¥ In each voxel, fit data Zn to a curve of the form
¥ Zn È a + b×tn + b×rn for n=0,1,É,N-1 (N=# time pts)
¥ a, b, b are unknown parameters
to be calculated in each voxel
¥ a,b are ÒnuisanceÓ baseline parameters
¥ b is amplitude of r(t ) in data = Òhow muchÓ BOLD
¥ Baseline model might be more complicated for long (> 150 s) continuous
imaging runs:
¥ 150 < T < 300 s: a+b×t+c×t 2
¥ Longer: a+b×t+c×t 2 + ŽT/200 low frequency
components
¥ Might also include as extra baseline components the estimated subject head
movement time series, in order to remove residual
contamination from such artifacts