Thank you very much for the response, Michael!
At the moment, no specific corrections have been performed, although I am not sure what kind of corrections are done during the subtraction process implemented in the scanner; in that context, what kind of corrections do you propose? We used a very short TE so I guess that the data should not be contaminated by BOLD signal that much; however, do you think that ICA would be a good choice to denoise the data?
Kind regards,
Jan
On Jan 2, 2013, at 11:04 AM, Michael Chappell wrote:
> Jan,
>
> That seems reasonable at least to first approximation. As you have spotted the numbers appear to be relative and there is no guarantee that numbers can be compared on different days (depending upon any other corrections that are or not being done). Whether what you are seeing is a genuine difference or due to noise will depend of course on how noisy the data looks.
>
> Michael
>
> On 28/12/2012 22:28, Jan B wrote:
>> Dear members,
>>
>> I would like to calculate the % difference between two resting-state PCASL data sets of one subject that were collected on different days. The scanner we use (Philips Achieva) calculates and outputs the subtraction images (although it is also possible to output the raw tag-control data...).
>> One idea I have to perform this analysis is to calculate the “perfusion mean” of each run, i.e. of the subtraction,
>> using the fslstats –M command. When I do this, I get numbers as 18.7 for day 1 and 20.9 for day 2 (I guess the numbers are nondimensional…).
>> My question is: Is it possible to say that (global) perfusion in this subject on day 2 (20.9) is 11,8 % higher than on day 1 (18.7)?
>>
>> I would be very greatful if someone could answer me this (probably easy) question.
>>
>> Thank you very much and kind regards,
>>
>> Jan
>>
|