Unfolding the Hippocampus | Hacker News Reader

Unfolding the Hippocampus

news.westernu.ca

54 pointsrbanffy5 comments

A little of topic, but does anyone know of an introduction to working with MRI and neuro-imaging?

Getting basic information from things like DICOM files is easy enough, but then the neuro-science and image processing stuff starts and basically every tool assumes (rightfully so) that you know the correct terminology to use them.

There are a few US based universities that have online courses and training, for free, but those are typically not on demand and all seat are take on the day registration opens.

I'm just looking something that will get me started on learning correct terminology and basic image processing.

It's because these tools are all built assuming you're a neuroscience researcher with access to an MRI machine. CONN is a great tool and has some tutorials https://web.conn-toolbox.org/tutorials but you really need to sit down with an intro to neuroscience textbook to learn the terminology. I don't think there's an easier way to get up to speed unfortunately.

Also the Human Brain Project has a ton of fMRI data you can download and play with, as well as a few brain atlases that are incredibly useful https://www.humanbrainproject.eu/en/explore-the-brain/ but to learn the terminology, you're still best off buying an intro to neuroscience textbook

It would be interesting to use this technique to:

1. Map the topology of the brain (what the researchers have done here)

2. Run an MRI with IV contrast

3. View the MRI "unwrapped" according to the topology

I suspect the topology can be composed with other current methods that will make it much easier to identify regions of interest.

Simultaneously and in parallel, is anyone MRI-scanning hundred of generations a year of nematodes and fruit flies to build basic physical correlations in neurology and genetics? Structurally, much simpler and cheaper, biochemically very similar.

They're much, much too small for MRI[0], but we do know a crazy amount about the structure of C. elegans: there's a complete wiring diagram of all the physical connections between its 302 neurons, a fate map of the embryonic origin and final role of every single cell, and so on.

[0] Each voxel (='volume element') in a "standard" MRI is cube 0.5-1mm on a side. A C. elegans is about a millimeter long, but only 100 µm wide, so you could fit about a hundred into one voxel. The resulting scan therefore wouldn't tell you much spatially, though I suppose you could do spectroscopy if you wanted to know what a worm was made of.

Reply on news.ycombinator.com