cursor control

Forum for discussion on different brain signals
Locked
emily
Posts: 40
Joined: 24 Mar 2008, 06:13

cursor control

Post by emily » 10 Jul 2008, 01:53

Hello,
Is it possible that a user could be controlling the vertical cursor movement with eye movement for the 1D cursor task. For example glancing to the left may make the cursor move upward while staring straight ahead makes it move down? Do you generally record HEOG and VEOG so that you can identify if this is how they are controlling the cursor?

I have one participant who achieved 80% accuracy in their initial session of 15 runs made up of 17 trials. Could this have been through learning to modulate brain rhythms or do you think they must be using something like eye movement to gain control so quickly.

thanks,
Emily

gschalk
Posts: 615
Joined: 28 Jan 2003, 12:37

Cursor Control ...

Post by gschalk » 10 Jul 2008, 12:14

Emily,

Congratulations to your successful experiments! It is definitely possible that people reach 80% accuracy after the first session. What you will always need to do is to make sure that control was produced by the brain and not by other sources (actually, eye movements will produce changes in frequencies that are much lower than those used for mu/beta rhythms, so it is unlikely that these were the source for cursor control, but muscle movements could be the reason). To do this, you should do topographical analyses similar to those done in the initial screening. If you still see activations that are focused sharply in the frequency domain and also sharply in space over sensorymotor cortex, you are fine.

Gerv

emily
Posts: 40
Joined: 24 Mar 2008, 06:13

Post by emily » 14 Jul 2008, 08:30

Thanks Gerv,
Can you clarify what I need to put in the offline analysis to get that information.

The states I think I need to look at are
states.TargetCode=1(bottom target),
states.TargetCode=2 (top target),
states.ResultCode=1 (hit bottom/miss top)
states.ResultCode=2 (hit top/miss bottom)

But I'm not how these need to be placed in the offline analysis to get the results I'm after.

Also is it enough to analyse each dat. file made up of 17 trials or should I analyse multiple files from the same session.

cheers
Emily

gschalk
Posts: 615
Joined: 28 Jan 2003, 12:37

analysis ...

Post by gschalk » 15 Jul 2008, 15:12

Emily,

What you need to compare is:

(states.TargetCode==1)&(states.Feedback==1)
(states.TargetCode==2)&(states.Feedback==1)

(i.e., up target on the screen and cursor is moving vs. down target on the screen and cursor is moving)

Gerv

emily
Posts: 40
Joined: 24 Mar 2008, 06:13

Post by emily » 18 Jul 2008, 03:25

Hi Gerv,

So do you mean that in the offline analysis window I need to enter the following?

Condition 1: states.TargetCode == 2
Label 1: Top target

Condition 2: states.TargetCode == 1
Label 2: Bottom target

Trial Change Condition: states.Feedback

Wouldn't I need to enter the ResultCode somewhere also? I know it won't matter for trials where participants have 100% accuracy, but it would need to be included in trial analysis when this isn't the case wouldn't it?

Thanks,
Emily

emily
Posts: 40
Joined: 24 Mar 2008, 06:13

Post by emily » 18 Jul 2008, 03:30

Hi Gerv,

So do you mean that in the offline analysis window I need to enter the following?

Condition 1: states.TargetCode == 2
Label 1: Top target

Condition 2: states.TargetCode == 1
Label 2: Bottom target

Trial Change Condition: states.Feedback == 1

Do I need to enter the ResultCode somewhere also? I know it won't matter for trials where participants have 100% accuracy, but it would need to be included in trial analysis where this isn't the case wouldn't it?

Thanks,
Emily

emily
Posts: 40
Joined: 24 Mar 2008, 06:13

Post by emily » 18 Jul 2008, 03:34

Hi Gerv,

So do you mean that in the offline analysis window I need to enter the following?

Condition 1: states.TargetCode == 2
Label 1: Top target

Condition 2: states.TargetCode == 1
Label 2: Bottom target

Trial Change Condition: states.Feedback

Wouldn't I need to enter the ResultCode somewhere also? I know it won't matter for trials where participants have 100% accuracy, but it would need to be included in trial analysis when this isn't the case wouldn't it?

Thanks,
Emily

gschalk
Posts: 615
Joined: 28 Jan 2003, 12:37

analyses ...

Post by gschalk » 18 Jul 2008, 05:29

Emily,

Condition 1 should be:

(states.TargetCode==1)&(states.Feedback==1)

Again, this means that the target was on the top (I believe 1 is top) and there was feedback, i.e., the cursor was moving. If you simply use states.TargetCode==1, then you include the period in which the subject is not actually controlling the cursor (i.e., before the cursor appears), which is not what you want.

Condition 2 should be:

(states.TargetCode==2)&(states.Feedback==1)

Trial Change Condition should be something like:

states.IntertrialInterval==1

or

states.TargetCode > 0

Gerv

Locked

Who is online

Users browsing this forum: No registered users and 1 guest