Getting deep into the brain

June 29, 2017

As we try to uncover how many cell types are in the brain, researchers are recognizing the importance of characterizing neurons by several features: their location, shape, gene expression and how they respond to electrical signals.

For neurons on the surface of the brain, there are powerful new tools to collect this information from just one cell at a time. But what happens when we need to learn about cells deeper in the brain—ones we cannot see right on the surface?


“Probing neurons past the surface of the brain is challenging even for well-trained humans who need to collect data from only a single cell and not damage cells around the target cell,” says Lu Li, Ph.D., Senior Scientist at the Allen Institute for Brain Science. “This makes large-scale collection of this data incredibly hard. Our goal was to create a robot that could do the work just as well or better than a human.”

One challenge the robot faced was how to find target cells and avoid tissue damage without an operator being able to see neighboring cells on the surface of the brain. Li and his team first developed a method to enable the robot’s probe to passively detect whether cells were firing nearby, and then to actively probe the space for hidden cells.

Once the probe reached its target site, researchers demonstrated that the probe could effectively deliver a molecule that distributed to the far ends of just a single neuron, making tagging the neuron possible for future experiments.

The robot, called the Automatic single-Cell Experimenter (ACE), is described this month in the journal Nature Communications.

“ACE has several distinct advantages over typical optical approaches,” says Li. “First, it can label individual cells deeper in the brain, not just the cortex, and the particular molecule we use to label the neuron’s shape only requires a small amount to visualize the entire cell.”

Since the ultimate goal of the work is to decipher cell types in the brain, this kind of highly specific labeling is the key first step in order to correlate data on how a neuron is shaped with functional data, including how it sends and receives signals in circuits.

“We hope that by creating this technology, we can automate and thereby increase the number of experiments done to isolate single cells, and to make the question of how many cell types are in the brain that much more accessible,” says Li.