Adaptive resonance theory (ART)
The handbook of brain theory and neural networks
A model of computation in neocortical architecture
Neural Networks - Special issue on organisation of computation in brain-like systems
Computation of pattern invariance in brain-like structures
Neural Networks - Special issue on organisation of computation in brain-like systems
Neural Assemblies, an Alternative Approach to Artificial Intelligence
Neural Assemblies, an Alternative Approach to Artificial Intelligence
On Intelligence
Neural associative memory for brain modeling and information retrieval
Information Processing Letters - Special issue on applications of spiking neural networks
| Hi-index | 0.01 |
This work presents first simulation results substantiating a previously proposed conceptual model of computation in neocortical architecture [E. Korner, M.-O. Gewaltig, U. Korner, A. Richter, T. Rodemann, A model of computation in neocortical architecture, Neural Networks 12 (1999) 989-1005]. This model gives a detailed functional interpretation of the six-layered columnar cortical architecture and related subcortical structures. It hypothesizes three interacting processing systems at each stage of the cortical hierarchy: The A-system (middle cortical layers) accomplishes fast bottom-up processing where the first spike wave traveling up the cortical hierarchy can activate a coarse initial hypothesis at each level. In the B-system (superficial layers) the initial hypothesis is refined by slower iterative processes involving feedback. Finally, the C-system (deep layers) represents the local hypothesis of a macrocolumn which is fed back to the B-system of a lower level inducing expectations and predictions for the present and future input signals. These ideas are illustrated by an example implementation of the microcircuitry in a single cortical macrocolumn based on cell assemblies and associative memories. In a second step we have integrated our model at the level of V4 into a large scale implementation of the visual system involving several primary and higher visual cortical areas as well as parts of the hippocampal formation, and subcortical structures involved in generating eye saccades. With this model we can demonstrate object classification and the learning of new object representations.