As I sit with countless articles with structural brain images, I get suprised how confident some authors sound about their findings, particularly in the fields of memory impairment and hippocampal atrophies. And in using a cross-sectional sample. Always be on the alert with the word cross-sectional data.
When ever I see an MRI (Magnetic Resonance Imaging) image I think two things; 1) my dear lord science is amazing, just the mare physics of this thing is breathtaking! 2) my god haven’t we come longer than this?
With many journalists recaping science inaccuratley, I thought I’d put you on the alert of the struggels we face, to make you dare to question.
Because you see, when ever a friend tells you they read a great cognition study showing relavance through MRI imaging, and they’ve read it in a pop-sci magazine. Think of this.
This is my hand.
This is my hand scanned
The pixels form a pattern. (This would be a typical T1 weighted image)
This is my hand as represented by a structural MRI. The key note being structural.
If you were to say something about the true state of my hand, it’s volume, structure, potential damadgase, would you dare to be so confident.
MRI’s are amazing, and with the right techniques, used with the right research question (and in clinics!!), very reliable. But to draw determenistic conclusion, is not optimal, most of us add such things as biomarkers and pathology (post mortem of course, we rarely get samples of a brain). And we try to think of new ways to image
So the next time your friend decides to flash with knowledge of a study a brain and its cognition flash back by taking out a magic marker, doodeling on their finger and place the finger on their forehead, ask them to look in to the mirror and decide weather their finger is broken. For such is the challanges of a structural MRI image.
And if by this time you haven’t noted and muttered angerly that many of my references are from Wikipedia, put some marker on your own finger, If you have, good work! I’m proud of you already.