Back to PLS Help
Posted on 11/25/13 13:15:56
Number of posts:
Hi,
Replies:
I'm Online
Posted on 11/27/13 10:35:35
Number of posts: 229
brainscores reflect the 'overall' strength of a difference between conditions across the brain. Presumably, the individuals with high brainscores on taskB generally show the largest A/B differences (i.e. lower brainscores on taskA)????
If, for example, your behavioural data are RT, then a positive correlation of behaviour with brainscores would indicate that those with the longest RTs have the largest global differences between the conditions.
A better way to address your question might be to run a behavPLS with your data - that will show you where the conditions are similar, and where they are different.
nancy
Posted on 11/27/13 10:53:16
Number of posts:
Thanks for the reply.
I've added the actual data below. In this example, there is a significant negative correlation between Brain scores on task B and RT. I guess what I am confused about is whether the person with the most negative brain score (subject 12) has the most or the least difference in activation between tasks B vs. A .
Thanks for the help,
David
| Subject | Brain Score | RT |
| 1 | -21.1581 | 2641.48 |
| 2 | -34.84845 | 2880.925 |
| 3 | -7.0453 | 2125.32 |
| 4 | -10.2969 | 2424.565 |
| 5 | -56.6165 | 3247.935 |
| 6 | 10.7915 | 2485.77 |
| 7 | -64.90575 | 3303.96 |
| 8 | -27.99575 | 2657.685 |
| 9 | -29.57055 | 3387.125 |
| 10 | -14.8221 | 2717.47 |
| 11 | -42.52235 | 2368.475 |
| 12 | -65.28255 | 3377.95 |
| 13 | 15.37685 | 2571.02 |
| 14 | -16.6738 | 3028.665 |
| 15 | -49.1765 | 2486.495 |
| 16 | -33.0588 | 3189.775 |
| 17 | -9.5861 | 2515.115 |
| 18 | -39.05465 | 2322.993 |
| 19 | -22.12 | 3119.96 |
| 20 | -23.77995 | 2353.378 |
I'm Online
Posted on 11/27/13 11:02:31
Number of posts: 229
If you want to use the gui window, you can click on a brainscore and it will label that subject for all conditions.
Otherwise, it looks like you've pulled out the brainscores from the results, so you could look at the brainscores for taskA in the same way.
n
Posted on 11/27/13 11:24:38
Number of posts:
I just did this, and there is a positive correlation between brain scores in task A task B (r=0.68). So it looks like the person with the most positive brain score in task A also tends to have the least negative Brain score in Task B.
And actually, in both tasks A and tasks B, there is a negative correlation between brain scores and RT.
So to summarize:
- I did a non-rotated PLS, which gives a significant LV in which positive saliences are more active in Task a vs. B and negative saliences are more activae in Task B vs. A.
- I extracted brain scores from both tasks A and B
- Brain scores for Task A are negatively correlated with RT in task A.
- Brain scores for Task B are negatively correlated with RT in task B
- Brain scores in Task A are positively correlated with brain scores in task B.
So for the interpretation: I would say that
- More activation in positive salience regions in Task A were related to faster responding
- But my confusion is for Task B. Can I interpret the negative correlation between brain scores and RT in the same way as for task A? In other words, is the interpretation of a negative correlation between brain score and RT the same for positive saliences and negative saliences, or should the interpretation be flipped? Was greater activation in negative salience regions in task B associated with faster or longer RT?
Thanks, and sorry for all these messages!
David
Login to reply to this topic.
