Abstract

A key requirement of brain modeling that goes beyond isolated applications is that it should interrelate stimuli, neural activity, and measurements in diverse situations and be testable by multiple experiments. This requires dynamics at many scales to be incorporated simultaneously, and the results to be integrated into a unified framework. Thus, integrative models need to combine salient features of physiology, anatomy, and dynamics at various temporal and spatial scales.

A fruitful approach to integrative modeling is via neural field theory in which microscopic neural properties are approximated in a way that enables quantitative predictions for comparison with experiment, plus interpretations of real data in terms of physiology and anatomy. In particular, neural field approaches provide unified theories of multiple phenomena, and make quantitative predictions of many types of experimental observations. This enables such theories to be tested against multiple data sets, and strongly constrains their parameters and characteristics in ways that cannot be done with theories of just one or a few phenomena.

This talk outlines the main ideas of neural field theories that span from synapses to the whole brain, and are parameterized in terms of measurable physiological quantities. A selection of quantitative applications will then be reviewed, including comparisons with EEG spectra, evoked responses, seizures, arousal dynamics, and fitting of predictions to data to infer brain states and physiological parameters. Finally, several very recent directions will be discussed, selected from areas including new hybrid methods that combine spiking-neuron and field-theoretic approaches, and study of cortical waves, plasticity, and networks.