作者: David Somers , Nancy Kopell
DOI: 10.1007/BF00198772
关键词: Modulation (music) 、 Artificial neural network 、 Oscillation 、 Computer science 、 Synchronization 、 Topology 、 Synchronization networks 、 Relaxation (approximation) 、 Relaxation oscillator 、 Control theory 、 Coupling (computer programming)
摘要: Synchronization properties of locally coupled neural oscillators were investigated analytically and by computer simulation. When in a manner that mimics excitatory chemical synapses, having more than one time scale (relaxation oscillators) are shown to approach synchrony using mechanisms very different from with sinusoidal waveform. The relaxation make critical use fast modulations their thresholds, leading rate synchronization relatively independent coupling strength within some basin attraction; this is faster for have conductance-based features caricatures such as the FitzHugh-Nagumo equations lack features. Computer simulations one-dimensional arrays show regime synchronize much rapidly same whose parameters been modulated yield We present heuristic explanation effect based on can affect way scales array length. These results suggest emergent behavior oscillating networks be dramatically influenced intrinsic network components. Possible implications perceptual feature binding attention discussed.