Concept formation in neural networks: Implications for evolution of cognitive functions

The main paradigm of cognitive neuroscience is the connectionist concept postulating that the higher nervous activity is performed through interactions of neurons forming complex networks, whereas the function of individual neurons is restricted to generating electrical potentials and transmitting signals to other cells. In this article, I describe the observations from three fields—neurolinguistics, physiology of memory, and sensory perception—that can hardly be explained within the constraints of a purely connectionist concept. Rather, these examples suggest that cognitive functions are determined by specific properties of individual neurons and, therefore, are likely to be accomplished primarily at the intracellular level. This view is supported by the recent discovery that the brain’s ability to create abstract concepts of particular individuals, animals, or places is performed by neurons (“concept cells”) sparsely distributed in the medial temporal lobe.

Download Full-text

Parallel Distributed Processing and Neural Networks: Origins, Methodology and Cognitive Functions

International Journal of Neuroscience ◽

10.3109/00207459109080640 ◽

1991 ◽

Vol 60 (2) ◽

pp. 195-214 ◽

Cited By ~ 5

Author(s):

Randolph W. Parks ◽

Debra L. Long ◽

Daniel S. Levine ◽

David J. Crockett ◽

Edith G. McGeer ◽

...

Keyword(s):

Neural Networks ◽

Cognitive Functions ◽

Distributed Processing ◽

Parallel Distributed Processing

Download Full-text

Resetting the brain

Philosophy of Psychedelics ◽

10.1093/med/9780198843122.003.0006 ◽

2021 ◽

pp. 110-123

Author(s):

Chris Letheby

Keyword(s):

Neural Networks ◽

Cognitive Functions ◽

Prediction Error ◽

Cognitive Process ◽

Large Scale ◽

Predictive Processing ◽

Therapeutic Benefits ◽

High Level ◽

The Brain

‘Resetting the brain’ examines the hypothesis that (i) large-scale neural networks become stuck in dysfunctional configurations in pathology, and (ii) psychedelics cause therapeutic benefits by disrupting these configurations, providing an opportunity to ‘reset’ the relevant networks into a healthier state. This chapter argues that this view is correct but limited; per Chapter 5, it needs to be supplemented with an account of these networks’ cognitive functions. To this end, the chapter introduces the predictive processing (PP) theory of cognition, which views the brain as an organ for prediction error minimisation. One PP-based theory of psychedelic action claims that (i) the networks targeted by psychedelics encode high-level beliefs, and (ii) psychedelic disruption of these beliefs provides an opportunity to revise them. This is the cognitive process that corresponds to the ‘resetting’ of neural networks. The chapter concludes by proposing that the beliefs most often revised in successful psychedelic therapy are self-related beliefs.

Download Full-text

Homogeneous neural networks cannot provide complex cognitive functions

Behavioral and Brain Sciences ◽

10.1017/s0140525x99341820 ◽

1999 ◽

Vol 22 (2) ◽

pp. 293-293 ◽

Cited By ~ 1

Author(s):

Alexey M. Ivanitsky ◽

Andrey R. Nikolaev

Keyword(s):

Neural Networks ◽

Cognitive Functions ◽

Cell Assemblies ◽

Functional Systems ◽

The Brain

Within the Hebbian paradigm the mechanism for integrating cell assemblies oscillating with different frequencies remains unclear. We hypothesize that such an integration may occur in cortical “interaction foci” that unite synchronously oscillated assemblies through hard-wired connections, synthesizing the information from various functional systems of the brain.

Download Full-text