Dear Alfonso and forum,
I am trying to run dFC for event-related design with two conditions A and B (group of 16 healthy participants). These two conditions came from different runs (e.g., ABABABAB). I have 4 ROIs where i would like to look at differences in dynamic connectivity between A and B across ROIs. I have 480 scans for each condition. Because of my task is event-related, I am in doubt how to run the analysis using sliding window approach:
1. in Setup, shall i add the sliding-window temporal decomposition to the A and B conditions?
2. In sliding-window onsets i enter onsets for each condition. What shall i enter into sliding-window length? Obviously, with event-related design i will have different numbers of onsets in the specified sliding window. How this may affect the results?

There is one question more regarding the dynamic factors. What is a dynamic factor? It is derived from my data, but how can interpret the factor? I am trying to look at the literature, but there is no obvious explanation. Could you suggest me any paper to look at?
thank you in advance
alla

Dear Alfonso and forum,

I would really appreciate too further information regarding the methods of Dynamic FC that are used in conn. Is there any reference paper or other more extensive manual from which we could base on?

Thank you in advance for your help and information.

Katherine

Dear Alla & Katherine,

We are working on both a methods paper and a more extensive manual for the new Dynamic FC option in CONN. This option is still under development and some of the method details may change by the time the full documentation and manuscript are released, but you are welcome to try it out and let us know any specific questions and/or feedback/suggestions you may have. Briefly, the method is an extension of PPI (PsychoPhysiological Interaction model) that automatically estimates the most explanatory "psychological factors" from the data. The procedure returns:

      a) a number of temporal modulation factors /timeseries (characterizing the largest dynamic changes observed in a given ROI-to-ROI connectivity matrix)
and b) the ROI-to-ROI loading matrices (interaction terms in PPI) associated with each factor

Jointly these two factors characterize the largest observed dynamic changes in ROI-to-ROI connectivity across the length of your sessions/runs. You can then use second-level analyses to explore potential within- or between- subject differences in any of these factors (the factors in (b) can be analyzed under the 'factor loadings' subheading in the second-level dynFC results tab, and some simple properties of the timeseries in (a) -i.e. average, variability, rate of change- can be analyzed under the 'factor scores' subheading in the second-level dynFC results tab).

Hope this helps and let me know if you would like me to further clarify any of this and/or any thoughts/comments/suggestions you may have (and I will add a note here as soon as additional documentation/references regarding this method become available).

Best
Alfonso

Dear Alfonso,

I am working in resting-state analysis for static and dynamic FC. I wanted to know if it is possible to work for Dynamic and ROI-voxel/ROI analysis in the same project, or if I should have one project separated for each type of analysis. What is the best way?
I've been playing with both analysis in the same project, but it looks like everything is mixed!
Additionally, could you please introduce me with the setup configuration for Dynamic FC analysis with sliding windows? 

Thank you, I am looking forward to your reply,

Best,

Katherine Baquero

Dear Alfonso,

I've started processing fMRI data with dynamic FC button in CONN. But it's not clear for me what is psychological factors and how they are calculated? Could you explain it for me in math terms (formulas or smth else)? Your reply to previous question is not quite clear for me.

Thank you in advance,
Victoria.

Dear Alfonso,

I wrote to you earlier question about the factors in the Dynamic FC.
But it's not clear for me what is psychological factors and how they are calculated? Could you provide a link to an explanatory article, for example?

Thank you in advance,
Victoria.

Hi,
i would like to understand how dynamic factors are calculated and how we can interpret them, in fact i have with two set of data (one resting state and one task related) and when looking at the section 'Explore factors' of the Dynamic FC, there are difference that i am not able to understand (CONN v 15g).
If I look at the time course of Factors scores in my task-related data, i observe an interruption of the plot, the gap correspond to the 'resting' blocks, that occurs in all subjects (here subjects= runs) at the same time, the resting blocks have the same onsets, while other conditions are randomized. Is it normal or is it revealing a problem in my data? onsets or else?

thank you,

Sherihane

p.s you can find attached the screen shot of the section Explore factor

another question if i want my baseline to be explicit , my fixation point blocks, do i have to name it ' rest' or shall i do smthg else?
thank you

Hi Alfonso, 

I am interested in carrying out dynamic FC analyses too. In your previous message, you mentioned a methods paper and a more extensive manual being underway for the dynamic FC option. I was wondering whether they have been released yet as I would appreciate any documentation that could help me interpret the findings.

Many thanks

Originally posted by msc_22: Hi Alfonso,

I would also be really interested in getting further documentation about Dyn ICA in Conn 19, it is an amazing tool.

Cheers,

I am also doing dynamic ICA analysis too. I am very interested in some suport to interpret the results and understand why the dynamic connectivity matrix are asymetric. 
Any support to understand this results will be very helpfull