I'm considering using PsPM for analysing skin conductance data but I'm unsure whether it is suitable for the stimuli used in this study. I have had a look at the PsPM manual and noticed that the models seem to be suited for short stimuli (<1s) or for longer stimuli of unspecified duration. However, stimuli (emotional pictures) in this study had a fixed duration of 5 seconds. I hope you can help me.
Thanks a lot,
Catarina
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
it seems that the non-linear models ("DCM") are excellently suited to analyse your experiment.
Linear models (GLM) assume short sudomotor firing with known latency, usually elicited by short stimuli. This might not be the case with a 5 s stimulus.
Non-linear models allow you to specify time windows during which sudomotor firing is assumed to occur - e. g. during the 5 s stimulus. These models were developed for fear conditioning where a CS extends in time - and this is absolutely comparable to your experiment with an extended picture presentation.
Tips:
- if you assume just one response, model the entire time window.
- if you think there may be more than one response during this time, you can either
(a) model the entire time window twice - this is very flexible but may mean that the estimated responses overlap strongly, so you may have to combine them somehow
(b) split your time window into two halves which you model separately - then the responses are forced to occur on one or the other half
- you don't need to account for the conduction delay, what you are modelling is the central (brain) sympathetic impulse, going out to the sweat glands
Hope this helps
Best wishes
Dominik
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Dominik,
I'm considering using PsPM for analysing skin conductance data but I'm unsure whether it is suitable for the stimuli used in this study. I have had a look at the PsPM manual and noticed that the models seem to be suited for short stimuli (<1s) or for longer stimuli of unspecified duration. However, stimuli (emotional pictures) in this study had a fixed duration of 5 seconds. I hope you can help me.
Thanks a lot,
Catarina
Dear Catarina
it seems that the non-linear models ("DCM") are excellently suited to analyse your experiment.
Linear models (GLM) assume short sudomotor firing with known latency, usually elicited by short stimuli. This might not be the case with a 5 s stimulus.
Non-linear models allow you to specify time windows during which sudomotor firing is assumed to occur - e. g. during the 5 s stimulus. These models were developed for fear conditioning where a CS extends in time - and this is absolutely comparable to your experiment with an extended picture presentation.
Tips:
- if you assume just one response, model the entire time window.
- if you think there may be more than one response during this time, you can either
(a) model the entire time window twice - this is very flexible but may mean that the estimated responses overlap strongly, so you may have to combine them somehow
(b) split your time window into two halves which you model separately - then the responses are forced to occur on one or the other half
- you don't need to account for the conduction delay, what you are modelling is the central (brain) sympathetic impulse, going out to the sweat glands
Hope this helps
Best wishes
Dominik