E11 Bio is excited to share a major step towards brain mapping at 100x lower cost, making whole-brain connectomics at human & mouse scale feasible (🧠→🔬→💻). Critical for curing brain disorders, building human-like & aligned AI systems, and even simulating human brains.
E11 Bio’s mission is to unlock connectomics at the scale of whole mammalian brains, first for mice, then for humans. We want to make connectomics radically cheaper and more accessible: here’s how we’re doing it: https://lnkd.in/gAvHDbUE
The big challenge for scaling up connectomics is proofreading the output of the AI models used to automatically reconstruct the connectome from large brain image datasets. These AI models are good, but still make many errors which must be manually fixed by human annotators, which makes up 95% of costs.
AI will improve, but it will need more and better data to do so. Our PRISM technology addresses this bottleneck by making the neurons “self-annotating.” Each cell makes a different combination of proteins which function as a cellular ID (barcode), giving neurons the potential to proofread themselves. We acquire image datasets of these cells using expansion microscopy and protein staining, and then analyze them using “self-proofreading” AI models we’ve developed to leverage barcodes for reduced human proofreading and lower costs.
We still have much work to do in order to scale up PRISM connectomics to large volumes or an entire mouse brain. But by addressing the biggest cost bottlenecks, an optical mouse connectome could come sooner than expected — we predict it could happen in just five years.
E11 Bio is a non-profit Focused Research Organization. In line with our open science philosophy, we are rapidly sharing these first details ahead of our initial preprint in 2025, which will document a pilot study as well as a beta version of our methods and AI models.
What we plan to do after that is to periodically release increasingly large-scale connectomic data drops and demonstrate increasing performance of our AI models. We are seeking partners in scaled imaging, compute, and data storage as well as automation and instrumentation, and are eager to discuss potential applications of our tools and datasets for AI, biomedicine and other key commercial applications – please reach out if you’re interested!
The hippocampal study was a collaborative effort between E11 Bio and Rosa Park (now at E11 Bio) and Sam Rodriques at the Crick Institute, Joergen Kornfeld at Max Planck/LMB Cambridge UK, and Bobae An, Daniel Leible, and Ed Boyden at MIT and HHMI.
We are incredibly grateful to E11 Bio’s funders including Eric and Wendy Schmidt, James Fickel, and others.
Here’s the link to the roadmap again if you read to the end: https://lnkd.in/gAvHDbUE We hope you will join us in our mission to map the mouse connectome (and beyond!)