I lol'd at this:
"Unlike with large commercial devices, such as MRI scanners, appropriate use of these tools requires in-depth knowledge of many technical details."
fMRI skeptics: "Oh my goodness, they are doing all the stats wrong". Also fMRI skeptics: "No knowledge required!"
Right?! I appreciate that (f)MRI might *seem* more plug-and-play to an electrophysiologist, but to actually push scientific boundaries and get the highest quality data, you really need to know what you're doing -- on the data acquisition side, but often even more importantly, on the analysis side.
I missed that line! @jvoigts.bsky.social MRI systems might seem plug&play, but every step from study design to acquisition to analysis benefits from professional support. An MRI machine without regular professional QC & maintenance will give bad data in subtle and obvious ways.
I was debating whether that was a good example or not. I agree with what you're saying, the major point I wanted to make was that there's a huge amount of complexity that's been packaged into a machine that a well trained non expert can use. This is not the case in many systems neuro experiments.
MRI is also a great success story of providing access to technical expertise, support, and amazing standardized analysis pipelines. Imo we can learn from this in other areas. Of course boundary pushing work (and analysis/stats) always requires depth on understanding how the sausage is made!
Thanks for your response @jvoigts.bsky.social (and sorry for not tagging you in before!). Bc other subfields of neuro can sometimes be disparaging of (f)MRI, comments like this are a bit triggering ;) but I appreciate your clarification, and also your larger point -- science needs more specialists!
Comments
"Unlike with large commercial devices, such as MRI scanners, appropriate use of these tools requires in-depth knowledge of many technical details."
fMRI skeptics: "Oh my goodness, they are doing all the stats wrong". Also fMRI skeptics: "No knowledge required!"