There are two data sets here. 

  resphantom2.mat
  3T_coronary.mat

The resolution phantom is a great data set to start with.  The
off-resonance blurring is easy to see, and the two data sets align
exactly.  The coronary data set is more challenging.  The two data
sets are shifted by about a pixel, which complicates computing a high
resolution map.

In each case the matlab variables are

  d1  -- data for echo time 1
  d2  -- data for echo time 2
  te1 -- echo time 1
  te2 -- echo time 2
  k   -- k-space trajectory
  w   -- weighting

The weighting is an analytic approximation, you may want to compute the 
Voronoi weighting. The sampling time is 4 us for each.

The phantom data should be reconstructed as a 160x160 pixel image, and
the coronary data as a 400x400 pixel image. Example coronary
reconstructions are included in that data set.