Transcript Cycle 2: Structure Defines Function

Cycle 2: Structure Defines Function
• Physiologist (function): if we could
document all the connections…would we
understand how they lead to behaviors?
• Anatomist (structure): we can never
discover brain computation without
revealing its basic connectivity.
Cycle 2: Hierarchy of Multiple
Parallel Loops
• Brains move bodies
• Sensors help make better movements
(or no movement)
• Simple sensation-to-action loop.
• Bigger brains ->multiple parallel loops
• Idea: phylogetically newer loops
modify old loops.
• But “age” of structures of given brain is the
same…
• phylogenetic age gives more opportunity
for older structures to be modified
• Modifying inputs of ‘old’ structures could
change the content of their output. (e.g.
Amygdala)
Cycle 2: Hierarchy of Multiple
Parallel Loops
[p.32] …there is nothing in the
physical world by itself that
would predict a priori the
response of a brain to a
stimulus. It is often largely
the state of the brain that
determines the behavioral
outcome…
[yet] we do not know the
neuronal processes
underlying the word “state”.
Cycle 2: Hierarchy of Multiple
Parallel Loops
[p.33] What is meant by
‘calibration’?
Cycle 2: Large-scale Organization
of the Brain Web
• [p.34] the frequency of various brain
oscillators are kept relatively constant
[across species] even as the numbers of
neurons have increased enormously
• RTs differ as do neuronal response
latencies…e.g. 3/5 rule.
Cycle 2: Scaling problems in brains
of various sizes
Cycle 2: Scaling problems in brains
of various sizes
• LOCAL: time to propagate to distant regions
• RANDOM:
Cycle 2: Scale-free systems
• No Bell curve; power law
• No one typical example in the mix
• No qualitative difference in the big middle
and small units
• Tensegrity – provides scale-free
architecture.
Cycle 2: Cortical column
Even in L4, the vast majority of
excitatory synapses arise
locally within cortex
40 %
30 %
<10 %
Thalamus
(20 % long corticocortical
excitatory connections)
Two competing world views:
How is perception shaped?
Empiricism
Feed Forward
Nativism
Feedback