The exact boundaries move depending on how researchers ask the question, but the review put the bump for important personal memories at roughly ten to thirty years of age, with word-cued memories skewing a little earlier and life-script style events spread a little wider.

The obvious explanation

The instinctive account, the one most people reach for without being prompted, is that this period is simply packed with firsts. First kiss, first heartbreak, first time living away from a family home, first real job, first serious argument with a friend that mattered enough to change the friendship. Adolescence and early adulthood front-load a life with events that, by definition, cannot happen again in quite the same way.

A first is, structurally, unrepeatable.

This is part of what researchers call the cognitive or novelty account, one of several explanations set out by the psychologist David Rubin and colleagues in the late 1990s. The logic holds up on its face: novel experiences are thought to be encoded more deeply than routine ones, and a stretch of life with an unusually high density of novel experiences should, on that logic, leave behind an unusually high density of retrievable memories. It is a tidy story, and it is not wrong to notice that this period is unusually front-loaded with firsts.

Where the tidy story runs into trouble

The trouble is that when researchers have gone looking for the fingerprint of novelty directly, it has not shown up cleanly. In an internet-based study of nearly 3,500 participants, the psychologists Steve Janssen and Jaap Murre gave people cue words and asked them to recall a personal memory for each one, date it, and rate how novel, emotional, positive or negative, and important it had felt. If the novelty account were the whole story, memories from the bump years should have scored higher on those ratings than memories from other periods. They did not. Published in the Quarterly Journal of Experimental Psychology in 2008, the finding was that the bump, in this dataset, actually contained relatively fewer memories rated as novel, emotional or important, not more.

This is one study, using one method, and it should not be read as settling the question. Other cueing methods produce different pictures. A 2016 review by Jonathan Koppel and David Rubin in Current Directions in Psychological Science found that the size and location of the bump shifts depending on how memory is prompted: word cues produce a smaller bump that sits earlier in life, asking people for their most important memories produces a larger bump that sits later, and odour cues, unusually, pull the bump earlier still. In our reading, that instability is the more interesting finding. The reminiscence bump is not one fixed thing waiting to be measured. It changes shape depending on what a researcher asks for.

What the material actually points to

If pure novelty does not carry the explanation on its own, two other accounts do more of the work in the literature. The first is the narrative or identity account: this stretch of life is when a person is actively assembling the story of who they are, and events that get folded into that story are the ones that stay retrievable, whether or not they were especially novel in the moment. The second is the cultural life script account, associated with the Danish psychologist Dorthe Berntsen and Rubin, which holds that every culture carries a rough, shared template of events expected to happen roughly in order: finishing school, leaving home, forming a serious relationship, having a first child. People recall these transitional events disproportionately from the age range in which their culture expects them to happen, and expect them to be positive, even when an individual life did not run that way.

These two accounts do not throw out the intuition behind the “firsts” idea so much as relocate it. A first job matters not because it is chronologically novel but because it gets used, later, as a landmark in the story a person tells about becoming an adult. The unrepeatability of a first experience is less about the brain logging something unprecedented and more about that experience being available for use as an identity marker for everything that follows.

Firstness matters less for what it does at the moment of encoding, and more for what it gets used for afterward.

What this does not show

None of this amounts to a single settled mechanism. The bump has been documented across cultures with different scripts, in clinical and non-clinical samples, and using methods ranging from cue words to diary studies to tests of which news events people remember best, and the exact temporal window still shifts depending on which of those methods is used. Researchers studying the phenomenon also rely heavily on retrospective self-report from adults looking back, which is a different kind of evidence to a real-time record of what actually got encoded in the moment. The pattern is well established. The single mechanism behind it is not.

What is left, on the available evidence, is something more modest than the popular version of the story. The mid-teens to mid-twenties do not appear to work better as a recording device than the decades either side of them. They appear to be doing different work: producing the material that a person will spend the rest of their life drawing on to explain who they became, and how.