Computational Maps in the Visual Cortex
     Figure 7.8
MiikkulainenBednarChoeSirosh
Home    
About the Authors
Back Cover    
Table of Contents 
Sample Chapter 
Figures    
References    
Errata    
Demos     
Talks/Courses 
Software    
Credits    
Purchase online at:

springeronline.com
amazon.com

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

Fig. 7.8. Changes in lateral inhibitory weights due to adaptation. The strengths of the inhibitory weights to the neuron marked with a small black square are shown in color coding as in Figure 5.12. This neuron prefers an orientation of +15o, as seen from its RF in (a), plotted in gray scale as in Figure 6.4. Before adaptation, the inhibitory connections mostly come from neurons with similar preferences along this same orientation (b). As a result of adapting to the vertical input in the center, the blue areas corresponding to the response to that input become stronger (c). This effect is summarized in (d), computed by subtracting the weights in (b) from those in (c), scaling the positive values up to a visible level, then labeling each connected neuron with the color corresponding to its orientation preference. Connections increased only from neurons with vertical preferences, i.e. those that were active in the settled response (Figure 7.9c,e, top row). As a result of normalization, the rest of the connections decreased, as seen by plotting the negative differences in (e). These connections include all orientations other than vertical. The orientation-specific changes are summarized in the difference histogram (f), which shows that the net connection strength to neurons with preferences around vertical increased, while connections to other orientations decreased. Together these changes give rise to the direct and indirect TAE, as shown in Figure 7.9.