Disentangling a complex pattern of interrelationships among multiple measures of Alzheimer neuropathology and clinical status: Comments on Cummings, Pike, shankle, and Cotman