#!/lusr/bin/php Demos: Computational Maps in the Visual Cortex
    Computational Maps in the Visual Cortex
     Demo 7.9
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
Loading...  Loading...
Loading...  Loading...
Loading...  Loading...
  (a) Retinal
activation
(b) Settled V1
response before
adaptation (SB)
(c)
SBH
(d) Settled V1
response after
adaptation (SA)
(e)
SAH
  
Play/Pause
Reset Animation
Step through
Speed Control
Depending on your connection speed, the animation may take a while to load; it is ready when images appear in the boxes above. The [<] and [>] buttons wrap around; [<] can be used to get to the final frame quickly. Original JavaScript animation source from webreference.com.
Demo 7.9. Direct and indirect tilt aftereffect. This animation of Figure 7.9 shows how the direct and indirect tilt aftereffects (described in Figure 7.1) emerge in the LISSOM model. The colors in the response plots (b,d) represent the orientation preference of each activated neuron according to the color key on top, and the saturation (i.e. brightness) represents its activation level. The initial V1 response (b) is based on the afferent weights only, before the lateral interactions, and is static in this demo. The settled V1 response (d) shows an animation of how the final stable activation pattern in response to the input shown in (a) looks after increasing amounts of adaptation to the vertical line. The animation repeats automatically once the full level of adaptation has been reached.

The top row (labeled "Adaptation") shows these responses to the same input as used for adaptation. After adaptation (iteration 7), the settled response is weaker, broader, and includes a wider range of orientations, but the perceived orientation stays approximately the same as before adaptation (iteration 0). For an input with a slightly different orientation (row "Direct effect"), more units encode orientations greater than 10o (green areas), and fewer encode those less than 10o (blue areas) in the settled response after adaptation than before. The net effect is a direct TAE, with the perceived orientation shifting away from the adaptation orientation, from 8.8o to 21.1o. For an input with an orientation very different from the adaptation pattern, the changes are more subtle (row "Indirect effect"). Only the neurons around 0o were activated during adaptation. Their inhibition from other vertical-preferring neurons increased, but decreased from those not active during adaptation. As a result, the green-colored neurons nearest 0o are now less inhibited by the rest of the neurons responding than before adaptation, and so they respond more strongly. The net effect is an indirect TAE, with the perceived orientation shifting toward the adaptation orientation, from 57.4o to 54.2o. Therefore, the LISSOM model explains computationally how the observed tilt aftereffects can arise from adapting lateral connections.

Previous demo; Next demo