Closed-form solutino using trapezoidal approximation

Pei, M., Nguyen, T.D., Thimmappa, N.D., Salustri, C., Dong, F., Cooper, M.A., Li, J., Prince, M.R., Wang, Y., 2014. Algorithm for fast monoexponential fitting based on Auto-Regression on Linear Operations (ARLO) of data. Magnetic resonance in medicine 73, 843–850.

../../_images/trapezoidal.png

Algorithm parameters overview

algorParam.r2s.

Description

s0mode

Method to extrapolate the T1-weighted magnitude signal (S0) from the multi-echo magnitude data

Usage

algorParam.r2s.s0mode

Method to extrapolate the T1-weighted magnitude signal (S0) from the multi-echo magnitude data. Options are '1st echo' (use the first echo intensity directly), 'weighted sum' (signal-intensity-weighted combination across echoes) and 'averaging' (average of the echo-time-corrected signal across echoes)

Default Value: ‘1st echo’

Examples:

algorParam.r2s.s0mode = '1st echo';

algorParam.r2s.s0mode = 'weighted sum';

algorParam.r2s.s0mode = 'averaging';