Menu ▾ ▴

Task Duration, Import, and Trimming

Help
Alex Francisco
2014-09-16
2014-12-01

  • Alex Francisco

    Alex Francisco - 2014-09-16

    Hi Dr. Bach,

    I have a couple questions regarding 1) task duration and event onsets and 2) trimming.

    1) I am currently running a standard GLM analysis of SCR’s during a 4 second decision phase. My task is roughly 27 minutes long, or about 1,620 seconds. However, when I import the files (which are .acq) into Matlab via the import function in SCRalyze, the duration is stated under the “infos” variable as (original duration/4). For example, the duration of one subject was exactly 27.3249 minutes, or 1,639.494 seconds. When imported to .mat format through SCRalyze, the duration is stated exactly as 409.8735 which is perfectly 1,639.494/4.

    To get to my question, first, is there a specific reason for this division? Is it simply a function of the .mat format? I am not extremely familiar with Matlab. Second, would this have an effect on the multiple condition files used during analyses? Since I currently have them specified in the original seconds (i.e. not divided by 4) would SCRalyze then not be able to properly identify the onsets?

    2) Keeping my previous question in mind, I have a few questions about data trimming. I am interested in doing the analyses separately for each third of my task. When trimming the imported SCR files, should I trim in regards to this duration/4 model? For example, taking my previous subject whose task was 27.3249 minutes, I could trim the file to the first third stating that the first third was from 0 seconds from the start of the file, to 551.367 from the start. Or, in keeping with the duration/4 method, I could state that the first third was from 0 to 137.84175 (551.367/4). Which one is correct? I get clearly different results when trimming with standard seconds vs. (standard seconds/4).

    3) Last question, still in regards to trimming. If I am interested in looking at the each third of the task separately, would you suggest trimming the files into 3 separate files with the trim function? This would be as opposed to simply using different multiple condition files to examine each third separately (i.e. one condition file that contains onsets for the first third, then run the GLM again with the untrimmed scr file using a condition file containing onsets of the second third, then repeat for the third third).

    I am using files that have been imported using the transfer function, so they are in microsiemens and do not necessarily have to be normalised for analyses. I would assume that the process of normalisation would affect SCR’s differently when considering whole task vs. trimmed task, but I may be wrong in that.

    Sorry for all of the questions, if it would help I can easily send you some of my data files.

    Sincerely,
    Alex

     

  • Dominik Bach

    Dominik Bach - 2014-12-01

    Dear Alex

    I apologise for the unacceptable delay in my reply. I had not been alerted to your contribution in this forum automatically, and I failed to check manually.

    The import problem you mentioned has been re-posted and I'm not trying to resolve it. Please bear with me for another week or two.

    I hope this answers also question 2. Wrt question 3, I would recommend splitting the file if there are long breaks with unmodelled responses.

    Best wishes
    Dominik

     

Log in to post a comment.