Sparse codes and spikes

In order to make progress toward understanding the sensory coding strategies employed by the cortex, it will be necessary to draw upon guiding principles that provide us with reasonable ideas for what to expect and what to look for in the neural circuitry. The unifying theme behind all of the chapters in this book is that probabilistic inference—i.e., the process of inferring the state of the world from the activities of sensory receptors and a probabilistic model for interpreting their activity—provides a major guiding principle for understanding sensory processing in the nervous system. Here, I shall propose a model for how inference may be instantiated in the neural circuitry of the visual cortex, and I will show how it may help us to understand both the form of the receptive fields found in visual cortical neurons as well as the nature of spiking activity in these neurons.