On the Revival of Neurocognitive Modeling: Testing Old Ideas with New Techniques

Abstract

AbstractThis article describes how synergy between cognitive psychology, computer science, and artificial intelligence led to a revival of 19th-century neurocognitive modeling in the form of 20th-century neurocognitive computational modeling. Scientific evidence about the mind and its relationship to the brain began to accumulate about two centuries ago. Neurocognitive modeling based on this evidence rose to prominence in the last quarter of the 19th century, with prominent examples being the models of Wernicke and Wundt, presented as diagrams. This work lost its influence after the First World War, but came back to life after the Second. In the 1950s, researchers began developing cognitive models, which became neurocognitive in the 1960s. Moreover, with the rise of computer science and artificial intelligence, modeling became computational. Today’s neurocognitive models, realized as diagrams or computer programs, revive the early modeling of Wernicke and Wundt in several respects. While predictions were derived and tested qualitatively from the early models, today they can be derived mathematically in computer simulations and statistically evaluated for quantitative agreement with data sets. I describe how recently 20th-century techniques have been used to test 19th-century ideas about attentional control (Wundt), choice in go/no-go tasks (Donders vs. Wundt), the role of the arcuate fasciculus in speech repetition (Wernicke), and focal behavioral symptoms in neurodegeneration (Wernicke vs. Pick).