The Brain's Birth Audit and the Deep-Sea Tracks That Never Healed

Read the episode (Narration script)

Before you were born, your brain was quietly running an audit. And according to new research, a surprising number of its own cells failed.

A team at Harvard Medical School and Boston Children's Hospital has read the DNA of individual human neurons, from before birth into later life, looking for copy number variants: places where a stretch of DNA has been accidentally duplicated or deleted. Think of a recipe book where some pages have been photocopied twice and others torn out entirely.

Neurons are fussy about this. Many genes need exactly the right number of copies, and inherited copy number variants are linked to autism, epilepsy and schizophrenia. These ones weren't inherited, though. They arose in single cells as the brain was being built.

They examined more than two thousand three hundred neurons from donated brain tissue. Before birth, an average of about a quarter of neurons in the cortex carried widespread changes across multiple chromosomes. In some individuals it was nearly half. After birth, that share fell roughly four and a half times, to around six per cent.

Those cells didn't repair themselves. They seem to have been removed. The timing lines up with a well-known period of programmed cell death around birth, when the developing brain discards large numbers of neurons. That culling is usually explained as competition: neurons that make good connections receive survival signals, and the rest fade away. Genetic quality appears to be part of the contest. In measurements from nearly forty-five thousand cells, neurons with heavy loads of copy number changes showed abnormal activity in the genes for building synapses, the junctions where neurons connect. And the changes that survived into later life were less likely to involve dosage-sensitive genes or genes linked to developmental disorders.

This is postmortem tissue from a limited number of people, and the synapse mechanism is a strong suggestion rather than a demonstrated cause. Whether this checkpoint works differently in conditions such as autism is the obvious next question, and still unanswered. But it changes a basic picture of how a healthy brain gets built: less like flawless construction, more like a busy workshop with a very strict inspector.

From the inside of a skull to the bottom of the Pacific.

In 1979, a mining company drove a test machine across the seafloor in the Clarion-Clipperton Zone, between Hawaii and Mexico, collecting potato-sized mineral lumps called polymetallic nodules. Those nodules contain metals used in batteries and electronics, which is why the region has become the focus of deep-sea mining plans.

The tracks are still visible, more than four kilometres down. Last year, researchers showed that animal numbers in those tracks remained lower, with only early signs of recovery. Now a University of Liverpool-led team has looked at what's on the menu. Comparing sediment inside the old disturbance with an untouched site about two kilometres away, they found less organic matter in the mined area, and what remained was poorer food: lower in the kinds of nutrients, including essential fats, that seafloor animals rely on to grow and reproduce. Both larger animals and single-celled organisms were scarcer there.

At these depths, food arrives as a slow drizzle of particles from the surface. Scrape away the top layer of sediment and you remove the pantry along with everything else. Forty-four years later, it hasn't been restocked. This is one site, but it's a rare chance to see the deep sea on its own timescale.

Next, a molecule with a job application in quantum technology.

Future quantum networks need parts that store quantum information and send it out as light. Today's leading candidates are tiny defects in crystals such as diamond, but stray electrical charges around them can make the colour of their light wobble.

A team from the German company NVision Quantum Technologies, reporting in Nature, tried building the part from chemistry instead. They placed a carbon-based molecule inside a carefully matched host crystal, cooled it to four and a half degrees above absolute zero, and addressed single molecules one at a time. The molecule's light stayed steady for more than an hour, and its spin, the quantum property storing information, held its delicate state for milliseconds when protected with microwave pulses.

That isn't a network. Linking molecules to each other through light is the next step. But chemists can manufacture identical molecules by the billion, and that's an appealing property for a technology that depends on every part behaving alike.

Finally, a lesson in checking your work.

Studies in mice have suggested that a single dose of psilocybin, the active compound in magic mushrooms, could produce lasting changes: less anxiety, less depression-like behaviour, better unlearning of fear. Five labs at Berkeley, UCSF and Stanford decided to run the same experiments in parallel, at the same dose, with around two hundred mice per experiment, and pool everything.

The immediate effects held up everywhere. Within minutes, the mice became more anxious and avoidant, and showed less fear. But a day later, the lasting benefits didn't reliably show up. Individual labs sometimes saw an effect; taken together, it largely disappeared.

That doesn't settle anything about people, whose experience of a psychedelic involves far more than a mouse in a maze can show. It does tell researchers which mouse results are solid enough to build on, and which deserve a second look.

That's the show for today. Somewhere in your head right now are the neurons that passed inspection. Treat them kindly.