Hi Mark, thanks for the answer. Could you check if what I did is
correct:
Define 2 EVs and run feat
EV1 - Sinus: Sinusoid with Period 32 and Phase 0.
EV2 - Cosinus: Sinusoid with Period 32 and Phase 24.
in the feat directory, look for pe1 and pe2, then do
fslmaths pe2.nii.gz -div pe1.nii.gz phase_angle.nii.gz
Correct?
How would I then proceed?
Thanks, Michael
On 21-May-09, at 2:08 PM, Mark Jenkinson wrote:
> Hi,
>
> Pretty much - you just need two regressors - one is the sin(a*t) and
> the
> other is cos(a*t). The relative values of the coefficients (parameter
> estimates)
> for these then encodes the phase. That is:
> tan(phase angle) = parameter estimate for cos() / parameter
> estimate for sin()
> which comes from: sin(a*t + phi) = sin(a*t)*cos(phi) +
> cost(a*t)*sin(phi)
>
> Note that the temporal derivative of sin(a*t) is a*cos(a*t), but I
> wouldn't
> use that as the factor in front is quite crucial and you need to know
> exactly how the derivative has been scaled. Instead I would just turn
> the
> temporal derivative off and explicitly make a cosine regressor with
> the
> correct amplitude.
>
> All the best,
> Mark
>
>
> On 21 May 2009, at 19:31, Michael Scheel wrote:
>
>> Dear fsl experts,
>>
>> i have data from a retinotopic mapping experiment - and I want to
>> analyse it using the so called 'travelling wave method', i.e. a
>> rotating checkerboard wedge, that jumps 45 degr. every 4 sec. After
>> 32 sec it has fully rotated and starts again. Therefore the stimulus
>> I'd choose is a sinusoid wave with a period of 32 sec. My question
>> is how to get a retinotopic map using fsl, i.e. how do I extract the
>> phase information - is this the same as the temporal derivative.
>>
>> Thanks, Michael
>>
|