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

Fig. 4.1. Architecture of the basic LISSOM model. LISSOM consists of a hierarchy of two-dimensional sheets of neural units, including an array of retinal receptors, ON and OFF channels in the LGN, and a cortical network representing V1. The LGN and V1 activation is shown in gray-scale coding from white to black (low to high). The activity on the retina (a single oriented Gaussian) is presented like natural images: Light areas are strongly activated, dark areas are weakly activated, and medium gray represents background activation. This input gray scale will be used for all models that include the LGN and which can therefore process natural images. Sample connections are shown for one unit in each LGN sheet and one in V1. The LGN afferents form a local anatomical receptive field on the retina, and cause ON-center LGN units to respond to light areas surrounded by dark, and OFF-center units to dark areas surrounded by light. Neighboring LGN neurons have different but overlapping RFs. Similarly, V1 neurons have afferent receptive fields on the LGN sheets. V1 neurons also receive lateral excitatory and lateral inhibitory connections from nearby V1 neurons; these connections are shown as dotted and dashed circles around the V1 neuron, respectively. V1 activity is patchy because only those neurons respond whose feature preferences match the orientation, eye of origin, and direction of movement of the pattern currently in their receptive fields.