Harvard University and Google have achieved a groundbreaking milestone in brain science by creating an incredibly detailed map of human brain connections. This remarkable feat focuses on a tiny cubic millimeter of the cerebral cortex, obtained from a patient during epilepsy surgery in 2014. Over a span of ten years, a dedicated team of biologists and machine-learning experts meticulously analyzed this small tissue sample, which contains around 57,000 cells and 150 million synapses. The work represents a significant advancement in understanding the brain’s intricate wiring.
The process began with staining the brain tissue using heavy metals, which bind to lipid membranes in cells, allowing them to be visible under an electron microscope. The tissue was then embedded in resin and sliced into extremely thin sections, each measuring just 34 nanometers thick. By converting the complex 3D problem into a more manageable 2D one, the team generated a massive dataset of 1.4 petabytes. To reconstruct these 2D slices into a unified 3D model, they utilized state-of-the-art machine-learning algorithms in collaboration with Google. This involved aligning the images and automatically segmenting the various cell types, with some manual adjustments made to enhance the accuracy of these segments.
The resulting brain map unveils a wealth of information regarding the cellular structure of the brain. It has identified neurons with more than 50 synapses, a detail that was previously overlooked but could be crucial for understanding cortical processing. Despite the project’s achievements, there are still challenges ahead, such as verifying the vast amount of data manually to correct segmentation errors. Certain cells, like unidentifiable egg-shaped structures and tangled cells, present intriguing mysteries that warrant further exploration.
The brain map is now accessible to the public, opening up new possibilities for research. It holds the potential to advance our comprehension of mental health disorders like schizophrenia and could potentially inspire advancements in artificial intelligence by mimicking brain functions. Looking ahead, future projects include expanding this research to cover entire mouse brains and additional regions of the human brain. These endeavors could lead to further breakthroughs in neuroscience and related fields.
