Transcript PowerPoint presentation for the lecture

Understanding the brain:
a work in progress
The brain performs an incredible range of functions
• Controls body functions and motivates us to
obtain appropriate resources to maintain life
• Movement
• Detect and interpret sensory information and
social cues
• Attend to specific things rather than others
• Learn and remember information and integrate it
with past knowledge
• Guide behaviour through emotional responses
• Generate conscious awareness of the external
environment, self and others
0.5
5
High speed supercomputers 2000-2010
• 2000 IBM ASCI White 7.226 TFLOPS DoE-Lawrence
Livermore National Laboratory USA
• 2002 NEC Earth Simulator 35.86 TFLOPS Earth
Simulator Center, Japan
• 2004 IBM Blue Gene/L 70.72 TFLOPS DoE/IBM
• 2005 136.8 TFLOPS DoE/U.S. National Nuclear
Security, Lawrence Livermore National Laboratory
280.6 TFLOPS
• 2007/8 478.2 TFLOPS IBM Roadrunner 1.026 PFLOPS
DoE-Los Alamos National Laboratory 1.105 PFLOPS
• 2009 Cray Jaguar 1.759 PFLOPS DoE-Oak Ridge
National Laboratory, USA
IBM Sequoia Supercomputer
20 PFLOPS speed
1.6 PFLOPS memory
318m2
96 racks
7megawatts
Neurons
Neuroglial cells
Astrocytes - anchor neurons to blood
vessels and transport of nutrients/
waste. Have receptors, produce
growth factors and modulate synaptic
transmission. Signal to one another via
gap junctions using calcium.
Microglia - defence against
pathogens and monitor the condition
of neurons.
Ependymal cells - line the fluid-filled
cavities in brain and spinal cord.
Produce, transport, and circulate the
cerebrospinal fluid.
Oligodendrocytes- produce the
myelin sheath in the CNS which
insulates and protects axons.
CREB Pathway
The molecular
brain!
Glu
C
2009
ProteinLounge.com
PresynapticTermi
nal
Ca++
Glu
Hormones/
Neurotransmitters
Glu
Growth
CaCn
Factors
iGLUR
GLUR
(Group-I)
Gb
Ga
SOS
PI
c-Raf
Gb
Gas
ATP
DAG
Ga
Gg
Gb
PKC
IP3
MEKs
P
Ca++
PLC
Calm
P
Gg
Ca++
PKC
PI3K
Akt
Ga
Gg
GRB2
Ca++
cAMP
CamK-II
CamK-II
CamK-IV
Neuron
ERK1/2
P
PKA
p90RSK
P
P
Elk1
SRE
P
S133 S133
C
C
R
R
E
E
B
B
CRE
p300
TBP
TATA
Co-factor
TFIIB
Gene
RNA
Pol-II
Expression
Major subdivisions of the brain
Reticular activating system
Neural plasticity
Learning – turning the gain up and the noise down
Imitating the actions of others (mirror neurons)
Control
Autistic
How is information represented in the brain?
Advantages/disadvantages of spatial encoding
Advantages and disadvantages of temporal encoding
Correlation and pattern changes
A
B
C
D
DABC
pre-stimulus
during stimulus
Inhale
247ms
5 sec
Incidence of
inhale-related sequences
Complexity of
inhale-related sequences
400
Number of
elements in sequences
Number of
sequences detected
16
300
200
100
14
12
10
8
6
4
2
0
0
Prestimulus
During stimulus
Prestimulus
During stimulus
14
Number of
elements in sequences
Number of
Sequences detected
400
300
200
100
0
1
2
3
Odor Concentration
4
13
12
11
10
1
2
3
Odor Concentration
4
Incidence of
inhale-related sequences
Complexity of
inhale-related sequences
400
Number of
elements in sequences
Number of
sequences detected
16
300
200
100
14
12
10
8
6
4
2
0
0
Prestimulus
During stimulus
Prestimulus
During stimulus
14
Number of
elements in sequences
Number of
Sequences detected
400
300
200
100
0
1
2
3
Odor Concentration
4
13
12
11
10
1
2
3
Odor Concentration
4
Combined spatial and temporal encoding
• Most robust solution, allowing brains to be a reasonable
size
• Makes it easier to both separate, integrate and decode
information
The Sensory Brain
Sensory maps - vision
Sensory maps - hearing
Somatosensory and motor maps
The somatosensory homunculus
Integration of sensory information
• Multisensory brain areas
• One sense can influence interpretation of
another one (see a mouth shape the word
“bait” and hear the word “gate”, you think you
hear “date”) – McGurk Illusion
• Facial expressions, even if not consciously
perceived, modify the perception of emotion
in the voice of the speaker
The brain as an interpreter
Illusions
Synaesthesia
Synaesthesia
Synaesthesia
Synaesthesia
We may all start off
experiencing the
world through
synaesthesia
Neural encoding of faces
"Who are you?",
"how do you feel?"
"do i like you"?”
Answers in <300
milliseconds!
Face processing in the brain
Face processing in the brain
Single cell vs population encoding
Quian-Quiroga et al (2005) Nature
Andrews et al
J Neurosci (2010)
The brain as an interpreter
Encoding face identity and face emotion cues simultaneously
Face discrimination learning
Operant discrimination between different faces
Brain rhythms and face recognition learning
30-120Hz
4-8Hz
Coupling between fast and slow oscillations (theta and gamma)
Phase locking between IT neuronal activity and theta
>75% of IT electrodes show coupling between theta phase and gamma amplitude
I 5µV
Correlations between discrimination
performance and altered theta/gamma activity
Neural network models
NL=0.0001
L= 0.00055
NL=0.002
L= 0.0035
Gamma ↑
Theta ↓
Theta ↑
Gamma ↓
Decreased synchronization as theta/gamma ratio increases
Model
Downstream neuron
IT
How desynchronization alone can produce potentiation
Excitatory
neurons
Synch
(1)
De-synch
(2)
Output
(1)
Downstream
neuron
(2)
How desynchronization alone can produce potentiation
Excitatory
neurons
Synch
(1)
De-synch
(2)
Output
(1)
Downstream
neuron
(2)
Excitatory
neurons
How desynchronization alone
can produce potentiation
Decorrelation
reduces
noise
Synch
(1)
De-synch
(2)
Output
(1)
Downstream
neuron
(2)
Decorrelation improves discriminability of
patterns
The problems of consciousness
• There is no single seat of consciousness in the
brain
• Many things are processed without conscious
awareness
• Often similar patterns of brain activation are seen
when information is processed with or without
conscious awareness
• There are different levels of consciousness
• Individuals may be aware even when they show
no obvious signs of consciousness
Assessing conscious awareness in “vegetative state” brain damaged patients
Spatial imagery
Motor imagery
Study found 10% of vegetative state patients could perform
motor/spatial imagery tasks
Monti et al (2010)
New Eng J Med
Using brain imaging to enable vegetative state patients to communicate
Monti et al
(2010)
New Eng J Med
PP
Effects of anaesthesia and sleep on cortical integration
Alkire et al (2008) Science
Effects of deep anaesthesia on cortical processing
Reduced
unidirectional
information flow and
long distance
connections, and
increased short-loop
feedback
How does consciousness emerge?
• Perhaps widespread and integrated flow of
activity in the neocortex generates a
metarepresentation.
• When information is processed unconsciously
a metarepresentation does not form due to
lack of integrated flow between cortical
processing nodes.
Establishing functional connections in the brain using Granger causality
Future progress
• Stronger links between mathematicians,
computer scientists and neuroscientists
• A greater emphasis on revealing key
functional connectivity changes in the brain
• Provide a better understanding of
temporal/patterning aspects of neural
encoding
• Further advances in technologies for
measuring the activity of the working brain