We re-examined the concepts of "disconnection" and "connectivity", which are used in both clinical neuropsychology and studies on schizophrenia. First, we examined if classical neuropsychological symptoms, which were originally proposed as disconnection syndromes, i.e. callosal disconnection syndrome, conduction aphasia, and pure alexia, should be referred to as "disconnection syndromes". We concluded that callosal disconnection syndrome and pure alexia can be regarded as "disconnection syndromes", but conduction aphasia only partially fulfills our provisional criteria for disconnection syndrome. We next evaluated whether disconnection/disconnectivity models of schizophrenia fulfill the requirements for disconnection syndromes, and concluded that these models do not satisfy any of our proposed criteria. Thus, the major disconnection/disconnectivity models of schizophrenia should not be called "disconnection syndromes" or even "hypothetical disconnection syndromes". In order to retain the disconnection hypothesis, the underlying pathology of schizophrenia may be regarded as partial and widespread disconnection, but this is not necessary to fulfill the criteria for "disconnection syndrome" in the sense of clinical neuropsychology. However, such a modification makes the concept of disconnection over-inclusive. We found that our bias in regarding schizophrenia as a disease of disconnection may originate, in part, from historical and accidental reports from a century ago such as the renaming of schizophrenia from dementia praecox to "schizo-phrenia(splitting of the mind)" and the conceptualization of its fundamental psychopathology as "loosening of association".
 <Authors' abstract>