The Work of Betty Tuller and Complex Systems

1
The Work of Betty Tuller and Complex Systems

• John Cadwallader
• PSB 4065

2
Complex Systems

• Composed of a large number of simple agents
  interacting with one another according to
  specified rules.
• Share universal patterns.

3
Complex Systems

• Examples
• Molecules in a liquid
• Genes in a cell
• Organisms in an ecosystem
• Buyers/sellers in an economy
• ISPs on the internet
• Neurons in a brain?!?!?

4
Complex Systems

• Share Common Properties
• Phase transitions
• Order parameter(s)
• Control parameter(s)
• Slaving
• Enhanced contrast
• Hysteresis
• Others

5
Phase Transitions

• Rayleigh-Bernard Instability

6
Order and Control Parameters
7
Slaving

• Analogy state of an army can be described
  without reference to any individual soldier. The
  state of every soldier is a slave to the state of
  the army. In understanding the system we dont
  have to worry about every single degree of
  freedom.

8
Hysteresis
9
How does all of this apply to behavior?
10
Speech Categorization Tuller et al. 1994

• How do we determine what a persons utterance
  means?
• Different people say the same word differently.
• The same word can sound different in different
  environments.
• Same-sounding words can have different meanings.

11
Speech Categorization Tuller et al. 1994

• Used a say stay continuum
• Varied the gap between the s and the rest of
  the word.

12
Speech Categorization Tuller et al. 1994

• Order parameter interpretation of stimulus
• Control parameter interval between s and rest of
  word
• Patterns
• Hysteresis
• Enhanced contrast
• Indicative of complex systems-like behavior

13
Speech Categorization Tuller et al. 1994
14
P. Kruse et al 1996

• The perception of multistable visual objects
• Different groups of lights flash alternately

15
P. Kruse et al 1996

• Percepts range from
• No apparent motion
• Alternating motion
• Unidirectional motion
• Bidirectional motion
• All lights appear on simultaneously

16
P. Kruse et al 1996

• Order parameter type of motion observed
• Control parameter interval between lights

17
P. Kruse et al 1996
18
Freeman 1997

• Olfactory coding
• How does olfactory bulb recognize different
  odors?
• No one-to-one correspondence between odorant and
  cells in the olfactory bulb.
• Even lesioning areas of the bulb that are most
  activated by a particular odorant does not
  prevent the animal from recognizing the odorant.

19
Freeman 1997
20
Conclusions

• Complex systems approach offers powerful method
  for understanding macro-scale brain function
• Also offers direct path for simulating brain
  function on computers, something the cognitive
  approach and behavioristic approach do not allow
  (in any straightforward manner)