A reviewer suggested that we reanalyzed our SCR data using the PsPM toolbox, a suggestion for which I'm him or her thankful as I was unaware of this very impressive toolbox. After reading the manual and the posts on the discussion forum, I remain with the following issue. Let me first provide you with some background.
Two brief (duration of 250 ms) stimuli of different intensities were presented randomly each 30 times in total, divided across three blocks (separated by a 3 min break). SCR responses were only recorded from 2s prior to stimulus onset until 8s after stimulus offset (so no continuous recordings). The interstimulus interval varied between 9.5 and 13.5 seconds. In addition, we recorded the perceptual categorization of the stimulus on every trial (two-forced choice task, high vs. low).
I read that PsPM assumes continuous data but that it could also work with data recorded in epochs. Is this indeed the case, especially for SCR? And if so, as you suggested in your reply to a previous post, should I just include a marker that accounts for the start of each epoch in addition for a marker that indicates the start of each block?
Many thanks in advance for your highly appreciated advice.
Best,
Jonas
Jonas Zaman, MSc, PhD
Health Psychology
Centre for the Psychology of Learning and Experimental Psychopathology
Psychology and Pedagogic Sciences
KU Leuven
Room 02.09 (PSI)
Tel: +32 16 3 25840
in principle this kind of analysis is possible, and given the 8 s post-stimulus interval, also very meaningful.
The power of the PsPM approach particularly for SCR is (1) to use the full data rather than just a peak, and (2) to account for previous responses that overlap with the current one. They will still overlap with a 13.5 s ITI.
To fully exploit point (2), I would recommend importing the data in continuous time, but setting the non-recorded segments to NaN. In other words, you work out when in time did the recordings happen, and fill the gaps with NaNs.
PsPM will then be able to account for overlapping responses by creating a design matrix in continuous time. The time points for which no data exist will be automatically removed for model inversion, after creating and filtering this design matrix. That means, all you inference will only rely on the recorded data segments - you don't create 'artifical' data.
Hope this helps
Dominik
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Dominik,
A reviewer suggested that we reanalyzed our SCR data using the PsPM toolbox, a suggestion for which I'm him or her thankful as I was unaware of this very impressive toolbox. After reading the manual and the posts on the discussion forum, I remain with the following issue. Let me first provide you with some background.
Two brief (duration of 250 ms) stimuli of different intensities were presented randomly each 30 times in total, divided across three blocks (separated by a 3 min break). SCR responses were only recorded from 2s prior to stimulus onset until 8s after stimulus offset (so no continuous recordings). The interstimulus interval varied between 9.5 and 13.5 seconds. In addition, we recorded the perceptual categorization of the stimulus on every trial (two-forced choice task, high vs. low).
I read that PsPM assumes continuous data but that it could also work with data recorded in epochs. Is this indeed the case, especially for SCR? And if so, as you suggested in your reply to a previous post, should I just include a marker that accounts for the start of each epoch in addition for a marker that indicates the start of each block?
Many thanks in advance for your highly appreciated advice.
Best,
Jonas
Jonas Zaman, MSc, PhD
Health Psychology
Centre for the Psychology of Learning and Experimental Psychopathology
Psychology and Pedagogic Sciences
KU Leuven
Room 02.09 (PSI)
Tel: +32 16 3 25840
Hi Jonas
in principle this kind of analysis is possible, and given the 8 s post-stimulus interval, also very meaningful.
The power of the PsPM approach particularly for SCR is (1) to use the full data rather than just a peak, and (2) to account for previous responses that overlap with the current one. They will still overlap with a 13.5 s ITI.
To fully exploit point (2), I would recommend importing the data in continuous time, but setting the non-recorded segments to NaN. In other words, you work out when in time did the recordings happen, and fill the gaps with NaNs.
PsPM will then be able to account for overlapping responses by creating a design matrix in continuous time. The time points for which no data exist will be automatically removed for model inversion, after creating and filtering this design matrix. That means, all you inference will only rely on the recorded data segments - you don't create 'artifical' data.
Hope this helps
Dominik