Is there training-dependent reorganization of digit representations in area 3b of string players?