Back Scatter: The firing neurons of a mouse on the move

Electrode arrays can pick up signals from the firing of individual neurons, which are small (about 10 µm across) and densely packed (about 100 000 per mm³). Turning such electrode arrays into useful neural probes requires a trade-off: Researchers can observe either fewer neurons at high spatial resolution or more neurons at low resolution. And if the goal is to study unrestricted animal behavior, the probe must also weigh less than 3 g and function over long time periods, a feat that’s proven difficult.

Matteo Carandini of University College London, Timothy Harris of the Howard Hughes Medical Institute campus in Virginia, and their colleagues have now designed a 1.1 g neural probe that can distinguish individual neurons over a 10 mm plane. Their device has a density of recording sites that is an order of magnitude higher than other neural probes. That low weight and high resolution is possible in part because data transmission channels are shared by up to 12 electrodes at a time.

The researchers implanted the probes in 21 rats and mice and successfully recorded neuron signals for months. The image shows data taken across a mouse’s cortex, hippocampus, and thalamus in four sequential recordings (different colors) of about two seconds. Each colored dot represents a signal spike measured at a position on the probe (vertical axis) at a specific time (horizontal axis).

To get long-term stable measurements, the researchers designed an algorithm to correct for the probe’s motion relative to the brain by tracking and position-correcting a neural fingerprint—that is, the distinctive response in the visual cortex to a set of images. (N. A. Steinmetz et al., Science 372, eabf4588, 2021 .)