Mapping the human brainstem has long been a "black box" for medical imaging due to its small size and constant interference from physical movements like breathing and heartbeats.

A collaborative team from MIT, Harvard, and MGH has introduced the BrainStem Bundle Tool (BSBT), an AI-powered software that automatically segments eight distinct nerve fiber bundles. By using a convolutional neural network—a type of AI modeled after the visual cortex—the tool processes diffusion MRI scans to track the movement of water along myelinated axons (nerve fibers) to create a high-resolution map of neural connectivity.

This precision allows clinicians to see how white matter, the brain’s communication cables, degrades during neurodegenerative diseases. In Parkinson’s and multiple sclerosis patients, BSBT identified specific patterns of volume loss and structural integrity changes that were previously invisible to doctors.

Beyond diagnosis, the tool offers significant prognostic hope; it successfully tracked the physical shifting and healing of nerve bundles in a traumatic brain injury patient during a seven-month coma. By making this tool public, researchers aim to turn the brainstem into a readable map for treating everything from sleep disorders to paralysis.

Mapping the human brainstem has long been a "black box" for medical imaging due to its small size and constant interference from physical movements like breathing and heartbeats.

A collaborative team from MIT, Harvard, and MGH has introduced the BrainStem Bundle Tool (BSBT), an AI-powered software that automatically segments eight distinct nerve fiber bundles. By using a convolutional neural network—a type of AI modeled after the visual cortex—the tool processes diffusion MRI scans to track the movement of water along myelinated axons (nerve fibers) to create a high-resolution map of neural connectivity.

This precision allows clinicians to see how white matter, the brain’s communication cables, degrades during neurodegenerative diseases. In Parkinson’s and multiple sclerosis patients, BSBT identified specific patterns of volume loss and structural integrity changes that were previously invisible to doctors.

Beyond diagnosis, the tool offers significant prognostic hope; it successfully tracked the physical shifting and healing of nerve bundles in a traumatic brain injury patient during a seven-month coma. By making this tool public, researchers aim to turn the brainstem into a readable map for treating everything from sleep disorders to paralysis.