We used a sigma of 200s for all tasks. In practice if you clean the data I doubt it matters whether you do anything more than a linear detrend.
Peace, Matt. From: <[email protected]<mailto:[email protected]>> on behalf of David Hofmann <[email protected]<mailto:[email protected]>> Date: Thursday, June 29, 2017 at 7:56 AM To: hcp-users <[email protected]<mailto:[email protected]>> Subject: [HCP-Users] How to choose high-pass filter cut-off period for the task data (block design) Hi all, I was wondering how to choose the high-pass filter cut-off period (in seconds) for the task data when the block onset files are used for the 1st level GLM (with spm12). As far as I understood the period should be around twice the period until a block repeats. Based on the .txt onset files (reading out the block duration) I determined the following cut-off frequencies (in seconds), but I'm confused about what to choose for the Motor and the Working memory task: Emotion: 72s (block length=18s -> 2*18*2) Gambling: 112s (block length = 28s -> 2*28*2) Language: 86s (block length math ~13s, block length story ~ 30s -> (13+30)*2) Motor: ? (block length = 12s, but there are several blocks, so not sure here) Relational: 64s (block length = 16s -> 2*16*2) Social: 92s (block length = 23s -> 2*23*2) WM: ? (the same blocks don't repeat here, so not sure) I'm using the already preprocessed volume data files (e.g. tfMRI_EMOTION_LR.nii from 100307_3T_tfMRI_EMOTION_preproc) Any help appreciated! greetings David _______________________________________________ HCP-Users mailing list [email protected]<mailto:[email protected]> http://lists.humanconnectome.org/mailman/listinfo/hcp-users _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
