Hi Christopher,
In our case, each session is just under 5 minutes, each on/off block is about 20 plus/minus 2 seconds. I am not sure what you meant by including motion parameters? Any suggestions on how I may reduce the "noise" in our FEAT results?
Thanks again!
Louis.
On Wed, 19 Jun 2013 12:20:01 +0000, Watson, Christopher <[log in to unmask]> wrote:
>I don't think I would consider that a lot of movement. However, I bet the spikes in movement coincide with the active blocks. So if you are currently including motion parameters in your design, you should try an analysis without doing that.
>
>And no, your result is not very similar to results I've seen. However, my experiment is 12 minutes long. on the other hand, another task I've used is a finger-tapping task that is only 5 minutes long, and I see pretty robust results as well. What are your scan parameters?
>
>Slice-timing correction probably isn't crucial.
>
>________________________________________
>From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Louis Shue [[log in to unmask]]
>Sent: Wednesday, June 19, 2013 6:45 AM
>To: [log in to unmask]
>Subject: Re: [FSL] FEAT analysis of single hand action
>
>Hi Christopher,
>
>Is this considered "a lot" of movement?
>
>https://dl.dropboxusercontent.com/u/8460189/disp.png
>
>Since you have mentioned that the motor result is robust, was the result I linked to earlier a typical one? If not, are there any suggestions on improving the result? Note that I did not include slice timing correction in the previous result.
>
>https://dl.dropboxusercontent.com/u/8460189/rendered_thresh_zstat1.png
>
>Cheers,
>Louis.
>
>On Tue, 18 Jun 2013 13:41:51 -0400, Chris Watson <[log in to unmask]> wrote:
>
>>The visual cue will not affect motor cortex activation (well, it will in
>>a sense, but should not diminish it). For example, a motor task I give
>>to subjects uses a flashing circle to indicate hand squeezes. The motor
>>result is *very* robust. You may need to extend your experiment (more
>>blocks). And I recommend keeping the same (or nearly the same) length
>>for rest blocks.
>>Although, I would think that 5 active blocks is enough. Did this subject
>>move a lot?
>>
>>Chris
>>
>>On 06/18/2013 03:59 AM, Louis Shue wrote:
>>> Hi Christopher,
>>>
>>> What I meant was how should one design the experiments to so that activations in the visual cortex can be minimised, or conversely so that activations in the motor cortex can be more easily observed? For example, are audio cues a better idea? Or using a block design with non-uniform intervals??
>>>
>>> Thanks again.
>>>
>>> Louis.
>>>
>>> On Mon, 17 Jun 2013 11:30:50 -0400, Chris Watson <[log in to unmask]> wrote:
>>>
>>>> Why do you call it an "unintended effect"? Visual cortex activation is
>>>> inevitable here.
>>>> Did that patient move a lot? There seems to be motion artifact.
>>>> I don't think it is symmetrical, at least looking at the first 3 in the
>>>> bottom row (and a couple before that).
>>>>
>>>>
>>>> On 06/17/2013 11:14 AM, Louis Shue wrote:
>>>>> Hi Andreas
>>>>>
>>>>> On Mon, 17 Jun 2013 12:49:59 +0200, Andreas Bartsch <[log in to unmask]> wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> if the subjects >sees< a cue and has to attend to it to get the on and
>>>>>> off, you expect visual activations, don't you?
>>>>> This is exactly what I was afraid might be the case but, since we were not involved in the experimental design, I was wondering how might one isolate this unintended effect during the experimental design? And is there anything that can be done to suppress visual cortex activations without introducing additional artifacts?
>>>>>
>>>>> >From what I can see there is also left M1 activation. Was the ball put
>>>>>> into the right hand?
>>>>> Yes the ball was in the right hand. Pardon my ignorance but how did you see the "left M1 activation"? From what I can tell from rendered_thresh_zstat1.png the activations were more or less symmetrical?
>>>>>
>>>>> Louis.
>>>>>
>>>>>> Cheers,
>>>>>> Andreas
>>>>>>
>>>>>> Am 17.06.13 12:45 schrieb "Louis Shue" unter <[log in to unmask]>:
>>>>>>
>>>>>>> Dear FSL experts,
>>>>>>>
>>>>>>> We are currently using FEAT (so far only first level analysis) to analyse
>>>>>>> some fMRI recordings as a part of study on stroke rehabilitation. The
>>>>>>> experimental protocol seems "conventional" enough from what I can gather
>>>>>> >from literature.
>>>>>>> The subject (both controls and stroke patients) sees a prompt at set
>>>>>>> intervals. When it's 'on', in our case the subject squeezes and releases
>>>>>>> a pressure ball continuously, using the same hand for a given recording
>>>>>>> session). When it's 'off' the subject relaxes. There are 5 'on' cycles in
>>>>>>> each session.
>>>>>>>
>>>>>>> However, the results from FEAT show very strong activations in the visual
>>>>>>> cortex region, instead in the motor cortex!? I am not sure if we are
>>>>>>> doing anything wrong here, but can someone suggest how we may
>>>>>>> isolate/enhance the desired effects somehow?
>>>>>>>
>>>>>>> I have included the raw data and the model file - in 3 column format - if
>>>>>>> anyone wants to take a look. In the model file, it's initially in the
>>>>>>> 'off' state, all measures in seconds.
>>>>>>>
>>>>>>> Any advice is greatly appreciated!
>>>>>>>
>>>>>>> data (for a control subject, performing the required action using the
>>>>>>> right hand) :
>>>>>>>
>>>>>>> https://dl.dropboxusercontent.com/u/8460189/Cn001_20110902_CIRC_Irvin_007_
>>>>>>> ep2d_bold_3.nii.gz
>>>>>>> https://dl.dropboxusercontent.com/u/8460189/bold3_active_right.txt
>>>>>>>
>>>>>>> (part of) FEAT output:
>>>>>>>
>>>>>>> https://dl.dropboxusercontent.com/u/8460189/rendered_thresh_zstat1.png
>>>>>>> https://dl.dropboxusercontent.com/u/8460189/tsplot_zstat1.png
>>>>>>>
>>>>>>> TR=3, High pass cutoff=100, BET brain extraction checked.
>>>>>>>
>>>>>>> Cheers,
>>>>>>> Louis.
>>>>
>>
>>
|
