help
help > RE: Average within / between network connectivity
Apr 28, 2018 03:04 PM | Alfonso Nieto-Castanon - Boston University
RE: Average within / between network connectivity
Hi Himanshu,
Sorry for the late reply. Yes, that significant result can be directly reported and it means that there is a group effect in the between-networks connectivity (i.e. the average connectivity between the two networks is different among the groups). In addition, after getting a result like this, typically you would want to follow it up performing post-hoc analyses to better characterize those between-group differences (e.g. by performing a series of two-sample t-tests between pairs of groups; or you could plot the individual average between-network connectivity values for each group -e.g. by loading the "between-networks" column in the csv file as a new second-level covariate in conn so that you can use the tools.calculator gui there to plot those between-network connectivity values)
Hope this helps
Alfonso
Originally posted by Himanshu Joshi:
Sorry for the late reply. Yes, that significant result can be directly reported and it means that there is a group effect in the between-networks connectivity (i.e. the average connectivity between the two networks is different among the groups). In addition, after getting a result like this, typically you would want to follow it up performing post-hoc analyses to better characterize those between-group differences (e.g. by performing a series of two-sample t-tests between pairs of groups; or you could plot the individual average between-network connectivity values for each group -e.g. by loading the "between-networks" column in the csv file as a new second-level covariate in conn so that you can use the tools.calculator gui there to plot those between-network connectivity values)
Hope this helps
Alfonso
Originally posted by Himanshu Joshi:
Hi Alfonso,
Thankyou for this wonderful script (conn_withinbeteenROItest). I have also used this script for exploring within and between network connectivity for three resting state networks; two networks at a time. I had four groups in my study and as per suggestions, using CONN's gui second-level results tab, ANOVA was implemented for between subjects contrast. After specifying the ANOVA model [1 -1 0 0; 0 1 -1 0; 0 0 1 -1] and selecting three networks in second level results tab, I ran the script "conn_withinbetweeROItest". This promted me to select, two networks which further resulted in
Within matrix Set-1 (4 x 4 ROIs; 12 valid ROI pairs; 99 subjects);
Within matrix Set-2 (7 x 7 ROIs; 42 valid ROI pairs; 99 subjects);
Between matrix (4 x 7 ROIs; 28 valid ROI pairs; 99 subjects;
Set-1 Within-network test (average effect between Set-1 ROIs):F(3 90 )=0.79 (p=0.505040)
Set-2 Within-network test (average effect between Set-2 ROIs): F(3 90 )=1.77 (p=0.159140)
Between-networks test (average effect between Set-1 & Set-2 ROIs):F(3 90 )=4.33; (p=0.006720).
The p-value for between network test was then converted into two-tailed p value: 2*min(p,1-p)= 0.0134.
Since the model specified here is ANOVA model, I was wondering, if we need to perform post-hoc tests. Or two tailed p value of 0.0134, which is significant, can be reported.
Thanks and Regards
Himanshu Joshi
Thankyou for this wonderful script (conn_withinbeteenROItest). I have also used this script for exploring within and between network connectivity for three resting state networks; two networks at a time. I had four groups in my study and as per suggestions, using CONN's gui second-level results tab, ANOVA was implemented for between subjects contrast. After specifying the ANOVA model [1 -1 0 0; 0 1 -1 0; 0 0 1 -1] and selecting three networks in second level results tab, I ran the script "conn_withinbetweeROItest". This promted me to select, two networks which further resulted in
Within matrix Set-1 (4 x 4 ROIs; 12 valid ROI pairs; 99 subjects);
Within matrix Set-2 (7 x 7 ROIs; 42 valid ROI pairs; 99 subjects);
Between matrix (4 x 7 ROIs; 28 valid ROI pairs; 99 subjects;
Set-1 Within-network test (average effect between Set-1 ROIs):F(3 90 )=0.79 (p=0.505040)
Set-2 Within-network test (average effect between Set-2 ROIs): F(3 90 )=1.77 (p=0.159140)
Between-networks test (average effect between Set-1 & Set-2 ROIs):F(3 90 )=4.33; (p=0.006720).
The p-value for between network test was then converted into two-tailed p value: 2*min(p,1-p)= 0.0134.
Since the model specified here is ANOVA model, I was wondering, if we need to perform post-hoc tests. Or two tailed p value of 0.0134, which is significant, can be reported.
Thanks and Regards
Himanshu Joshi
Threaded View
Title | Author | Date |
---|---|---|
Ruth Shaffer | Apr 27, 2016 | |
Alfonso Nieto-Castanon | May 4, 2016 | |
Chabran Eléna | Sep 20, 2018 | |
Maria Beser | Feb 11, 2022 | |
max345 | Feb 1, 2022 | |
Alfonso Nieto-Castanon | Feb 3, 2022 | |
max345 | Feb 4, 2022 | |
Alfonso Nieto-Castanon | Feb 4, 2022 | |
max345 | Feb 9, 2022 | |
max345 | Feb 5, 2022 | |
Himanshu Joshi | Feb 16, 2018 | |
Alfonso Nieto-Castanon | Apr 28, 2018 | |
Himanshu Joshi | Apr 9, 2018 | |
Himanshu Joshi | Apr 28, 2018 | |
Himanshu Joshi | Mar 13, 2018 | |
Himanshu Joshi | Feb 26, 2018 | |
Jon Dudley | May 4, 2016 | |
Alfonso Nieto-Castanon | May 5, 2016 | |
Helene Veenstra | Feb 8, 2018 | |
Alfonso Nieto-Castanon | Feb 14, 2018 | |
Helene Veenstra | Feb 15, 2018 | |
Stephen L. | Feb 8, 2018 | |
Helene Veenstra | Feb 12, 2018 | |
Mariah DeSerisy | Apr 30, 2020 | |
Elizabeth Olson | May 16, 2020 | |
Alfonso Nieto-Castanon | May 19, 2020 | |
Nadine Smith | Feb 23, 2021 | |
ryanw412 | Dec 16, 2021 | |
Jose Maximo | Apr 28, 2021 | |
Olivier Roy | Oct 15, 2021 | |
Karli Carston | Nov 9, 2021 | |