Around the time of birth, the developing human brain mysteriously tidies up by purging itself of a vast number of mutation-riddled neurons, according to a painstaking analysis of more than 2,300 nerve cells.

We are all mosaics: vast genetic diversity found between cells in a single person

The analysis, published in Nature¹ on 28 October, found that more than one-quarter of the sampled neurons from the brains of human fetuses had either gained or lost several segments of their chromosomes. By contrast, only about 6% of neurons from the brains of babies after birth had such abnormalities. “It suggests that the neurons that survive may be the ones that have kept the most intact genomes during development,” says Christos Proukakis, a neurologist at University College London, who was not involved in the study.

The findings reveal a new mechanism for regulating how many neurons a person has and for deciding which fetal neurons are allowed to persist in the brain, says Diane Shao, a paediatric neurogeneticist at Boston Children’s Hospital in Massachusetts and first author on the paper.

Learning more about this process could be key to understanding neurological conditions that often appear in childhood, such as autism spectrum disorder (ASD) and epilepsy.

Quality control

Studies in mice² have suggested that, during early development, mammals eliminate many brain cells by activating a molecular programme that leads to the cells’ deaths. Shao and her colleagues analysed human brain development by studying the brains of deceased individuals ranging from fetuses at 14 weeks of gestation to nonagenarians. The team focused on neurons in the cerebral cortex, a region of the brain responsible for memory, learning and consciousness, among other functions.

How to make a brain: new experiments challenge existing picture

The team analysed the genomes of individual neurons to look for those that had either lost or gained some chromosomal segments, a phenomenon called copy number variation. The analysis revealed that a stunning number of cells with widespread copy number variation were present during fetal development. There were so many that Shao and her colleagues spent years trying to work out whether the results were real. But even after repeating the experiment using a different single-cell sequencing method, the results were similar.

After birth, however, the proportion of these abnormal neurons dropped by more than fourfold, suggesting that they had been culled during development.

How does the brain know which neurons are normal and which are not? One possible clue, says Shao, is that those brain cells with high levels of copy number variation also tended to show abnormal activity of genes involved in forming connections, or synapses, with other neurons.