Back to PLS HelpPosted on 05/06/13 13:26:49
Number of posts: 9
Hi,
I have run a seed (hippocampus)/behaviour (performance) PLS for a group of participants that completed three tasks. I am having trouble wrapping my head around the results.
There is a significant LV that shows a pattern for which all of the three tasks seed contributes, but only one tasks' performance behaviour contributes - which I take to mean that hippocampal circuit is involved in two of the tasks all of the time and one of the tasks (in which performance contributed) dependent on performance (i.e., more performance, more hippocampal connectiviity).
The other significant LV shows a pattern that invovles beahviour (perfomrance) for all three tasks, but only seed for one task - does this mean that the hippocampal connectivity is only involved in that one task?
Thanks for any feedback!
Best,
Signy
Replies:
I'm Online
Posted on 05/06/13 22:06:12
Number of posts: 229
Sounds like a very cool dataset :-)
I might reword your description of the results like this:
LV1 is showing that the dominant HPC functional connectivity pattern is with regions XXX, regardless of task demands.
For one task, the HPC functional network also relates to (some aspect of) behaviour.
LV2 is showing that the activity in a second set of regions is related to (some aspect of) behaviour, but those regions are functionally connected to HPC in only one condition.
You didn't indicate whether the LV1 behav correlations and the LV2 HPC correlations are in the same task or not, but that would be important to consider as you move forward.
nancy
Posted on 05/07/13 07:31:46
Number of posts: 394
another important addition to Nancy's - that there is more than one pattern of hippocampal functional connectivity (this emphasizes that functional connectivity is not static)
one pattern that supports behavior in Task X only (LV1)
one pattern the support behavior in Task Y (LV2)
Posted on 05/07/13 10:22:28
Number of posts: 9
another important addition to Nancy's - that there is more than one pattern of hippocampal functional connectivity (this emphasizes that functional connectivity is not static)
one pattern that supports behavior in Task X only (LV1)
one pattern the support behavior in Task Y (LV2)
Thanks for your feedback!
To Nancy's comments, the behaviour correlations in LV1 are for a different task than the HPC correlations in LV2. If the interpretation for LV1 is there is a HPC functional connectivity pattern (with regions XXX) that is there regardless of task demands (but for one task, the HPC functional network also relates to behaviour) - can I look at these seed/behav simulatenously (multiblock?) to see if task demands interact with HPC connectivity in each of these conditons. For example, if some aspect of behaviour is related to this LV1 for one task - how can I see if greater prefomrance is driving this particular HPC pattern, which is not present with lower performance?
Randy - I really like the idea of finding multiple hippocampal connectivity patterns! So, can I say that a hippocampal connectivity pattern in LV2 only supports behavour in Task Y (the task in which both seed and beahviour were significant) ?
Again, thanks for you help!
Best,
Signy
Posted on 05/09/13 08:44:26
Number of posts: 394
HI Signe,
For your first point, isn't the fact that the pattern of HPC functional connectivity varies with performance enough to answer the question? From my I understanding of your results, if HPC functional connections and behavior vary together on one task (LV1), then that suggests the stronger performance is related to stronger HPC connectivity (ignore the sign of the correlation for the time being). If you want to do high vs low performance, you would need to split the sample, which I would not recommend as it changes the data distribution. For the second LV, the same logic holds, only that its a different set of regions for HPC connectivity - which is cool.
If you want to talk more about this, let me know and we can find a time to meet.
Login to reply to this topic.
