Hello all,
I am new to Optseq2 and I have a question about generating 2-way design orders using the program and about a question about temporal jitter. I am working on a design with 7 conditions. I have run the following design through optseq2 and gotten out reasonable looking orders.
--ntp 160 --tr 3 --psdwin 0 21 --ev imi_touch 3 20 --ev imi_separate 3 20 --ev obs_touch 3 20 --ev obs_separate 3 20 --ev exe_touch 3 20 --ev exe_separate 3 20 --ev ident 3 20 - -o pilot_order_evrltd --nsearch 10000 --nkeep 3
1) The stimuli are hand signs presented for 3 seconds each and there are 20 repetitions for each of the conditions. The conditions that have seperate or touch in their labels involve presentation of the same set of hand signs in each condition; the different conditions just involve doing different things in response to seeing these hand signs, e.g. imitating or observing them, etc. For example, ev imi_touch, obs_touch and ev exe_touch involve seeing the same set of 20 signs and either observing, imitating or executing them. So, while I'm not interested in analyzing responses to the 20 different hands signs separately, I also don't want the same sign to appear in different conditions successively or even close in time, i.e. imitate and then observe the same sign close in time. I'm wondering if there is a way to tell optseq2 all of this information so that it will generate an order specifying not only my conditions but also randomize the specific repetition of the condition. Basically I want all of the considerations such as psdwin taken into account for condition order and only randomization taken into account for the order of trial repetitions. Somehow I need to let optseq know that the same repetition numbers in different conditions are linked. Is there any way to do this with Optseq2? If not, can you suggest some way I can achieve this, perhaps using optseq2 for condition order and then another program for trial order? Thanks!
2) The second question I have is regarding ISI jitter. I have heard that it is optimal in fast event-related designs that the stimulus duration NOT be a multiple of the TR, yet it seems that this is what optseq is looking for. It doesn't seem that optseq is generating jitter between successive stimuli but only between successive presentations of the same condition. Is this all that is necessary for deconvolution of the signal during post processing, or is a jittered ISI between successive stimuli also necessary (or desirable), and if so, how can this be achieved? One strategy I have heard of is moving the stimulus onset time around within a larger TR. For example shifting a 1.5 sec stimulus around within a 3 sec TR to create temporal jitter. Is this type of timing something optseq could generate? Thanks.
Liz