Is there "feedback" during visual imagery? Evidence from a specificity of practice paradigm.