Inter-subject normalisation using mu

1

Hello,
In my analysis, I have calculated the SC using the “tcksift2” command and subsequently multiplied the streamline weights by a proportionality coefficient (mu). I read this paper titled “Quantitative Streamlines Tractography: Methods and Inter-Subject Normalization,” where different methods for calculating mu have been discussed.
I am confused about which equation is utilized to calculate mu in the “tcksfit2” command. Is the adjusted mu used in this calculation?
If not, please suggest how to calculate the adjusted mu for comparing SC across participants.

Regards
Pratika

2

I still need clarification on the above query.

I will be very thankful for the reply.

3

Hi,

I’m not an expert on this, but I think the mu you obtain from tcksift2 is the one you are looking for. So you can build your connectome with the sift2 weights and then multiply it by mu (basically what you are already doing).

Best regards,

Manuel

1 Like
4

Hey @mblesac

Thank you for the reply. I would like to know if there is a difference between the “mu” provided by the “tcksift2” command and the adjusted “mu.” If they are different, how can I calculate the adjusted “mu” to compare the structural connectivity (SC) across subjects?

Regards
Pratika

5

Dear @Pratika_siwatch,
Although you asked this a long time ago, my understanding is that the adjusted μ (eq. 4 in the paper) is identical to the μ provided by tcksift2 as long as the following three conditions hold:

  1. You used a common response function across all subjects (this could be the average response function in the population, or the response function of a representative subject).
  2. You normalized the intensity of the DWI data (step 13 in this paper).
  3. The voxel size is identical across subjects (this isn’t actually necessary if you’re using SIFT2).

The relevant part that explains this in the paper you referenced is:

“The global intensity normalisation and group average response function components of the recommended pre-processing pipeline for AFD analysis are tailored to make equivalent across subjects the values of DWIref and AFDref, respectively. As such, if this pipeline is followed,
the term AFDref/DWIref is identical across subjects by construction, and simply multiplying the sum of streamline weights within a pathway of interest (e.g. a connectome edge) by μi permits direct quantitative comparison of FBC between subjects”

Hope this helps you and future users,

Roey