Does the Adult Human Brain Make New Neurons? Part One: Reading the Case for Persistence

Abstract

In the spring of 2018, two research groups published findings on the same question, in the same species, within weeks of each other, and reached opposite conclusions. One reported that the formation of new neurons in the human hippocampus falls to undetectable levels after childhood. The other reported that it continues into the eighth decade of life. Neither group was careless, and neither result was quietly withdrawn.

This article examines the response written by eighteen researchers working in the field, published in Cell Stem Cell later that year. Their argument is not that the negative finding was wrong in what it observed, but that what it observed cannot carry the weight of the conclusion drawn from it. Along the way they set out something more broadly useful than a verdict: a description of how tissue, timing, and choice of marker can each erase a signal that was present in life.

We treat this as the first of three readings. The second takes the opposing reply on its own terms. The third steps outside the hippocampus altogether. Our purpose is not to settle the question — it is not settled — but to show what a genuine scientific disagreement looks like from the inside.

1.Two Papers, One Question, Opposite Answers

The two reports at the centre of this dispute are easy to state. Sorrells and colleagues, publishing in Nature, examined human hippocampal tissue across a range of ages and concluded that recruitment of new neurons declines steeply during childhood and becomes undetectable in adults. Boldrini and colleagues, publishing in Cell Stem Cell, examined a different set of human hippocampal tissue and concluded that the process persists throughout aging.

The disagreement is not about interpretation of a shared dataset. Each group looked at different brains, with different histories, prepared in different ways, and stained for overlapping but not identical sets of proteins. That is the crux of the matter, and it is why the response from the wider field concentrates almost entirely on method rather than on conclusions.

The authors of that response are worth naming as a group rather than individually, because the composition is itself part of the argument: eighteen investigators, including several whose own laboratories produced the foundational human studies of the preceding two decades. Their position is stated without hedging. There is, they argue, currently no reason to abandon the idea that adult-generated neurons contribute to plasticity and cognition across the human lifespan.

2.Why a Marker Can Disappear Without the Cells Disappearing

The negative finding rested largely on the absence of two proteins, doublecortin and PSA-NCAM, together with the absence of the elongated morphology typical of young neurons. The response argues that each of these can fail to appear for reasons that have nothing to do with whether the cells existed.

Doublecortin is fragile. Controlled work in rodents shows that staining for it weakens within hours of death. Human postmortem tissue is rarely obtained within hours. In the two studies at issue, the permitted interval between death and fixation differed — up to twenty-six hours in one, up to forty-eight in the other — and that difference alone sits comfortably inside the window in which the signal degrades.

Fixation introduces a second problem in the opposite direction. Human samples are sometimes held in formalin for years, and prolonged fixation masks antigens rather than destroying them: the protein is still there, but the antibody can no longer reach it. The authors point specifically to PSA-NCAM as susceptible to this, which would produce exactly the pattern reported — a marker that appears absent because it has been rendered invisible.

A third factor is stranger and, to us, the most striking detail in the paper. In bats, doublecortin staining drops sharply within thirty minutes of capture. The proposed mechanism is stress hormones. If that holds in humans, then the circumstances of a person's final hours — illness, intensive care, physiological stress — become a variable in what their tissue will show, and one that no laboratory protocol can correct after the fact.

3.What Counts as Evidence for Something You Cannot Watch Happen

Marker staining is not the only way to ask the question, and the response leans on the alternatives. Two are worth understanding, because they approach the problem from a completely different direction: instead of looking for cells that appear young, they establish when a cell was born.

The first method arrived by accident of clinical practice. In 1998, Eriksson and colleagues examined hippocampal tissue from cancer patients who had received bromodeoxyuridine, a compound that incorporates into DNA during cell division and was given for diagnostic purposes. Cells that divided during treatment carried a permanent label. New neurons were found. The sample was five brains.

The second method is more ingenious still. Atmospheric nuclear testing during the Cold War raised global carbon-14 levels, which then declined in a known curve after the test ban. DNA synthesised during those decades carries an isotopic signature that dates it. Spalding and colleagues used this to estimate the rate of neuronal addition in the adult human hippocampus, arriving at a figure of roughly seven hundred new neurons per day per dentate gyrus.

Neither approach depends on a protein surviving postmortem delay. Both are, in the response's framing, evidence about the end product rather than about intermediate stages — a demonstration that new neurons are present, rather than an inference from cells that look immature.

4.The Arithmetic That Reconciles Sparse Staining With Real Turnover

There is a numerical argument in the paper that deserves more attention than it usually receives, because it addresses an objection that seems, at first hearing, decisive: if new neurons are genuinely being added every day, why do studies find so few cells expressing markers of immaturity?

The answer turns on how long a cell stays in the labelled state. If doublecortin expression spans roughly three weeks, then a daily addition rate of several hundred cells implies a standing population of immature cells numbering in the thousands at any given moment — not the millions that intuition might suggest. The authors note that even allowing a wide margin of error, the counts of doublecortin-expressing cells reported in the literature fall in the same order of magnitude as the carbon-dating estimates predict.

This is a modest claim and it is presented as one. It does not prove the marker studies were correct. It establishes something narrower and more useful: that sparse staining and continuous neurogenesis are not in contradiction, and that a low count cannot be read directly as a low rate.

5.What the Response Concedes

A document written by eighteen people defending a position could easily contain no concessions at all. This one contains several, and they are more interesting than the defence.

The authors accept that the individual variation reported in the positive study is large, with some subjects showing very low counts. They accept that proper stereological counting — the statistical machinery that turns observations in a tissue section into an estimate for a whole structure — is difficult to apply rigorously to the kind of human samples typically available. They accept, explicitly, that it is conceivable, if unlikely, that the subjects in the negative study genuinely had little or no ongoing neurogenesis.

Most consequentially, they raise the possibility that the human timeline simply differs from the rodent one. If maturation of a new neuron takes months in primates rather than weeks, and if the period of heightened plasticity is not fully captured by the window in which doublecortin is expressed, then a marker validated in mice may be measuring the wrong interval in humans. They label this speculation. It is the kind of speculation that reframes a field.

6.Absence of Evidence

The logical spine of the response is an old principle applied to a specific case. A study that fails to detect something has two possible explanations: the thing was not there, or the method could not see it. Distinguishing between them requires showing that the method would have detected the thing had it been present.

This is precisely the point at which the two camps part company, and it is worth flagging now because it becomes the central issue in the second part of this series. The opposing group does not accept the framing. They argue that they demonstrated their method's sensitivity, and that the burden has been discharged. Whether it was is the substance of the disagreement.

7.What This Episode Teaches About Reading Science

We find this exchange valuable for a reason that has little to do with neurons. It is a clean demonstration that a published, peer-reviewed, methodologically competent study in a leading journal can produce a finding that other competent researchers regard as an artefact — and that this is normal, not scandalous.

It also demonstrates how much depends on invisible decisions. How long the tissue waited. What it was fixed in and for how long. Which antibody was used. What the person was dying of. None of these appear in a headline, and most do not appear in a summary. They determine the result.

The honest position at the end of this first reading is not that adult human neurogenesis has been confirmed. It is that the strongest negative finding to date is compatible with several explanations, only one of which is that the phenomenon does not occur — and that the case for its occurrence rests on methods, particularly isotopic birthdating, that the negative finding does not address. The reply to that argument is the subject of Part Two.

Editorial Comment

MindHeaven® makes no claim that any product influences the formation of new neurons in the human brain, and nothing in this literature would support such a claim. We publish this series because the question of how the adult brain changes is foundational to everything else we write about, and because the disagreement itself is instructive. A field capable of arguing this openly about its own core finding is a field worth reading carefully.