Deep Dive into a Cubic Millimeter of Brain Matter

This piece, by Onno Berkan, was published on 09/11/24. The original text, by Cariss Wong, was published by Nature on 05/09/24.

One cubic millimeter is about as large as a grain of pink Himalayan salt, about a millionth of the brain. Contained in there is around 57,000 cells and 150 million synapses, and about 1.4 petabytes of data, as found by a Google study. Further examination unearthed neurons that made up to 50 connections with each other, as well as “neurons with tendrils that formed knots around themselves”, writes Wong. Nobody had seen anything like this before.

This bite-sized portion of brain matter was cut into 5,000 slices and imaged by an electron microscope at Harvard. Those images were then put back together by the Google team’s AI. It was this reconstruction that led to the neurons that synapse onto each other fifty times, the neurons with tendrils that formed knots around themselves, and even to neurons that seemed to be “near-perfect mirror images of each other”, as Wong describes it. Of course, we have no clue why any of these fascinating features happen, or what purpose they serve. It is also important to note that this is a piece of the brain the size of a grain of salt.

Additionally, we don’t even know if these features are actually real. That miniscule piece created such a gigantic map that only a few per cent of the neurons have been manually checked– the rest remains in need of proofreading. Until we can fully check for errors caused by staining or just piecing the images back together, we cannot be sure of the map’s accuracy. 

This is a difficult process, and mapping out the whole brain may be a little ways away, but scientists remain optimistic. Viren Jain, a co-author on the central paper for this story, is a part of the AI-driven brain research movement, which has been gaining momentum. Using machine learning to simulate and model brain activity, to create better maps of the brain, and processing large amounts of data could have a large stake in the future of neuroscience research, and I’m all here for it.

If you’re into connectomics, the USC Center for Integrative Connectomics (CIC) “seeks to develop a multimodal multiscale connectome and cell-type map of the mammalian brain using advanced tracing, imaging, and computational methods.” Go check them out! 

Thank you for reading. Reminder: Byte Sized is open to everyone! Feel free to submit your piece.

All submissions to berkan@usc.edu with the header “Byte Sized Submission” please. Thank you!