Download Human Vision: Electrophysiology and Psychophysics

Human Vision:
Electrophysiology
and
Psychophysics
Why study biological
systems?
Pattern recognition
A A A
A A T A
A A A
Artificial neural networks are
used to solve problems in fields
as diverse as
• pattern recognition
• optimisation
• locomotion and spatial recognition
• sequence prediction
In contrast to previous efforts in artificial
intelligence, there is focus not only on the
algorithm necessary for solving a problem,
but also the underlying brain architecture
and physiology that solves this problem.
The brain is made up of neurons
At one time there was a debate as to the question of whether
the brain was composed of a continuous system of wires or
whether it was a discontinuous network made up of individual
neurons
Soma – contains the nucleus
Dendrite – receives input
Axone – sends output
The neuron is a superchip
The neuron is the basic building block for all the
activities in the brain – these activities are as diverse as
• sensory function – vision, audition, touch, smell
• motor functions – muscle contraction
• perception eg
• cognition
For these functions, one neuron can receive upto 100000
contacts and send signals to hundreds of other neurons.
Many levels of complexity in the models
The simplest
The McCullogh Pitts neuron
The most complex model I know of
The multicompartment model of the cerebellar
neuron.
The McCullogh Pitts neuron
1 or 0
1 or 0

a
1 or 0
a = threshold
Output function - Heavyside
step function
The McCullogh Pitts neuron
X1

X2
AND, a=2
Input X1
0
0
1
1
Input X2
0
1
0
1
Output
0
0
0
1
The McCullogh Pitts neuron
X1

X2
OR, a=1
Input X1
0
0
1
1
Input X2
0
1
0
1
Output
0
1
1
1
Shelton DP (1985) Neuroscience 14: 111-31
A model of the cerebellar Purkinge neuron with
1000 compartments
Uptil now we have discussed the conditions of a neuron at
rest. But the neuron is in a network and the information
processing capacities of the neuron lie in its capacity to
communicate with the other neurons.
Uptil now we have discussed mainly passive
conductances ie The movements of current which do
not require the expenditure of energy. The leak current
primarily consisting of K+ dominates this state.
Neuronal communication is via the action potential.
This state is active and is dominated by the Na+ ions.
Two sorts of synapses in the body
Synapses between neurons
Synapse from neuron to muscle
The elements of the network communicate with each other.
This is done by the transmission of chemicals called
neurotransmitters.
Presynaptic neuron
Neurotransmitter (6)
Neuroreceptor (3,4)
Postsynaptic neuron
A neuron sends a message via an action potential
The action potential leads to the
liberation of neurotransmitter
molecules
The action potential
Action potential: Na channels
are open and neurotransmitter
is released in synapse
Threshold: energy
expenditure to
open Na channels
Na channels close
Response to the
presynaptic action potential
The depolarization at the presynaptic terminal leads to
the liberation of neurotransmitters
presynapatic
postsynaptic
The neurotransmitter acts like a key. It opens the neuroreceptor
channel and permits the passage of ions across the synaptic
membrane.
2 types of postsynaptic effects
a) Excitatory
b) Inhibitory
Effects of an excitatory neurotransmitter
Presynaptic excitatory neuron
PRESYNAPTIQUE
Postsynaptic response is positive. The
postsynaptic neuron is depolarized and has
an increased probability of firing an action
potential
POSTSYNAPTIQUE
Excitatory neurotransmitters usually work by opening
postsynaptic Na+ or Ca2+ channels.
Effects of an inhibitory neurotransmitter
Inhibitory presynaptic neuron
Postsynaptic response is
negative. The neuron becomes
more polarized. There is a
decreased probability for the
postsynaptic neuron to fire.
PRESYNAPTIQUE
POSTSYNAPTIQUE
Inhibitory neurotransmitters work by opening K+ or Cl- channels
Spatial and temporal summation 
Each neuron has thousands of synapses. The signals come in
at different times and at different locations over the surface of the
neuron
The integration of signals at different locations is called spatial
summation. The integration at different times is called
temporal summation