FBA asymmetric analysis left verus right hemisphere

1

Hi Mrtrix3 users:

I am interested in doing left vs right hemispheric FBA on an individual subject and running group-level analysis afterwards.
I have already performed FBA on FD, FDC, and log(FC) according to the https://mrtrix.readthedocs.io/en/latest/fixel_based_analysis/mt_fibre_density_cross-section.html
, but not able to find significant difference on group-level .

Any suggestions would be appreciated.

Thanks.

Ping

2

Hi Ping,

Performing a fixel-based analysis of brain asymmetries requires a bit of deviation from the standard FBA pipeline. We will be submitting a manuscript on such an analysis imminently, which will describe the requisite steps in greater detail.

Rob

3

Hi @pinghongyeh

Not sure if the latter (unable to find significant results) refers to the former (FBA on an individual subject), but if so, this is probably entirely expected. Performing FBA on an individual subjects is not very likely to have sufficient power to yield any results at all…

Maybe that’s the way to go, but not afterwards, but out of the box. Any reasons why you’d first run FBA on an individual and then on a group (based on that?). Why not simply a group-level analysis? In an case, starting point of the latter is flipping the brains. Then think a bit about what that means, and if there’s anything you’d want to control for in the stats (not per se; depends on what you’re after). If you’re stuck on anything specifically, feel free to ask in more detail. Your current question is a bit open-ended… :wink: . Take into account that not finding a significant result is just what it is most of the time: there might simply not be an effect, or you might lack the power (e.g. number of subjects, etc…) to identify it as being significant. If power is an issue, think if you might have more specific hypothesis (e.g. an anatomical structure, etc…), so you can focus on that, yet with more statistical power; e.g. by defining some ROIs, etc… :slightly_smiling_face:

Cheers,
Thijs