Physics 380, 2011: Lecture 27

Back to the main Teaching page.

Back to Physics 380, 2011: Information Processing in Biology.

Presentation: Clay Owens, Fairhall et al, 2001 paper.

Class summary

See the paper Nemenman, 2011, for the summary.

• We concluded that information processing in biological systems is probabilistic.
  • We introduced concepts from probability theory to study stochastic processes.
  • We spent a lot of time on random walks, and memorized the mantra: "means add and variances add"
  • We studied different numerical/analytical methods for representing stochastic processes: master, Langevin, Fokker-Plank
• Information-theoretic characterization of biological signal tranduction
  • We introduced all the necessary quantities to quantify biological signal transduction
  • We studied properties of these quantities.
  • We paid particular attention to the Data Processing Inequality due to its importance in understanding the structure of biological signaling networks.
  • We concluded that information and evolutionary fitness are tightly related.
• We studied dynamics of biological information processing
  • We introduced the tools for quasi-stationary analysis of biological networks
  • We introduced the tools for linear analysis (Fourier analysis) of biological networks
  • We introduced some ideas from dynamical systems theory
  • We studied how information processing in the dynamical context can be improved
• One of the best tools for improving information transduction properties is adaptation
  • We classified different means of adaptation and showed how they emerge from information maximization arguments
  • We discussed how higher cognitive concepts such as learning can be quantified in similar means.

What's next for the field?

Computational and experimental tools will allow us, in the next few years, to start addressing the following questions.

• Can we show, experimentally and theoretically, that better information processing in spatiotemporally complex situations leads to better fitness?
• Can we incorporate dynamical signals and long-term tracking of dynamical responses into the analysis?
• Can we look at multivariate behavioral responses to multivariate signals, focusing on active changing of the environment by the organisms?
• What are the mechanisms for variance and time scale adaptation?
• What is the effect of changes in the environment and of natural environments on the animal information processing functions?