The proposal involves the design and implementation of electronic systems which use physiologically realistic models for neurons and neural structures like the cortex, hippocampus, thalamus, basal ganglia etc., and the generation and experimental test of predictions of the model in both physiological and cognitive domains. The project will create an electronic system with waking cognitive processing capabilities including episodic, semantic, procedural and working memory, and cognitive support processes during sleep (slow wave and REM). The model will use physiologically realistic models for pyramidal, spiny stellate, and various interneurons in the cortex, and subcortical neurons such as spiny cell and dopaminergic neurons. Physiological realism means modeling topology, connectivity and integration of signals over time within and between neurons. The project will generate experimental predictions of the model in the areas of psychology, brain imaging and physiology and test those predictions. Ideas for the design of commercial systems for performing cognitively complex functions will be developed.

The project context is a system theoretical approach developed by staff in the ANU Department of Computer Science which demonstrates that any system that learns to perform a complex combination of functions is constrained within some architectural bounds by the need to conserve resources combined with the ongoing need to learn new functions without undesirable side effects upon previously learned functions. These bounds constrain general architectural form, how system operations are separated into modules, and even device algorithms. The systems theory 'predicts' a wide range of known physiological and psychological features of the human brain, and some as yet unobserved physiological and psychological predictions. Staff in the ANU Department of Computer Science have created electronic implementations within these architectural bounds with general resemblances to physiology and have demonstrated that a qualitative account for cognitive processing can be provided in terms of observed physiology.

In more detail the project will involve:

Phase 1

• a. Carrying out experiments to test some predictions previously generated by the system approach in the areas of physiology, brain imaging, and psychology.
• b. Implementation of physiologically realistic models for cortex pyramidal, spiny stellate and interneuron (basket, chandelier etc) cells, and integration into cortex layers, columns, arrays and areas based on existing more prototype implementations. Subcortical structures will be emulated in this phase.The implementation will require guidance from neurophysiological, psychological and brain imaging expertise both for model parameters and to design test scenarios for the model.

Phase 2

• a. Use of model test scenarios to develop more detailed experimental predictions
• b. Design and performance of physiological, psychological and brain imaging experiments to test predictions
• c. Extension of the model with more physiologically realistic subcortical models based on existing limited prototypes. This extension will again require guidance from neurophysiological, psychological and brain imaging expertise.

Phase 3

• a. Use of extended model test scenarios to develop experimental predictions
• b. Design and performance of physiological, psychological and brain imaging experiments to test predictions

The expertise required for this project includes system design and implementation, neurophysiology, brain imaging, sleep research, and experimental psychology (memory and cognitive processing).

Type of Collaboration

System design and implementation would primarily be performed by ANU Department of Computer Science. Collaborators have provisionally been identified in the areas of neurophysiology, brain imaging and sleep research. We are therefore particularly interested in possible collaborators in the area of experimental psychology.

Contact Information

If interested, please contact

Tom Gedeon
tom.gedeon@anu.edu.au
T: 02 6125 5141

Andrew Coward
andrew.coward@anu.edu.au
T: +02 6125 5694
Mob: +61 0431 526 197