help
help > RE: second-level analysis in Conn
Oct 16, 2015 04:10 PM | Alfonso Nieto-Castanon - Boston University
RE: second-level analysis in Conn
Hi Diana,
Could you also please double-check that the Group2 and Score2 variables are correctly defined? i.e. please check if there are any non-zero values in the Score2 variable for any subject who is not included in group 2 (for whom the Group2 variable has a zero value). One simple way to check this would be to create a new second-level covariate and enter in the 'values' field "~Group2 & Score2" (without the quotes and changing Group2 and Score2 to your actual second-level covariate names). If that new covariate contains any 1's then those subjects are not within group2 (they have a zero value in the Group2 variable) yet they have a non-zero Score2 value. That could be a potential reason why Calculator would not identify your model as a proper regression equation and would refuse to show you the associated R2 values. If that does not seem to be the issue please send me your conn*.mat file and I will take a quick look to see if I can figure out the source of this issue.
Thanks
Alfonso
Originally posted by Diana Parvinchi:
Could you also please double-check that the Group2 and Score2 variables are correctly defined? i.e. please check if there are any non-zero values in the Score2 variable for any subject who is not included in group 2 (for whom the Group2 variable has a zero value). One simple way to check this would be to create a new second-level covariate and enter in the 'values' field "~Group2 & Score2" (without the quotes and changing Group2 and Score2 to your actual second-level covariate names). If that new covariate contains any 1's then those subjects are not within group2 (they have a zero value in the Group2 variable) yet they have a non-zero Score2 value. That could be a potential reason why Calculator would not identify your model as a proper regression equation and would refuse to show you the associated R2 values. If that does not seem to be the issue please send me your conn*.mat file and I will take a quick look to see if I can figure out the source of this issue.
Thanks
Alfonso
Originally posted by Diana Parvinchi:
Hi Alfonso,
Yes, in the Calculator display, I select my group 2 & score 2 from the predictor variables, enter the contrast [0 1] and select my FC variable as my outcome variable. Thanks very much for looking into this! Looking forward to hearing back from you.
Best,
Diana.
Originally posted by Alfonso Nieto-Castanon:
Yes, in the Calculator display, I select my group 2 & score 2 from the predictor variables, enter the contrast [0 1] and select my FC variable as my outcome variable. Thanks very much for looking into this! Looking forward to hearing back from you.
Best,
Diana.
Originally posted by Alfonso Nieto-Castanon:
Hi
Diana,
That is curious, the Calculator display will attempt to somehow "smartly" only display R2 values when they are relevant (i.e. when you have a regression-like equation) so perhaps the logic behind that code is not as "smart" as intended. I will double-check the code to see if I can figure out what might be causing this behavior; in the meantime could you please confirm that for your Group2 analyses you are selecting "Group2" and "Score2" as your predictor variables (contrast [0 1]) and your FC variable as your outcome variable?
Thanks
Alfonso
Originally posted by Diana Parvinchi:
That is curious, the Calculator display will attempt to somehow "smartly" only display R2 values when they are relevant (i.e. when you have a regression-like equation) so perhaps the logic behind that code is not as "smart" as intended. I will double-check the code to see if I can figure out what might be causing this behavior; in the meantime could you please confirm that for your Group2 analyses you are selecting "Group2" and "Score2" as your predictor variables (contrast [0 1]) and your FC variable as your outcome variable?
Thanks
Alfonso
Originally posted by Diana Parvinchi:
Hi
Alfonso,
Thank you very much for writing back. In the second level results window, I select my groups (group 1, group 2, group 3) and their scores (score1, score2, score3), enter the contrast [0 0 0 1 0 0;0 0 0 0 1 0; 0 0 0 0 0 1], select all of our ROIs (36) and leave the default eye (n) in the between sources contrast (this is to investigate association between FC and symptom severity in three age groups of cohorts with ASD). I don't enter "all subjects" in the Subject effects list, is that ok? I also looked in Setup.CovariatesSecondLevel and did not see NaN values for other subjects. Why do you think the R2 is missing in the calculator window for group 2 only (we get stats and the regression graph, but not R2 for this group)? Many many thanks for your help.
Best,
Diana.
Thank you very much for writing back. In the second level results window, I select my groups (group 1, group 2, group 3) and their scores (score1, score2, score3), enter the contrast [0 0 0 1 0 0;0 0 0 0 1 0; 0 0 0 0 0 1], select all of our ROIs (36) and leave the default eye (n) in the between sources contrast (this is to investigate association between FC and symptom severity in three age groups of cohorts with ASD). I don't enter "all subjects" in the Subject effects list, is that ok? I also looked in Setup.CovariatesSecondLevel and did not see NaN values for other subjects. Why do you think the R2 is missing in the calculator window for group 2 only (we get stats and the regression graph, but not R2 for this group)? Many many thanks for your help.
Best,
Diana.