Slide 1

Report
Fundamentals of Neuroscience
Neuroimaging in Cognitive
Neuroscience
James Danckert
PAS 4040 [email protected]
Functional Neuroimaging
• Electrical activity
– Event-related potentials (ERP), visual evoked potentials
(VEP) all derivative from EEG
• Stimulation
– Trans-cranial magnetic stimulation – single vs. rapid pulse
TMS
• Metabolism
– Positron Emission Tomography (PET) and Blood
Oxygenated Level Dependent (BOLD) functional MRI
(fMRI)
EEG
• Large populations of neurons
firing produce electrical
potentials that can be
measured at the scalp
• Signals are passively
conducted through the skull
and scalp and can be
amplified and measured
• Difference between reference
(ground) and recording
electrodes are measured to
give the electrical potential –
electroencephalogram (EEG)
ERPs and VEPs
•
•
•
•
EEG tends to record global brain activity
ERPs (and VEPs) are a special case of EEG
average EEG trace from a large number of trials
align signal to onset of a stimulus or response – hence event-related
potential (ERP)
Pros and cons of ERPs.
Pros
• Good temporal resolution
• Linked to specific physiological markers (e.g., N1, P3
etc. which in turn can be linked to known cognitive
processes)
Cons
• Poor spatial resolution
• Difficult to get at some brain regions (OFC,
temporal cortex)
Transcranial Magnetic Stimulation (TMS)
• Thompson (1910)
placed head between
two coils and stimulated
at ~ 42 Hz
• saw flashing lights –
magnetophosphenes
• was probably
stimulating the retina
and not the visual
cortex
Cowey and Walsh, 2001
TMS
• TMS applies a magnetic pulse to a certain brain region to
temporarily modulate the function of that region
TMS
circular coil
induced current
• the induced current in the tissue is in the opposite direction to that of the coil
• the intensity of the signal drops off towards the centre and outside of the coil
Cowey and Walsh, 2001
TMS
little or
no change
maximum
hyperpolarization
maximum
depolarization
• the flow of the current must cross the axon to cause
stimulation or interruption of function (N3 will not be stimulated)
Cowey and Walsh, 2001
TMS
Spatial extent of TMS
• spatial extent of induced electric field
– drops ~ 75% within 10 mm
– affects 600 mm2 of neural tissue
Rapid vs. Single Pulse TMS
• for single pulse TMS duration of stimulation = 1
msec, but affects motor cortex for up to 100 msec
• for rapid or repetitive pulse TMS stimuli are
delivered in trains with frequencies from 1 to 25 Hz
(1 – 25 times per second)
• duration of after-effects for rapid pulse TMS
anywhere from msec to several seconds
Transcranial magnetic stimulation (TMS / rTMS)
• excitatory or inhibitory reversible effects depending on site
and parameters of stimulation (e.g. frequency of pulses)
-> facilitates or slows down cognitive process/behavior
• when inhibitory, referred to as ‘virtual lesion technique’
• can give precise timing information (msec level) due to
transient nature of effects
• rTMS is beginning to be used as a treatment for depression
(focus is on DLPFC)
TMS
• Poor spatial localisation – how focal is the
stimulation?
• Can’t stimulate certain areas (e.g., temporal lobe)
and can only stimulate cortical surface
• Good temporal resolution
• Can presumably disrupt individual processes within
a task.
• Distance effects – changed interactions due to
stimulation
• Can induce seizures (particularly rTMS)
Frameless stereotaxy and fMRI
• areas can be identified
functionally and then used
to position the coil in a TMS
study using the frameless
stereotaxy method
• Paus is attempting to
directly combine fMRI and
TMS – with TMS pulses
delivered in between fMRI
runs
Metabolic Imaging
• Two main techniques – positron emission
tomography (PET) and functional MRI (fMRI)
• Activity in cells requires energy (oxygen and
glucose)
• Increased neural activity will lead to changes in
cerebral blood volume (CBV), cerebral blood flow
(CBF) and the rate of metabolism of glucose and
oxygen (CRMGl and CRMO)
• These changes in blood flow and metabolism can
be measured using PET and fMRI
Positron Emission Tomography (PET)
• Measures local changes in
cerebral blood flow (CBF) or
volume and can also be used
to trace certain
neurotransmitters (but can
only do one of these at a time)
• Radioactive isotopes are used
as tracers
• The isotopes rapidly decay
emitting positrons
• When the positrons collide
with electrons two photons (or
gamma rays) are emitted
• The two photons travel in
opposite directions allowing
the location of the collision to
be determined
Positron Emission Tomography (PET)
PET and subtraction
• Run two conditions –
stimulation (e.g., look
at visual images) vs.
control (e.g., look at
blank screen)
• Measure the
difference in
activation between
the two images (i.e.,
subtract control from
stimulation)
• This provides a
picture of regional
cerebral blood flow
relative to visual
stimulation.
Motion vs. colour.
• Subject views
coloured screen (left)
vs. moving random
black and white dots
(right)
• Both task activate
early visual areas (V1
and V2)
• Subtracting the two
images reveals
different brain areas
for colour (V4) vs.
motion (V5)
processing
PET vs. fMRI
• PET allows you to track multiple metabolic
processes so long as the emitted photon can be
detected – allows imaging of some
neurotransmitters
• PET is invasive – radioactive isotopes can only be
administered (at experimental levels) every 4 – 5
years
• fMRI has much greater spatial resolution (~ mms)
• fMRI has greater temporal resolution – can detect
activation to stimuli appearing for less than a second
(PET is limited by the half life of the isotope used)
fMRI
Magnet safety
• very strong magnetic fields – even large and heavy
objects can ‘fly’ into the magnet bore
Cerebral blood supply.
• Arterioles
• Capillaries
– Y=95% at rest.
– Y=80% at rest.
– Y=100% during activation.
– Y=90% during activation.
– 25 mm diameter.
– 8 mm diameter.
– <15% blood volume of cortical
– 40% blood volume of cortical
tissue.
tissue.
• Venules
– Primary site of O2 exchange
– Y=60% at rest.
with tissue.
– Y=90% during activation.
– 25-50 mm diameter.
Neurons
– 40% blood volume of cortical
tissue.
Artery
Vein
• Red blood cell
Art erioles
Veneoles
– 6 mm wide and 1-2 mm thick.
– Delivers O2 in form of
Capillaries
oxyhemoglobin.
1 - 2 cm
Transit Time = 2-3 s
Cerebral blood supply.
fMRI
• Deoxyhaemoglobin is paramagnetic
• When neural activity increases more oxygenated
blood than is needed is delivered to the site
• This leads to an imbalance in oxyhaemoglobin and
deoxyhaemoglobin – more oxy than deoxy
• fMRI is able to measure this difference due to the
different magnetic properties of oxy and
deoxyhaemoglobin
fMRI and BOLD
• blood oxygenated
level dependent
(BOLD) signal is
actually a complex
combination of:
– rate of glucose
and oxygen
metabolism
– CBV
– CBF
• same subtraction
logic used in PET is
used in fMRI
fMRI – block design
• fMRI (like PET) began examining brain activity using block designs
colours
rest
colours
motion
rest
rest
rest
fMRI – event-related design
• allows randomization of stimuli (not possible in PET)
fMRI – event-related design
• BOLD response has a predictable form
• In rapid event-related designs the signal to a given trial type is
deconvolved using models of the BOLD response
Linearity of BOLD response
Dale & Buckner, 1997
Linearity:
“Do things really add up?”
red = 2 - 1
green = 3 - 2
Sync each trial response
to start of trial
Not quite linear but good
enough !
Fixed vs. Random Intervals
If trials are jittered,  ITI  power
Source: Burock et al., 1998
fMRI spatial resolution
• images can be co-registered to the subject’s own brain (not an
average brain as in PET)
PET
fMRI
fMRI and topologies
• Using fMRI to “map” different brain functions
Penfield’s maps
Servos et al., 1998
red = wrist; orange = shoulder
Retintopy
EXPANDING RINGS
• 8 Hz flicker (checks reverse contrast 8X/sec)
• good stimulus for driving visual areas
• subjects must maintain fixation (on red dot)
ROTATING WEDGES
Source: Jody Culham
EXPECTED RESPONSE PROFILE OF AREA RESPONDING TO
STIMULUS
To analyze retinotopic
data:
time = 0
Analyze the data with
a set of functions with
the same profile but
different phase
offsets.
time = 20 sec
For any voxels that
show a significant
response to any of
the functions, color
code the activation by
the phase offset that
yielded maximum
activation (e.g.,
maximum response
to foveal stimulus =
red, maximum
response to
peripheral stimulus =
pink)
time = 40 sec
STIMULUS
time = 60 sec
0
20
40
TIME 
60
Source: Jody Culham
Retintopy: Eccentricity
calcarine
sulcus
left occipital
lobe
• foveal area represented at occipital pole
• peripheral regions represented more anteriorly
right occipital
lobe
Retinotopy
Source: Sereno et al., 1995
Other Sensory “-topies”
Audition:
Tonotopy
Sylvian fissure
temporal lobe
cochlea
Saccadotopy
•delayed saccades
•move saccadic target
systematically around the
clock
Source: Sereno et al., 2001
http://kamares.ucsd.edu/~sereno/LIP/both-closeup+stim.mpg
Marty Sereno’s web page
Break
Finding the human homologue of monkey
area X!
• recent research has used monkey neurophys
to guide fMRI in humans
Dukelow et al. 2001
Problems with the search for homologues
•
•
•
•
Absence of activation doesn’t mean the absence of function
Presence of activation doesn’t imply sole locus of function
But our brains are different!
Confirmatory hypotheses
Dukelow et al. 2001
fMRI and diagnosis
• fMRI is starting to be used in patients with
epilepsy
• one goal is to use this as a tool to localise
language, memory etc. prior to surgery
• another goal would be to use fMRI to
study the propogation of seizures
• in stroke patients fMRI can be used to
chart recovery of function
Patient SP – congenital porencephalic cyst
Left
Right
SP - motor strip
motor strip
sequential tapping
% signal change
alternating tapping
3
2
1
0
-11
21
images
41
SP – somatosensory strip
somatosensory
strip
sequential tapping
% signal change
alternating tapping
3
2
1
0
-11
21
images
41
superior
parietal
sequential tapping
6
6.0
% signal c ha ng e
% signal c hange
alternating tapping
4.0
2.0
0.0
-2.0 1
21
images
41
5
4
3
2
1
0
-1
images
Broca’s
area?
name animals
name objects
% sig n a l c h a n g e
non-word sounds (‘ba’)
2
1
0
-1
1
21
images
41
left occipital
name animals
name objects
% sig n a l c h a n g e
non-word sounds (‘ba’)
2
1
0
-1
1
21
images
41
fMRI and cognition
• What not to do – poorly designed tasks!
• What is the right inferior parietal lobe’s
contribution to movement control?
– spatial component of movements
– compare imagined movements with only a spatial
component vs. movements with a sequential
component
spatial only
complex sequence
% signal c hange
Bilateral FEF
1.0
0
-1.0
1
21
41
61
images
81
101
% signal c hange
Supplementary Motor Area (SMA)
1.0
0
-1.0
1
21
41
61
images
81
101
% signal c hange
Bilateral superior parietal
1.0
0
-1.0
1
21
41
61
images
81
101
Inferior parietal cortex
L
R
complex sequence only
Anterior
alternation only
both
Left
Right
Posterior
R
Design Problems
• What could the right vs. left parietal difference be due to?
– Attention? – possibly!
– Differences in eye movements? – maybe!
• Were the tasks really different in the intended way?
– Perhaps – both tasks were spatial in nature and both tasks had a
sequential component, so…
• So did you just test task difficulty?
– Maybe, what of it?
• Why is right parietal more active for less difficult tasks then?
– I don’t know and I don’t care, piss off! I’m gonna start again!
Boy, you must be rich then!!!
Confirming modularity
• Nancy Kanwisher and the parahippocampal
recliner region!
Is that all there is to it?
• Alex Martin and co. suggest that the FFA responds to
other kinds of objects too
• Isabelle Gauthier and co. suggest that it is expertise with
faces which drives the activation
Exploring behaviours
• Prism adaptation ameliorates neglect – how?
• First, explore the direct effects of prism
adaptation in the healthy brain.
• Clower et al 1996 used PET to do this but
reversed the direction of prismatic shift every 5
trials.
Prism Adaptation – Rossetti and colleagues
• prisms shift world
further to the right
(into the patient’s
‘good’ field)
• patient’s movements
compensate for the
prismatic shift – in the
opposite direction
• after effects lead to
better processing of
previously neglected
stimuli
Setup.
targets
fixation
gaze
head
coil
arm brace
scanner bed
rightward
shifting
prism
flexible arm
(mounted to scanner
bed on left side)
Subject moved prism in front of right eye (left eye was patched) prior to prisms run and
moved the prism to the side at the end of the run.
Protocol I.
2 sec
2 sec
0.5 sec
11.5 sec
2 sec
2 sec
5 runs with prisms (50 trials)
5 runs without prisms (50)
Protocol II.
post-target
fixation
fixation target
2 sec volumes – so 2 sec for critical stimulus (the target) and 12 sec
for post stimulus return to baseline (a la Bandettini).
Protocol III.
4T scanner at Robarts
17 pseudo-axial slices
5 mm thick
TR 2 sec
2-shot EPI sequence
Co-registered to 128 slice
anatomical
Adapting in the magnet.
PRE
POST
group mean
group mean
mean shift = 27.73 mm
-50 -40 -30 -20 -10
0
10
20
30
40
50
Finding ROIs.
med frontal (SMA)
left sup par
right sup par
Modeling the peak activation across trials.
Linear
Logarithmic
Left and right superior parietal cortices.
Left Superior Parietal
2.5
max % signal change
r2=0.13
2
1.5
1
0.5
0
-0.5
1
2
3
4
5
6
7
8
9
10
9
10
trial
Right Superior Parietal
2.5
max % signal change
r2=0.15
2
1.5
1
0.5
0
-0.5
1
2
3
4
5
6
trial
7
8
Cerebellar ROIs.
Medial Cerebellum
max % signal change
2.5
r2=0.14
2
1.5
1
0.5
0
-0.5
1
2
3
4
5
6
7
8
9
10
9
10
trial
Right Lateral Cerebellum
max % signal change
2.5
r2=0.42
2
1.5
1
0.5
0
-0.5
1
2
3
4
5
trial
trial
6
7
8
SMA.
max % signal change
(group mean)
transverse
sagittal
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
r2 = 0.37
1
2
3
4
5
6 7
trial
8
9
10
Bottom line?
• Difficult to image the direct effects of adaptation in normals.
• Image “good adaptors” OR change protocol to look at after
effects of adaptation – with all its problems…
Conclusions?
• fMRI should be used for good and not evil!
I wonder if fMRI
could be used to
cure cancer?
Acknowledgements
• fMRI of epilepsy
patient
–
–
–
–
–
–
Stacey Danckert
Seyed Mirsitari
David Carey
Mel Goodale
Ravi Menon
Jody Culham
• fMRI of prism
adaptation
–
–
–
–
Susanne Ferber
Stacey Danckert
Mel Goodale
Yves Rossetti
• fMRI of imagined
movements
– it was all my fault!
End of Lecture

similar documents