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

Fig. 4.4. Example input and response. At each self-organization iteration in LISSOM, the photoreceptors in the retina are activated according to the gray-scale values of the input image (as described in Figure 4.1). In this example, two unoriented Gaussians with width &sigma_u = 3.0 were drawn on random spatially separated locations on a 54 × 54 array of receptors (a). The 36 × 36 ON and OFF cell responses are plotted in the central square of (b) by subtracting the OFF cell responses from the ON: Dark areas represent higher activity in the OFF channel, light areas higher activity in the ON channel, and medium gray represents equal activation in both channels. The LGN responds to edges and lines in the input, with high ON cell activations where the input is brighter than the surround, and high OFF cell activations where the input is darker than the surround. Before self-organization (i.e. iteration 0), the 142 × 142 V1 map initially responds broadly and unspecifically to the input patterns (c); V1 activations are represented in the central square in gray scale from white to black (low to high). The lateral connections focus the response into discrete activity "bubbles" in (d), and connections are then modified. After 10,000 input presentations and learning steps, the initial and settled V1 responses are more focused, forming a sparse representation of the input (e and f ). This figure also illustrates how the retina is mapped to the LGN and the LGN mapped to V1 in LISSOM. The LGN and V1 networks are drawn to the same scale as the retina (as indicated by the outside squares), so that activity at a given location in the LGN and V1 corresponds to a stimulus at the corresponding location in the retina. The retina is larger than the LGN and the LGN larger than V1 so that all LGN and V1 neurons have complete receptive fields (i.e. they are not cut off by the network boundary; Figure A.1). In subsequent activity figures, such padding is omitted and only the retinal and LGN area that matches the V1 network is shown. An animated demo of the map response can be seen at ...