Question about Stimuli Set

[Xiaowei Zhuang]

Hi Dr. Stark,

We are considering running the MST task in the MR scanner and collect the fMRI data. We currently use stimuli set 1 to 6. 

I noticed some pairs of stimuli are differed in orientation more than object properties, for instance image 056a and 056b in Set 6. And some other pairs of stimuli differs in the spatial location more than object properties, for instance image 097a and 097b in Set 6.

I am wondering will the differences in object properties and differences in spatial location or orientation induce a confound in this task? 

Thank you for your time and response. 

Best,

Xiaowei. 

Xiaowei Zhuang   |  Research Engineer      |  Lou Ruvo Center for Brain Health  Cleveland Clinic  |  888 W Bonneville Ave.   |  Las Vegas, NV 89106  | (702) 685-6051

[Craig E.L. Stark]

It's possible.  We try to match the sets, but it's never perfect and it's never along every dimension (even in the 1-6, it's starting to look like one set is a bit different...)

But, the biggest aspect about them differing along multiple dimensions is what the subject perceives.  If the subject thinks the task can be solved by just looking at the location, then they'll adopt that strategy.  Given that: a) there are still many dimensions that change, b) subjects really aren't that perceptive, and c) they'd need to know these dimensional biases before the study phase to enact a strategy, I'm not terribly worried.  Heck, I remember back when we all did blocked fMRI designs how we'd do things like blocks of targets and foils but have it be say an 80/20 mix for 10 or so items.  So 8T+2F, 2T+8F, 8T+2F, ...  Not a single participant ever picked up on the notion that these weren't randomly distributed in there.

Craig