I'm planning to analyze a dataset consisting of fear acquisition and extinction. I was wondering if you have any experience in applying the model based method on fear extinction. In my experiment, there are 25 trials in acquisition and 22 trials in extinction. The CS are presented relatively long (3.5 s) so I was planning on using DCM. Do you have any comments on extinction? Would I just go ahead and model the omitted US as if it was acquisition?
Thank you,
Jon
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Another question I had was about acquisiton. I tried DCM in a few of my participants for the acquisition phase of the fear learning experiment and consistently found a considerable spike of at least 2 orders of magnitude in the flexible response amplitude for the first trial in each subject. Note that the first trial was a CS-. I cannot really figure out why I am getting this--could you help me out here?
Thank you so much.
Jon
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm happy to help. First, your plan of modelling the extinction experiment appears reasonable. I would model the US response for all trials to capture possible US omission responses in the CS+ trials (but you also have to do this for the CS- trials so as not to bias the model).
Regarding the "spike" on the first trial, I assume that this is a large CS response on the first trial? You could check this by the eye - is the response really there or is the DCM picking up something else (i. e. something that occurs before the first trial). Your data should ideally start at least a few seconds before the first trial. If the response is really there, maybe the first CS is just very salient - perhaps you could remove such effects by giving a "warning" a few seconds before the first trial (something like a screen that says "starting soon ...")
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:
Hi,
I'm planning to analyze a dataset consisting of fear acquisition and extinction. I was wondering if you have any experience in applying the model based method on fear extinction. In my experiment, there are 25 trials in acquisition and 22 trials in extinction. The CS are presented relatively long (3.5 s) so I was planning on using DCM. Do you have any comments on extinction? Would I just go ahead and model the omitted US as if it was acquisition?
Thank you,
Jon
Another question I had was about acquisiton. I tried DCM in a few of my participants for the acquisition phase of the fear learning experiment and consistently found a considerable spike of at least 2 orders of magnitude in the flexible response amplitude for the first trial in each subject. Note that the first trial was a CS-. I cannot really figure out why I am getting this--could you help me out here?
Thank you so much.
Jon
Hi Jon
I'm happy to help. First, your plan of modelling the extinction experiment appears reasonable. I would model the US response for all trials to capture possible US omission responses in the CS+ trials (but you also have to do this for the CS- trials so as not to bias the model).
Regarding the "spike" on the first trial, I assume that this is a large CS response on the first trial? You could check this by the eye - is the response really there or is the DCM picking up something else (i. e. something that occurs before the first trial). Your data should ideally start at least a few seconds before the first trial. If the response is really there, maybe the first CS is just very salient - perhaps you could remove such effects by giving a "warning" a few seconds before the first trial (something like a screen that says "starting soon ...")
Hope this helps
Best wishes
Dominik