For informational purposes, the means of the reversals was noted. Following
7 days of training, the base-line threshold determinations were repeated. There followed another 7 days of training, using the same patterns and training procedure, but this time the challenge to target recognition was not blur but added noise (proportion of corrupted pixels in otherwise sharp images).
A final repeat of the base-line threshold measurements enabled some
conclusions about the stability of training improvement, if any,
and whether it might be attributed to the perception of blur or
to form discrimination of the set being used.