Click on the image to see a PDF version (for zooming in)

Fig. 8.1. Architecture of the HLISSOM model. Each sheet of units in the model visual pathway is shown with a sample activation pattern and the connections to one example unit. The activities are shown in gray scale as in Figure 4.1. Visual inputs are presented on the retina, and the resulting activity propagates through afferent connections to each of the higher levels. Internally generated PGO input propagates similarly to visual input. Activity in the model at any one time originates either in the PGO sheet or the retina, but not both at once. As in the LISSOM model, the activity in the cortical levels (V1 and FSA) is focused by lateral connections, which are initially excitatory between nearby neurons (dotted circles) and inhibitory between more distant neurons (dashed circles). The final patterns of lateral and afferent connections in the cortical areas develop through an unsupervised self-organizing process, as in LISSOM. After self-organization is complete, each stage in the hierarchy represents a different level of abstraction. The LGN responds best to edges and lines, suppressing areas with no information. The V1 response is further selective for the orientation of each contour; the response is patchy because neurons preferring other orientations do not respond. The FSA represents the highest level of abstraction -- a neuron in the FSA responds when there appears to be a face in its receptive field on the retina.