Chapter11.pdf - Chapter 11 Solutions to Review Questions 11.1 All cellular systems face the noise saturation dilemma If the network is sensitive to

# Chapter11.pdf - Chapter 11 Solutions to Review Questions...

• 28

This preview shows page 1 - 4 out of 28 pages.

Chapter 11 Solutions to Review Questions 11.1 All cellular systems face the noise saturation dilemma. If the network is sensitive to large inputs, the network tends to ignore small input as noise. If the network is sensitive to small inputs, large inputs tend to saturate the system. A limited operating range seems incompatible with an unlimited dynamical range of inputs. To solve this noise saturation dilemma first a simple additive model was introduced. Additive dynamics resemble linear dynamics. The change in neuronal activation equals a sum of activation and signals, not a product or other non linear function. In the additive model, each node receives an external input I i and that the individual nodes do not interact with each other. The basic requirement of the network is to remain sensitive to the input ratios θ i . the i th neuron receives input I i . the convex coefficient θ i defines the “reflectance” I i : I i = θ i I I I i i = θ The simplest additive activation model combines the feed forward input I i , the passive decay rate A and the activation bound [0, B ]. Then the additive model is as follows: ( ) i i i i I x B Ax x + = (1) where if B is the total number of cells then x i is the number of cells which are turned ON and ( B – x i ) cells are turned OFF. I i is the intensity which makes the OFF cells ON. Equation (1) can be written as, ( ) i i i i BI x I A x + + = x 1 x 2 x i x n + + + + I 1 I 2 I i I n
Under equilibrium conditions i x = 0 ( ) 0 = + + i i i BI x I A or i i i I A BI x + = Now using I I i i = θ , we have I A I B x i i i θ θ + = B x I As i , Thus we see that x i approaches B even if the relative input intensity θ i is small. This is called saturation. Thus we see that additive models saturate. Now to preserve the sensitivity to the input signals, in the network, we consider a multiplicative model suggested by Grossberg. In it, the i th input excites the i th cell and inhibit all others. This is called an ON centre OFF surrounded kind of network or a cooperative competitive net. We know that, + = = i j j i i i i I I I I I θ Thus the equation of a cooperative competitive net is given by the following equation. ( ) n i I x I x B Ax x i j j i i i i i , 2 , 1 , L L = + = The term i j i i I x shows how the inputs I j j i inhibit the other excited nodes. - - - x 1 x 2 x i x n I 1 I i I n +
( ) i i i i i j j i i BI x I A BI x I I A x + + = + + + = Under equilibrium condition 0 = i x 1 + = + = + = I A i i i i B I A BI I A BI x θ θ i i B x I θ as There we see that at equilibrium the neuronal activation measures the scaled pattern information θ i . Thus pattern sensitivity replaces activation saturation. Thus we see that according to the Grossberg Saturation Theorem: Additive Models saturate, multiplicative models do not.

#### You've reached the end of your free preview.

Want to read all 28 pages?

• Summer '18
• Shiv
• Pattern matching

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern