Here’s something that, when you first hear it, tends to change how you think about adolescence — your own if you’re living through it, your children’s if you’re raising them, or your memories of it if you’re past it.
The human brain doesn’t finish developing until somewhere around the mid-twenties.
The number varies a bit depending on the study, and the word “finish” is doing some work — different parts of the brain mature on different schedules, and there’s no clean cut-off. But by the standard of the rest of your body, which is essentially done by about sixteen, the brain is a strange outlier. The last parts to come online — the regions responsible for long-term planning, impulse control, weighing risks against rewards, and resisting peer influence — are still under construction well past the age at which society tends to assume a person should be acting like a fully-formed adult.
This finding isn’t controversial in neuroscience. It’s one of the most robust discoveries of the last thirty years. What’s changed, as a result, is how researchers think about nearly everything associated with the teenage years: risk-taking, emotional intensity, the weight of social acceptance, the tendency to act in ways that older adults find baffling. The picture that has emerged is more interesting than the usual one, and worth understanding carefully.
The MRI studies that made it visible
The breakthrough came from a neuroscientist named Jay Giedd, working at the US National Institute of Mental Health in the 1990s and 2000s. Giedd had a simple and audacious idea: scan the brains of hundreds of healthy children, adolescents and young adults, and see how the structure changed over time.
He recruited thousands of participants and scanned them, some repeatedly, over years. What he found transformed developmental neuroscience. The brain wasn’t, as previously assumed, largely done developing by late childhood. Instead, major structural changes were continuing through the teens and into the early twenties. Grey matter — the part of the brain where most computation happens — peaked in volume around age eleven for girls and age twelve for boys, and then slowly declined as the brain pruned unused connections and reinforced used ones.
More importantly, Giedd’s scans showed that different regions were maturing on different schedules. The amygdala and limbic system — the older, emotionally-loaded parts of the brain responsible for feelings and drives — were largely mature by early adolescence. But the prefrontal cortex — the newer, executive part of the brain responsible for planning, impulse control, and weighing long-term consequences — wasn’t fully developed until the mid-twenties.
This asymmetry matters enormously. Between roughly thirteen and twenty-three, humans have an essentially adult emotional and reward system coupled to a still-developing system of regulation. The accelerator has arrived before the brakes. This is not a character flaw or a deficiency of willpower. It’s the predictable consequence of the schedule on which the human brain physically matures.
Why teenagers take risks
The most influential interpretation of Giedd’s findings came from a developmental psychologist named Laurence Steinberg at Temple University. Steinberg’s research, conducted over decades, made a specific and important argument: teenagers take risks not because they underestimate danger or overestimate their abilities, but because they experience rewards — especially social rewards — with unusually high intensity.
Steinberg’s experiments found something counterintuitive. When teenagers were alone and asked to take calculated risks, they behaved roughly as adults did. They understood probabilities. They weighed consequences. They made sensible choices. But when teenagers were placed in the same decision-making scenarios in front of peers, their behaviour changed sharply. They became much more willing to take risks, to pursue short-term rewards, to ignore long-term consequences. The same adolescent brain that was sensible in solitude became reckless in company.
Adult brains, by contrast, didn’t show this effect. Adults made similar choices whether alone or watched by peers. The difference wasn’t that teenagers didn’t know any better. It was that the reward system in the adolescent brain responds more intensely to social approval than the adult brain does — and the still-developing prefrontal cortex isn’t yet able to reliably override those intense signals.
This framing explains much of what’s otherwise puzzling about adolescence. The group that drives too fast together; the public displays of bravado; the exaggerated loyalty to friends against family; the apparent willingness to do things alone that look unwise and together that look genuinely dangerous. None of this is evidence of impaired reasoning. It’s evidence of a brain wired, for a specific developmental window, to weight social approval and novelty unusually heavily.
Why rejection hurts more
A parallel finding, from the British neuroscientist Sarah-Jayne Blakemore at University College London, extended the picture further. Blakemore’s research, using functional MRI, found that the teenage brain responds more strongly to social rejection than the adult brain does. Images of being left out, of conversations without them, of small signs of exclusion, produce larger and longer-lasting activation in the brain regions associated with emotional pain in teenagers than in adults.
This isn’t an adolescent overreaction. It’s a physiological fact. Teenagers don’t just claim to feel rejection more intensely than adults; they actually process it differently at the neural level. The same snub that a forty-year-old can shrug off will keep a fifteen-year-old awake, genuinely suffering, in a way that’s often invisible to the adults around them.
Blakemore’s work has clarified something that used to be dismissed as teenage melodrama. It isn’t melodrama. It’s a developmental feature. The human brain is wired, during the years when young people are forming their identities and place in the world, to take social information with maximum seriousness. This wiring was presumably adaptive in our evolutionary past — young humans had to navigate complex social hierarchies, build coalitions, and integrate themselves into communities to survive. A brain calibrated to feel social rejection intensely was a brain that worked hard to avoid producing it.
The dual-systems model
A neuroscientist named B.J. Casey pulled these threads together into what’s now called the dual-systems model of adolescence. The basic idea: two systems develop on different timescales. A hot system — responsible for emotional responses, reward-seeking, sensitivity to peers — matures early and runs at full intensity through the teens. A cool system — responsible for regulation, planning, and override — matures later and isn’t fully online until the mid-twenties.
The gap between the two, roughly ages 13 to 23, is the window of adolescent vulnerability to risk-taking, emotional intensity, and impulsive decisions. It’s also, importantly, the window in which humans are developmentally primed for learning, identity formation, exploration, and the intense relationships that shape the rest of a life. The same architecture that produces the problem produces the adaptive features.
This last point is worth emphasising, because it’s often missed. The adolescent brain isn’t a broken adult brain. It’s a differently-tuned brain, optimised for a specific developmental task. Its features are not bugs. Heightened sensitivity to peer approval helps young people integrate into adult society. Heightened emotional responsiveness drives the deep relationships that define lifelong identities. Willingness to take risks enables exploration of options older adults wouldn’t consider. Strip away these features and you’d get a compliant, unexploring young person — which is not the kind of young person societies actually need.
The counter-thread worth hearing
A measured counter-argument has come from the developmental psychologist Dan Romer at the University of Pennsylvania, who argues that the popular “immature brain” framing has been overstated and sometimes misleading. Romer’s research points out several things the simple story gets wrong.
First, adolescent risk-taking isn’t uniform. Most teenagers don’t actually engage in serious risk behaviours. The popular narrative often conflates the tail of the distribution — the minority taking serious risks — with the average teenager, who is mostly quite cautious and rule-following. Data on teenage crime, substance use, and dangerous behaviour generally shows most teenagers not engaging in these things.
Second, Romer argues that framing adolescence primarily as a period of neural deficiency can be self-fulfilling. When adults treat teenagers as incapable of responsible judgement, teenagers often rise to those expectations in the negative direction. Several cross-cultural studies have found that in societies where young people are given substantial adult responsibility earlier, the risky behaviours associated with adolescence in Western industrialised societies are much less pronounced.
Third, Romer points out that the “incomplete prefrontal cortex” story, while true, can overlook what the adolescent brain is actually good at. Adolescence is a period of unusual neural plasticity, openness to new experiences, and capacity for fast learning. A brain that’s still being built is also a brain that can still be shaped by experience in ways that the more settled adult brain cannot.
The honest picture, integrating the main research and Romer’s critique, is probably this: the adolescent brain really is structurally different from the adult brain; the differences really do predispose toward specific patterns of behaviour; but those differences also produce real strengths; and how adolescents actually behave depends heavily on the social environment surrounding them, not just their internal neurobiology.
What to do with this
The practical implications of this research are broad and worth holding.
For people currently in or just past adolescence, it offers a kind of permission: the intensity of your feelings, the importance of your friendships, the power of social approval in your decisions, are not deficiencies of your character. They’re features of a specific developmental stage. Many of the judgements you’re making now, you won’t quite recognise in ten years. This isn’t because future-you will be wiser. It’s because future-you will have a more fully-integrated brain. The judgements that feel obvious right now may look different later, and that’s normal rather than alarming.
For people past adolescence — especially parents, teachers and employers working with young adults — the research suggests patience calibrated to neurobiology rather than expectations of adult-equivalent functioning. Young people who make impulsive choices, get too caught up in their relationships, pursue questionable-looking passions or agonise over socially trivial setbacks are not failing at being adults. They’re succeeding at being adolescents. The developmental task of this period is not to act like a thirty-five-year-old. It’s to develop into one.
The question that remains
Perhaps the most useful take-away is a reframe. The “teenage brain” isn’t a broken version of the adult brain. It’s a stage, with its own logic, its own affordances, and its own limits. Understanding this doesn’t excuse dangerous behaviour, but it changes what we make of ordinary adolescence — the intensity, the risks, the friendships, the sense that everything matters this much right now. All of these are, in a sense, what the brain is for at this age.
The question worth holding, if you’re young enough for it to apply, or thinking about someone who is:
What does the developmental work of being this age actually look like — and is it being respected, or treated as a problem that you should have already outgrown?
Key research referenced: Jay Giedd’s MRI studies at NIMH (Giedd et al., 1999; 2008); Laurence Steinberg’s research on adolescent risk-taking (Steinberg, 2008; Age of Opportunity, 2014); Sarah-Jayne Blakemore’s work on the social brain in adolescence (Inventing Ourselves, 2018); B.J. Casey’s dual-systems model; Dan Romer’s counter-research on adolescent cognition.