[color=#000000]Hi Max,[/color]
[color=#000000]Yes, that is perfectly correct, and the two approaches:[/color]
[color=#000000] a) first computing [/color]newConditionA = ConditionA - ConditionC, and newConditionB = ConditionB - ConditionC; and then entering those newConditionA and newConditionB images as part of your second-level analyses (e.g. entering newConditionASession1 newConditionBSession1 newConditionASession2 and newConditionBSession2 and defining a [1/2 -1/2 -1/2 1/2] interaction analysis)
and b) directly defining your second-level analysis on the original images incorporating that subtraction as part of the contrast vector (e.g. entering newConditionASession1 newConditionBSession1 newConditionCSession1 newConditionASession2 newConditionBSession2 and newConditionCSession2 and defining a [1/4 1/4 -1/2 -1/4 -1/4 1/2] interaction analysis)
[color=#000000]are exactly equivalent. (that said, this is not the same as simply removing/disregarding Condition C, as the above analyses are using Condition C as a common baseline/reference condition across the two sessions, while something like entering ConditionASession1 ConditionBSession1 ConditionASession2 and ConditionBSession2 and defining a [1/2 -1/2 -1/2 1/2] interaction analysis would not)
[/color]
[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Max Shizzle:[/i][quote]Hello Everyone,
for a gPPI analysis, I need to remove a condition (Condition C) since it is only "task control" and not really relevant to the main research question. In a corresponding SPM task-analysis, this condition was subtracted at the first level.
If I two groups and these six conditions are set up at the second-level in CONN:
Group1
Group2
1. CondA - Session1
2. CondB - Session1
3. CondC - Session1
4. CondA - Session2
5. CondB - Session2
6. CondC - Session2
Considering the GROUPxTIME interaction: Is it correct to do a between conditions contrast at the second level in this way to get the same result as when removing condition C in the first level: 1/4 1/4 -1/2 -1/4 -1/4 1/2 (GROUPxTIME minus Condition C)?
Or should this condition be removed at the fist level, like in the corresponding SPM - Task analysis (to potentially make the results more comparable)?
Thank you in advance!![/quote]
[color=#000000]Yes, that is perfectly correct, and the two approaches:[/color]
[color=#000000] a) first computing [/color]newConditionA = ConditionA - ConditionC, and newConditionB = ConditionB - ConditionC; and then entering those newConditionA and newConditionB images as part of your second-level analyses (e.g. entering newConditionASession1 newConditionBSession1 newConditionASession2 and newConditionBSession2 and defining a [1/2 -1/2 -1/2 1/2] interaction analysis)
and b) directly defining your second-level analysis on the original images incorporating that subtraction as part of the contrast vector (e.g. entering newConditionASession1 newConditionBSession1 newConditionCSession1 newConditionASession2 newConditionBSession2 and newConditionCSession2 and defining a [1/4 1/4 -1/2 -1/4 -1/4 1/2] interaction analysis)
[color=#000000]are exactly equivalent. (that said, this is not the same as simply removing/disregarding Condition C, as the above analyses are using Condition C as a common baseline/reference condition across the two sessions, while something like entering ConditionASession1 ConditionBSession1 ConditionASession2 and ConditionBSession2 and defining a [1/2 -1/2 -1/2 1/2] interaction analysis would not)
[/color]
[color=#000000]Best[/color]
[color=#000000]Alfonso[/color]
[i]Originally posted by Max Shizzle:[/i][quote]Hello Everyone,
for a gPPI analysis, I need to remove a condition (Condition C) since it is only "task control" and not really relevant to the main research question. In a corresponding SPM task-analysis, this condition was subtracted at the first level.
If I two groups and these six conditions are set up at the second-level in CONN:
Group1
Group2
1. CondA - Session1
2. CondB - Session1
3. CondC - Session1
4. CondA - Session2
5. CondB - Session2
6. CondC - Session2
Considering the GROUPxTIME interaction: Is it correct to do a between conditions contrast at the second level in this way to get the same result as when removing condition C in the first level: 1/4 1/4 -1/2 -1/4 -1/4 1/2 (GROUPxTIME minus Condition C)?
Or should this condition be removed at the fist level, like in the corresponding SPM - Task analysis (to potentially make the results more comparable)?
Thank you in advance!![/quote]