I have one question regarding the interpretation of some FBA results using the log_FC quantitative metrics.
Lower magnitudes of fibre-bundle cross-section (FC) is typically found in neurodegenerative diseases, such as Alzheimer’s disease and Motor Neurone Disease, when these clinical populations are compared to healthy ones. To make it short, the common interpretation is that: lower FC = lower capacity of information transfer for the bundle of interest.
My question is:
Is this ‘always’ the case? Also for small interindividual differences on a population of healthy individuals? And can it ‘that’ be the only functional interpretation (pls, don’t hit me )?
In my study, I have noticed that interindividual differences in log_FC are particularly pronounced in the midsagittal line of the genu and splenium of the corpus callosum - the point where fibers from the two hemispheres converge. Here, I found that log_FC relate negatively (negative_ pFWE < 0.05) with auditory performances in an experimental task that might require interhemispheric interaction and communication. So, looking at that ‘interpretation’, my results seem to be a bit counterintuitive.
I’d like to have an opinion from you regarding the question (if not the results of my study, which I guess would harder to interpret). Also rising a doubt on this maybe ‘simplistic’ interpretation, the multiple factors that can affect FC (besides axon loss), or also pointing to relevant literature, could be of help.
To make it short, the common interpretation is that: lower FC = lower capacity of information transfer for the bundle of interest.
That specific interpretation is what we associate with the FDC metric. I would advise against using it in the context of FC - even if only wanting to establish a simple directionality of relationship - as it may lead to confusion. For FC the interpretation should strictly be one of morphology. This is importantly entirely a feature of image registration, and therefore the limitations of such need to be taken into consideration.
In my study, I have noticed that interindividual differences in log_FC are particularly pronounced in the midsagittal line of the genu and splenium of the corpus callosum - the point where fibers from the two hemispheres converge.
There’s a couple of aspects of the wording here that raised my eyebrow, so I need to make sure there’s a common understanding:
Are the differences very specific to the midsagittal line? We expect effects to manifest along the white matter trajectories, and explicitly process the data accordingly. If the effect you’re commenting on is highly specific to the midsagittal line, it may be indicative of some other issue; e.g.:
Describing this location as “where fibres from the two hemispheres converge” is quite unusual. Both biological neuronal fibres and reconstructed streamlines traverse the whole trajectory from one hemisphere into the other; these are not two separate distinct sets that “meet in the middle”. Not sure if it’s just a wording thing, but sometimes clumsy wordings have turned out to be evidence for erroneous processing so I prefer to clarify.
… in an experimental task that might require interhemispheric interaction and communication.
If the task only “might” require inter-hemispheric communication, then it’s difficult to assign any confidence to the expected effect, and therefore whether or not the observations are or are not congruent with such. For instance, is it possible that the task might require strong intra-hemispheric communication more so than inter-hemispheric communication, and therefore sacrifice the latter for the former?
… the multiple factors that can affect FC (besides axon loss)
As mentioned previously, FC is based entirely on registration. As shown in the FBA manuscript, it’s also possible to envisage scenarios where a change can be observed in this metric that does not correspond to information-carrying capacity due to corresponding changes in FD. So the interpretation of any changes in specifically FC, especially if in isolation to any other imaging metrics, needs to be quite conservative.
)?