Transcript Week 2 Lecture Notes

Brain and Cognition Week 2
Parts of Neurons
Dendrites and spines
Synapses
Action potentials
Synaptic transmission
1
The Neuron
2
Dendrites
Dendrites of a single neuron =
dendritic tree.
Dendrites are covered with
thousands of synapses (see
previous picture).
Postsynaptic membrane (part of
dendrite) contains receptors.
Many (not all) dendrites are covered
in spines.
3D reconstruction of dendritic spines
3
Spines
Electron micrograph of a synapse on
a dendritic spine
Light microscope image and 3-D reconstruction of spines on a
neuron in the hippocampus
4
Spines
Copyright © 2002 Cell Press.
Neuron, Vol 35, 1019-1027, September 2002
Spine Motility: Phenomenology, Mechanisms, and Function
Tobias Bonhoeffer and Rafael Yuste
Max Planck Institut für Neurobiologie, Martinsried, Munich, Germany; Department of Biological Sciences, Columbia University, New York.
5
Spines
In live neurons, spines are constantly in moving and changing shape.
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
Videos from Maria Fischer, Stefanie Kaech, Darko Knutti &
Andrew Matus, 1998, "Rapid Actin-Based Plasticity in Dendritic
Spines". Neuron 20, p. 847-854.
6
The Neuron
Neurons are polarized cells:
Dendrite (post-synaptic)
Soma (cell body)
Axon (pre-synaptic)
Axons and dendrites contain quite
different sets of molecules and
organelles.
Sizes of neuronal components:
Cell body
20 m
Nucleus
5-10 m
Axonal diameter <1m-25m
Synapse
800 nm
Spines
200-500 nm
Synaptic Vesicle 50 nm
7
Classifying Neurons
pyramidal neuron (cortex)
Purkinje neuron (cerebellum)
8
An Engineering Approach
9
Neuronal Morphology
Neuronal shape is highly variable and complex!
CA3
CA1
DG 10
Glia
Astrocytes: involved in regulating extracellular space around
neurons, regeneration
Myelinating Glia: oligodendrocytes, nerve conduction
There are more glial cells than
neurons in the brain …
11
The Membrane Potential
Cell membrane, extracellular and intracellular space
12
The Membrane Potential
13
The Potassium Channel
The Structure of the Potassium Channel: Molecular Basis of K+ Conduction
and Selectivity
Declan A. Doyle, João Morais Cabral, Richard A. Pfuetzner, Anling Kuo,
Jacqueline M. Gulbis, Steven L. Cohen, Brian T. Chait, Roderick MacKinnon –
Science 280, 5360, 1998.
14
The Potassium Channel
15
The Resting Potential
Ek = (RT/zF) loge([K+]o/[K+]i)
Where T is temp in kevin, z is valence (+ is 1) F
is faraday's constant, [K+]o and [K+]i are
concentrations outside and inside the axon.
Ek = 59.8 log10(20/400) = -75 mV
K+
Most important:
Na+
There is more K+ inside the
cell than outside (20×)
Ca++
There is less Na+ inside the
cell than outside (10×)
Cl-
16
The Resting Potential
The voltage difference across
a neuronal membrane is called
the membrane potential.
The membrane potential is
measured by using a
microelectrode.
At rest, the membrane
potential of a typical neuron is
about:
-65 mV (millivolts)
This is called the resting
potential of the cell.
17
Passive Decay of the
Membrane Potential
18
Action Potential
19
Recording Action Potentials
20
The Patch Clamp Method
21
The patch clamp consists of an electrode inside a glass pipette. The pipette, which
contains a salt solution resembling the fluid normally found within the cell, is lowered to
the cell membrane where a tight seal is formed. When a little suction is applied to the
pipette, the "patch" of membrane within the pipette ruptures, permitting access to the
whole cell. The electrode, which is connected to specialized circuitry, can then be used
to measure the currents passing through the ion channels of the cell. Furthermore, we
can use our electrical circuitry to "clamp" the membrane potential to any voltage that we
desire: very handy when measuring the activity of voltage-dependent channels.
Recording Action Potentials
Intracellular recording: measures the potential difference between the tip of
the electrode (inside the cell, i.e. intracellular) and another electrode outside
the cell (ground).
Extracellular recording: electrode is close to the cell, but does not impale 23
it.
The Shape of the Action Potential
≈ 2 ms
Phases of the action potential:
≈ 1 ms
Rising phase → Overshoot → Falling phase → Undershoot
24
Triggering the Action Potential
What causes the start of an action potential?
Analogy: pressing the shutter button on a camera
Action potentials are caused by depolarization of the
membrane beyond a threshold.
This depolarization can be the result of:
Na+ entering the cell after a neurotransmitter has
been released by another neuron
Injecting current through a microelectrode in an
experiment
After the depolarization crosses the threshold, the action
potential unfolds automatically, “all-or-none”.
25
Modeling the Action Potential
In the 1950s Alan Hodgkin and Andrew Huxley proposed an
ionic model for the action potential and conducted experiments
to test it. They received the Nobel Prize in 1963.
(a) The first action potential ever recorded (from squid giant axon).
(b) Voltages and conductances according to Hodgkin/Huxley, 1952.
26
The Chemical Synapse
27
Chemical Neurotransmission
“At rest”, the synapse (presynaptic side) contains numerous synaptic
vesicles filled with neurotransmitter, intracellular calcium levels are very
low (1).
Arrival of an action potential: voltage-gated calcium channels open,
calcium enters the synapse (2).
Calcium triggers exocytosis and release of neurotransmitter (3).
Vesicle is recycled by endocytosis (4).
28
Chemical Neurotransmission
Once released, the neurotransmitter molecules diffuse across the
synaptic cleft (about 20-50 nm wide).
When they “arrive” at the postsynaptic membrane, they bind to
neurotransmitter receptors (“lock-and-key” mechanism).
Two main classes of receptors:
Transmitter-gated ion channels
G-protein-coupled receptors
Transmitter-gated ion channels: transmitter molecules bind on
the outside, cause the channel to open and become permeable to
either Na+ (depolarizing, excitatory effect) or Cl– (hyperpolarizing,
inhibitory effect).
G-protein-coupled receptors have slower, longer-lasting and
diverse postsynaptic effects. They can have effects that change
an entire cell’s metabolism.
29
Excitatory Effects of Neurotransmitters
EPSP =
Excitatory
PostSynaptic
Potential
30
Inhibitory Effects of Neurotransmitters
IPSP =
Inhibitory
PostSynaptic
Potential
31
Integration of Synaptic Inputs
In the CNS, many EPSP’s are needed to generate
an AP in a single neuron.
A single EPSP has, in general, very little effect on
the state of a neuron (this makes computational
sense).
On average, the dendrite of a cortical pyramidal
cell receives ~10000 synaptic contacts, of which
several hundred to a thousand are active at any
given time.
The adding together of many EPSP’s in both space
and time is called synaptic integration.
32
Synaptic Integration
(a)Single input → single EPSP.
(b)Three APs arriving simultaneously
at different parts of the dendrite
add together to produce a larger
response (spatial summation).
(c)Three APs arriving in quick
succession in the same fiber can
also result in a larger response
(temporal summation).
33
Integration of Synaptic Inputs
Distal and proximal synaptic inputs:
34
Synaptic Plasticity
Synaptic efficacy (strength) is changing with time.
Many of these changes are activity-dependent, i.e.
the magnitude and direction of change depend on
the activity of pre- and post-synaptic neuron.
Some of the mechanisms involved:
-
Changes in the amount of neurotransmitter released.
Biophysical changes in ion channels.
Morphological alterations of spines or dendritic branches.
Modulatory action of other transmitters.
Changes in gene transcription.
Synaptic loss or sprouting.
35
Hebb’s Postulate
“When an axon of cell A is near enough to
excite a cell B and repeatedly and persistently
takes part in firing it, some growth process or
metabolic change takes place in one or both
cells such that A’s efficiency, as one of the cells
firing B, is increased.”
Donald Hebb, “Organization of Behavior”, 1949
36
Animal Models of Plasticity
Long-Term Potentiation (LTP)
Cross-section of the hippocampus:
Cajal’s drawing
37
Animal Models of Plasticity
Brain slice preparation of the hippocampus:
38
LTP
Typical LTP experiment: record from cell in hippocampus
area CA1 (receives Schaffer collaterals from area CA3). In
addition, stimulate two sets of input fibers.
39
LTP
Typical LTP experiment:
record EPSP’s in CA1 cells
(magnitude)
Step 1: weakly stimulate input
1 to establish baseline
Step 2: give strong stimulus
(tetanus) in same fibers
(arrow)
Step 3: continue weak
stimulation to record increased
responses
Step 4: throughout, check for
responses in control fibers
(input 2)
40
LTP
LTP is input specific.
LTP is long-lasting (hours, days, weeks).
LTP results when synaptic stimulation coincides with
postsynaptic depolarization (achieved by cooperativity of
many coactive synapses during tetanus).
The timing of the postsynaptic response relative to the
synaptic inputs is critical.
LTP has Hebbian characteristics (“what fires together wires
together”, or, in this case, connects together more
strongly).
LTP may produce synaptic “sprouting”.
41
The NMDA Receptor
(a)At the resting potential
(postsynaptic neuron),
glutamate binds to the NMDA
channel, the channel opens,
but is “plugged” by a
magnesium ion (Mg2+).
(b)Depolarization of the
postsynaptic membrane
relieves the magnesium block
and the channel open to
allow passage of sodium,
potassium and calcium.
42
The Associative Nature of LTP
Old(er) view: Associative requirement is
mediated by the voltage-dependent
characteristics of the NMDA receptor.
New discovery (1994): Active conductances in
dendrites mediate back-propagation of AP’s into
the dendritic tree.
43
Spike-Timing Dependent Plasticity
Basic Idea: Change in synaptic strength
depends on the precise temporal difference
between pre- and post-synaptic neuronal
firing (causality!).
44
The Neuron:
Integrator or Coincidence Detector?
Synchronous inputs really matter!
45
Data Analysis in Neurophysiology
Spike train
data sets:
Neuron in MT
Colby and
Duhamel, 1991
46
Data Analysis in Neurophysiology
Neuron in IT (object selective)
Desimone et al., 1984
47
Data Analysis in Neurophysiology
Neurons in V1 (orientation selective)
PSTH (firing rate)
Cross-Correlation
Auto-Correlation
Shift Predictor
Engel et al., 1991
48
Neural Coding
Rate coding versus temporal coding
One major mechanism of how neurons encode information is
through their firing rate (number of AP’s per second). –
Example: orientation selectivity.
Another major mechanism is synchronization (AP’s occurring
together in time). – Example: perceptual grouping.
Synchrony could affect other neurons (e.g. through spatial
summation – see unit 1).
49
Computational Neuroscience
Components of (most) neural models:
-
Units and connections
Inputs and outputs
Activation function
Learning rule
50
The McCullogh-Pitts Neuron
51
The McCullogh-Pitts Neuron
52
“Why the Mind is in the Head”
“Why is the mind in the head? Because there,
and only there, are hosts of possible
connections to be formed as time and
circumstance demand. Each new connection,
serves to set the stage for others yet to
come and better fitted to adapt us to the
world, for through the cortex pass the
greatest inverse feedbacks whose function
is the purposive life of the human intellect.”
Warren S. McCullogh, Hixon Symposium 1951.
53
54
55
gyrus
sulcus
Cortical Anatomy
Motor Cortex
Broca’s
Area
Representation of Imaging Data Sets:
- 3-D based
- Surface based
Auditory
Cortex
Visual Cortex
56
Cortical Anatomy
The macaque monkey cortex - unfolded.
Felleman and Van Essen, 1991.
57
Cortical Anatomy
58
“Brain Flattening”
59
Source: Marty Sereno, UCSD
60
Source: Marty Sereno, UCSD
61
“Brain Averaging”
62
Standardized Brain Atlas
63
Cortical Anatomy
macaque, chimpanzee, human
64
Cortical Anatomy
65
Methods of Cognitive Neuroscience
Neurobiology:
Neuroanatomy
Neurophysiology
Neuroimaging Techniques
PET
MRI/fMRI
EEG
MEG
Evidence from Dysfunction
Lesions
Diseases of the CNS
Cognitive Psychology
Computational Approaches
66
Methods of Cognitive Neuroscience
67
Neurobiology
Neuroanatomy and neurophysiology
are often conducted in animal
systems (monkey, cat, etc.).
Neuroanatomy:
-
Large-scale anatomy of the brain
Subdivisions of the cortex
Neuronal subtypes, layers
Morphology of single neurons
68
Neurobiology
Neurophysiology:
- Recording of neural activity, often
in the context of a stimulus or task
(single-cell recording, local field
potential recording, multi-electrode
recording).
- Electrical stimulation of neurons to
study their role in perception or
movement (micro-stimulation,
surface stimulation)
69
Methods Comparison
Method Space
Time
Neural Correlate
------------------------------------------------------------PET
coarse
coarse
“brain activation”, metabolic
(5 mm) (sec.)
rate of tissue, incorporation
of glucose, oxygen utiliz.,
receptor distribution, 2D-3D
capable.
fMRI
coarse
(2 mm)
coarse
(1 sec.)
“blood oxygenation state”,
“regional blood flow”,
oxygenated/deoxygenated
hemoglobin, 2D-3D, in
register with struct. Scan.
EEG
coarse
(?)
fine
(msec)
“electrical (field) potentials”,
surface electrodes arranged
on skull, limited depth, inverse
problem, unknown source
locations, allows correlation
measures.
MEG
coarse
(?)
fine
(msec)
“magnetic (field) potentials”,
SQUIDs arranged around head,
sensitive to noise, similar
advantages and drawbacks
as EEG.
70
A New Method…
Transcranial Magnetic Stimulation (TMS)
71
EEG
The electroencephalogram (EEG) measures the activity of
large numbers (populations) of neurons.
First recorded by Hans Berger in 1929.
EEG recordings are noninvasive, painless, do not interfere
much with a human subject’s ability to move or perceive
stimuli, are relatively low-cost.
Electrodes measure voltage-differences at the scalp in the
microvolt (μV) range.
Voltage-traces are recorded with millisecond resolution –
great advantage over brain imaging (fMRI or PET).
72
EEG
Standard placements of electrodes on the human scalp: A, auricle; C, central;
F, frontal; Fp, frontal pole; O, occipital; P, parietal; T, temporal.
73
EEG
74
EEG
75
EEG
Many neurons need to sum their activity in order to be detected by EEG
electrodes. The timing of their activity is crucial. Synchronized neural activity
produces larger signals.
76
The Electroencephalogram
A simple circuit to generate rhythmic activity
77
The Electroencephalogram
Two ways of generating synchronicity:
a) pacemaker; b) mutual coordination
1600 oscillators (excitatory cells)
un-coordinated
coordinated
78
EEG
EEG potentials are good indicators of global brain
state. They often display rhythmic patterns at
characteristic frequencies
79
EEG
EEG suffers from poor current source localization
and the “inverse problem”
80
EEG
EEG rhythms correlate with patterns of behavior (level of attentiveness,
sleeping, waking, seizures, coma).
Rhythms occur in distinct frequency ranges:
Gamma:
20-60 Hz (“cognitive” frequency band)
Beta:
14-20 Hz (activated cortex)
Alpha:
8-13 Hz (quiet waking)
Theta:
4-7 Hz (sleep stages)
Delta:
less than 4 Hz (sleep stages, especially “deep sleep”)
Higher frequencies: active processing, relatively de-synchronized
activity (alert wakefulness, dream sleep).
Lower frequencies: strongly synchronized activity (nondreaming sleep,
coma).
81
EEG
Power spectrum:
82
EEG - ERP
ERP’s are obtained after averaging EEG signals
obtained over multiple trials (trials are aligned by
stimulus onset).
83
MEG
The MEG laboratory
Images courtesy of CTF Systems Inc.
84
MEG
Measures changes in magnetic fields that accompany
electrical activity.
85
MEG
An example (auditory task):
86
87
88
89
MEG
Viktor Jirsa (FAU): http://www.ccs.fau.edu/~jirsa/Imaging.html
Three task conditions: MEG results
1
2
3
1 -- Listening to tones that were delivered with a delay of about 5s. A random
time was added to prevent stimulus prediction. The signal is an average
over about 80 stimulus presentations.
2 -- Reacting to acoustic stimuli. The same stimulus presentation as in (1) but
now the subject was told to press on an air cushion as soon as possible
after the tone was heard.
3 -- Synchronizing with a rhythm. Here the tones were presented regularly
with a frequency of 1 Hz. The subject was told to press the air cushion 90
in
synchrony with the stimulus.
PET
Positron Emission Tomography
Requires the injection of a positron-emitting
radioactive isotope (tracer)
Examples:
C-11 Glucose analogs (metabolism)
O-15 water (blood flow or volume)
C-11 or O-15 carbon monoxide
PET tracers must have short half-life, e.g.
C-11 (20 min.), O-15 (2 min.). Cyclotron!
Positron + electron  2 gamma ray beams.
Gamma radiation is detected by ring of
detectors, source is plotted in 2-D producing
an image slice.
91
PET
92
PET - Examples
In cognitive studies, a subtraction
paradigm is often used.
93
PET - Examples
Another example of control and task
states, and of averaging over subjects:
Marc Raichle
94
PET - Examples
(a) Passive viewing of nouns; (b) Hearing of nouns;
(c ) Spoken nouns minus viewed or heard nouns;
(d) Generating verbs.
M. Raichle
95
PET - Examples
M. Raichle
96
PET - Examples
PET images taken at different times,
e.g. during learning, can be compared.
M. Raichle
97
PET - Examples
PET images are pretty to look at ...
… and can be combined with other
imaging modalities, here MRI.
98
Functional Magnetic Resonance Imaging
Typical MRI Scanner
99
MRI - fMRI
The Physics (sort of) ...
Subjects are placed in a strong external magnetic field.
Spin axes of nuclei orient within the field. External RF pulse
is applied. Spin axes reorient, then relax. During
relaxation time, nuclei send out pulses, which differ
depending on the microenvironment (e.g. water/fat ratio).
fMRI – functional MRI
Allows fast acquisition of a complete image slice in as little
as 20 ms. Several slices are acquired in rapid succession
and the data is examined for statistical differences.
Hemoglobin is “brighter” than deoxyhemoglobin.
Oxygenated blood is “brighter” - active areas are “brighter”.
BOLD-fMRI
100
PET - MRI in Comparison
101
fMRI - Examples
102
Stimulus:
“checkerboard
pattern”
V1 responses
Jezzard/Friston
1994
103
104
fMRI High-Resolution Mapping
BOLD fMRI in cat cortex (level of
area 18)
Kim et al., 2000
105
fMRI High-Resolution Mapping
Kim et al., 2000
106
PET and fMRI Similarities and Differences
- Different biological signal. Yet, both pick up a signal
related to bulk metabolism (not electricity).
- fMRI has better temporal (<100 ms) and spatial
resolution (1 mm and less)
- fMRI does not involve radioactive tracers and subjects
can be measured repeatedly, over many trials.
- PET images generally represent “idealized averages”.
fMRI images are often registered with structural
scans to show individual anatomy.
- For both, images can be aligned for multiple subjects.
- fMRI is widely available, PET is not.
- fMRI does not allow localization of neurotransmitters or
receptors etc.
- For both, it can be tricky to get stimuli to the subject.
107
Data Analysis Issues
Neuroimaging (PET/fMRI): Activation values, spatial
resolution, averaging, image alignment and
registration.
EEG/MEG: Current source localization (inverse
problem), time domain data sets, frequency power
spectrum, correlation and coherency.
108