Neural Computation and the Computational Theory of Cognition

Abstract

We begin by distinguishing computationalism from a number of other theses that are sometimes conflated with it. We also distinguish between several important kinds of computation: computation in a generic sense, digital computation, and analog computation. Then, we defend a weak version of computationalism-neural processes are computations in the generic sense. After that, we reject on empirical grounds the common assimilation of neural computation to either analog or digital computation, concluding that neural computation is sui generis. Analog computation requires continuous signals; digital computation requires strings of digits. But current neuroscientific evidence indicates that typical neural signals, such as spike trains, are graded like continuous signals but are constituted by discrete functional elements (spikes); thus, typical neural signals are neither continuous signals nor strings of digits. It follows that neural computation is sui generis. Finally, we highlight three important consequences of a proper understanding of neural computation for the theory of cognition. First, understanding neural computation requires a specially designed mathematical theory (or theories) rather than the mathematical theories of analog or digital computation. Second, several popular views about neural computation turn out to be incorrect. Third, computational theories of cognition that rely on non-neural notions of computation ought to be replaced or reinterpreted in terms of neural computation.Keywords: Analog computation; Computational theory of cognition; Digital computation; Neural computation; Spike trains The brain is an analogical machine, not digital. (Lashley, 1958, p. 539) [A]ny [psychological] theory that is incompatible with known facts of neurology must be, ipso facto, unacceptable. (Fodor, 1965, p. 176) Correspondence should be sent to Gualtiero

Beyond cognitive myopia: a patchwork approach to the concept of neural function

“…3 The predominant assumption is that information processing contributes to cognitive functions: neural systems process information about the environment to affect the behavior of the organism (deCharms and Zado 2000, pp. 631f., Piccinini andBahar 2013, p. 456f., Panzeri et al 2017, p. 471).…”

The cognitive neuroscience revolution

“…First, whether nervous systems are universal, program-controlled digital computers is an empirical question; such a question cannot be settled independently of neuroscience. Further, and more importantly, evidence from neuroscience suggests that neural computation, in the general case, is in fact not a form of digital computation (Piccinini and Bahar 2013). Since digital computation is a necessary (though insufficient) condition for a system to be a universal, program-controlled digital computer, current best evidence suggests that nervous systems are in fact not such systems.…”

Computing Mechanisms and Autopoietic Systems

“…My full thoughts on this must be saved for another occasion, but in brief I think that computation can offer enactivism and autopoietic theory a convincing mechanistic base, and that in return enactivism and autopoietic theory can help computation escape from the metaphysical baggage that representationalism has burdened the received view with. In addition, paying attention to enactivism and autopoietic theory might help the development of nontraditional models of "sui generis neural computation", as hinted at by Piccinini and Bahar (2013), and considered more explicitly by Friston (2013).…”