Dear Ben,
Some additional reflections:
> (1) Are there methodological/technical issues arguing against the comparison?
It might well be biased. Differences between NEU VALPARA and NEU TRAPARA might be due to simple block effect (fatigue affecting the 2nd experiment to a larger extent), it might be related to the design (three conditions in VALPARA vs. two conditions in TRAPARA, thus different number of time points, different intervals between blocks of the same type, thus different frequency spectra). The second aspect could be eliminated by adding a non-trauma (positive?) condition. The first aspect could be controlled by randomizing the order of the two tasks. But I'm not sure whether this is useful in your case (if subjects start with TRAPARA the VALPARA task might be affected).
> (2) Could the different ratings after each block (arousal/trauma) between the paradigms cause problems? Arousal & trauma will be related but not identical
This is another bias. Contrasting different tasks might result in trivial effects. E.g. for healthy controls none of the stimuli should (?) be trauma-inducing, thus it might be easier to judge the stimuli during that task (no idea, maybe it's the opposite im case the task is to imagine whether the stimuli *could be* TRA-inducing). The two tasks might also interact with the conditions, as the tasks might be more similar for the neutral condition (not arousing, not TRA-inducing = nothing much to think about), alternatively it might be more demanding to judge the arousal of neutral stimuli vs. the TRA-related effects (maybe some NEU stimuli are arousing to some extent but not at all TRA-inducing).
In general, as you are interested in the arousal ratings / the task is not just to have people think about something, what about valence? And what about arousal/valence ratings for the TRAPARA? I'm definitely not into trauma literature, but if you think TRA affects NEU, wouldn't it be interesting to know whether this results in e.g. a more extreme arousal / more negative valence rating of the neutral stimuli?
> (3) Would you recommend to acquire both paradigms within one session or two sessions (basically stop & re-start the scanner)?
It's never bad to have a long session, you can still split it afterwards. In your case I'm not sure whether it is really useful to go with a long session though. In case the single-subject models cover both the tasks the high-pass filter might result in funny effects, as e.g. the VALPARA conditions might not just be "off" during the TRAPARA task but show some "ripples". And contrasting the NEU blocks might be ineffective when put into a single model, as it would be contrasting (parts of) the first half or so vs. (parts of) the second half. Corresponding effects could be affected by scanner drifts and/or true effects might be filtered depending on HPF setting.
Now, having said that the current design is suboptimal in some aspects, I'm not sure how to properly solve the issue. It would be good to have another condition in which you're absolutely sure there are no differences between tasks. Again, I don't know whether it's possible at all. Thinking about having to judge trauma effects vs. arousal might already result in differences in activation well before you present the stimuli, especially in patients. On the other hand, maybe these are trivial effects you're not interested in and you rather assume that TRA induces some effects which affect the processing of stimuli but not e.g. fixation periods. If that's not the case it might be worth a try to jitter the intervals between TRA and NEU blocks (sort of "after 15 s there's still some effect but not after 30 s"). In case TRA blocks have a rather strong effect then this could also cumulate over time, resulting in some regions activated all the time from a certain point in time onward.
Hope this helps a little,
Helmut
|