[Skip to Content]
[Skip to Content Landing]
December 1987

Computer Modeling of Mental Processes-Reply

Author Affiliations

Yale University School of Medicine PO Box 12A Yale Station New Haven, CT 06520

Arch Gen Psychiatry. 1987;44(12):1107-1108. doi:10.1001/archpsyc.1987.01800240083013

In Reply.—  Garfield raises useful questions about the parasitic states produced by overloading associative memory systems with stored information. This pathologic phenomenon has been demonstrated on the basis of computer-generated models.1,22 Whether parasitic states also occur during human cognition, and how such pathology would be experienced, are entirely open questions.Using the terminology of AI, these computer-simulation efforts reflect a version of parallel, distributed Processing (PDP). There is an increasingly widespread belief that the human cerebrum also reflects PDP Principles.3,4 While these artificial models are trivially simplistic compared with actual brains, it is likely that the energetics of all PDP systems—perhaps including those embodied by human brains—share common features despite major differences in gross architecture and/or individual neuron behavior. One likely common feature is a vulnerability to forming parasitic states under storage overload conditions.The human cerebrum seems to be organized as separate, interconnected components. For instance, AI and neurolinguistic