A better brain probe

SCIENCE is a mixture of the intellectual and the practical. And the practical requires tools. Until the invention of the telescope, astronomy had been stuck in a rut for millennia. Until the invention of the microscope, microbiology did not exist.

Neuroscience, too, has advanced recently on the back of some powerful tools, particularly techniques for scanning whole brains. But the devices that look at the nitty-gritty of how nerve cells themselves work are still Heath-Robinson affairs. These are the electrodes that record the impulses of individual cells, ideally simultaneously with lots of others, in order to try to work out how networks of cells process information.

That may change with a device described this week in #Nature. The business end of Neuropixels, as the new tool is known, is a probe made in the way that computer chips are made, by photolithography. This probe (see picture) is 1cm long and 70 microns across—about the width of a human hair. It is capable of recording signals from 384 nerve cells at the same time. These signals are gathered individually by electrodes 12 microns across that cover the probe’s surface. The electrodes are made from titanium nitride, a material chosen because it is both amenable to photolithography and can survive for up to six months inside a brain.

Each electrode records voltage changes in the nerve cell nearest to it and transmits those changes to the top of the probe through a conductive channel (one per electrode) that acts like a wire. There, the signals are picked up and deciphered by a chip that converts the analogue electrical impulses from each of the 384 channels into digital bits, and then pushes those bits out to a computer for analysis through just four wires. The whole set-up weighs just 250mg. That means it can be carried around without too much trouble by an experimental animal, such as a rat, while it is making recordings of what is going on inside the animal’s brain. Moreover, in a small brain like a rat’s, a 1cm probe can traverse several brain areas. It can therefore record interactions between those areas, and thus start to get a grip on how the brain works as a whole.

Devices that record the simultaneous doings of hundreds of nerve cells do already exist, but they are complicated and temperamental. They come with a crown of hundreds of fine, fiddly wires that stick out of the skull and have to be manipulated individually. Neuropixels, by contrast, can just be plugged into the brain being studied. It is this ease of use that is expected to give it its edge.

#Neuropixels is the brainchild of a collaboration between several medical-research foundations. The project was led by Tim Harris of the Howard Hughes Medical Institute’s Janelia Research Campus, in Virginia. He and his colleagues worked with teams from the Allen Institute for Brain Science, in Seattle, and the Gatsby Charitable Foundation and the Wellcome Trust, both in London. The device itself was fabricated by imec, a boutique, not-for-profit microelectronics shop in Belgium.

It all sounds very charitable—and it is, for the intention is to advance the field as a whole rather than to garner profit for a particular organisation (imec will continue to manufacture it, but will sell it at cost, rather than marking the price up). But the goal of the project was always, according to Dr Harris, to design a robust, commercial device, rather an academic laboratory widget. This, the team believe, they have done. Dr Harris says it is likely to sell for around €1,400 ($1,600) and will be available towards the end of next year.