Individual cells in the cortex, they found, responded not to the presence of light, but rather to the presence of edges in their region of the visual field. Furthermore, cells were found which would fire only in the presence of a vertical edge at a particular location in the visual field, while other nearby cells responded to edges of other orientations in that same region of the visual field. These orientation-sensitive cells were called "simple cells", and were found all over the visual cortex.
This suggested that simple cells posessed a patterned receptive field, with excitatory and inhibitory regions (represented by white and gray regions below) so that the cell would fire only if it received input (due to light) in the excitatory portion of its receptive field in the absence of input from the inhibitory portion. In other words the cell responds to spatial features which correspond to the spatial pattern of excitatory and inhibitory synapses in its own receptive field.
This operation is analogous to the operation of "edge detectors" in image processing, which process an image by spatial convolution with an edge "kernel" consisting of adjacent positive and negative values.
Even higher level "hypercomplex" cells were discovered which would respond to more complex combinations of the simple features, for example to two edges at right angles to each other in a yet larger region of the visual field.
All of this suggested a hierarchy of feature detectors in the visual cortex, with higher level features responding to patterns of activation in lower level cells, and propagating activation upwards to still higher level cells. This model of a hierarchy of feature detectors performing successive processing of the visual input remains the dominant view of visual processing in the visual cortex. It has inspired a number of computational models of visual processing involving successive stages of image convolution using patterned spatial kernels.