Larry Abbott Talk

“Harnessing the Dynamics of Neural Networks”

Large, strongly coupled neural networks tend to produce chaotic spontaneous activity.  This might appear to make them unsuitable for generating reliable sensory responses or repeatable motor patterns. However, this is not the case.  Inputs can induce a phase transition, leading to responses uncontaminated by chaotic “noise”.  Likewise, appropriately trained feedback units can control the chaos, resulting in a wide variety of repeatable output patterns.  Recently, these ideas have been extended to more realistic network models with spiking outputs.