Correlated Neuronal Activity Can Represent Multiple Binding Solutions

The correlated activity of neurons is being regarded as candidate mechanism for binding the responses of different feature selective cells. Binding, however, is rarely unique, especially at levels beyond feature detection. In this article the dynamical properties of a network of coupled oscillators as a model for a mental representation of visual stimuli are investigated. Oscillator networks have been analyzed previously, but here a new coupling scheme and a biologically more relevant transfer function are used. A principal mode analysis of the activity shows that the network sensibly handles ambiguous segmentations of a visual stimulus