Outline

• Brief Overview of Neuroimaging Modalities

– EEG / MEG
– PET
– fMRI
– fNIRS
• Cognitive Neuroscience study of deductive
reasoning
 Basics
• Structural vs Functional Imaging
• Precision in imaging
– Temporal Resolution
– Spatial Resolution
• Relative changes rather than absolute measurement
– Difference in brain activity b/w baseline (control) condition vs
task condition
• Indicators of brain activity - Basic brain physiology
– Neural activity generates electric potentials and magnetic fields
– Neuronal activation requires energy
• Oxygen cannot be stored by neurons
• Oxygen is supplied by the vascular system
• Brain consumes 20% of the oxygen in the blood stream
 What is Functional Brain Imaging
• The use of technology to determine how the
brain functions during a particular task
• Done while the subject is performing a specific
task
– i.e., linguistic, motor, sensory, attention
• One method of getting evidence for large-scale
localization of cortical functions
– The other major method: lesion studies
– Additional methods
• Intraoperative mapping
• Transcranial magnetic stimulation (TMS)
 Functional Brain Imaging Tools
• Electroencephalography (EEG)
• Magnetoencephalography (MEG)
• Positron Emission Tomography (PET)
• Functional Magnetic Resonance Imaging
(fMRI)
• Functional Near-Infrared Spectrescopy (fNIRS)
 EEG/MEG - Basic Logic
• Electrical activity of neurons produces currents
spreading through the head.
• These currents reach the surface of the scalp, in the
form of voltage changes and magnetic fields, both of
which can be measured non-invasively.
• Measured voltage changes at the scalp are called the
electroencephologram (EEG).
• Measured magnetic fields at the scalp are called the
magnetoencephologram (MEG).
 Electrical Activity in the Cortex
• Action Potential • Excitatory postsynaptic
– Presynaptic potential (EPSP)
– Axonal – Postsynaptic
– Generally notmeasurable – Dendritic
by EEG/MEG – Measurable by EEG/MEG
• Apical dendrites are
oriented in parallel along
the cortical sheet
• Apical dendrites thought to
contribute strongest signals
measurable with EEG/MEG
• Axons are more randomly
located, resulting in
currents from presynaptic
action potentials cancelling
each other out
• Postsynaptic electrical
activity (EPSP) sums,
creating large “dipole”
 EEG and MEG Signals
• EPSPs of parallel dendrites
in cortical columns
creates:
– Primary current (what we
want to know about)
– Secondary/volume currents
• Measured by EEG
• Influenced by intervening
tissue
– Magnetic field
perpendicular to primary
current
• Measured by MEG
• Unaffected by intervening
tissue
 These cells are well-aligned and tend to fire in synch

EEG signal reflects the summation of the synchronous activity of thousands

or millions of neurons that have similar spatial orientation
 Spatial Resolution
• Single synapse on dendrite contributes ~20 fA-m
(femto ampere meter = 10-15 = one quadrillionth)
• Empirical observations suggest EEG/MEG signals are
typically ~ 10 nA-m (nano = 10-9 = one millionth)
• Therefore, typical EEG/MEG signals reflect summed
activity of ~ 500,000 – 1,000,000 neurons
• ~ 1-5 mm2 of cortex forms lower bound of spatial
resolution
• In practice, the inverse problem further limits the
ability to spatially pinpoint EEG/MEG signals.
 MEG Signals
• MEG measures the fluctuations of frequency
(Hz) and amplitude (T) of the brain magnetic
signal
• 10 fT (10-15) to about several pT (10-12)
• Earth’s magnetic field ~ .5 mT
• Requires:
– Preposterously sensitive magnetometer (SQUID)
– Shielding from external noise
 The SQUID
• Superconducting Quantum
Interference Device (SQUID)
 Pros and Cons of EEG & MEG
EEG MEG
Signal magnitude Large signal (10 mV), easy to Tiny signal (10 fT), difficult
detect to detect

Cost Cheap Expensive

What does signal index? Measures secondary (volume) Measures fields generated by
currents primary currents
Signal purity Affected by skull, scalp, etc. Unaffected by skull, scalp,
etc.
Temporal Resolution ~ 1 ms ~ 1 ms

Spatial Localization ~ 1 cm ~ 1 mm

Experimental Flexibility Allows some movement Requires complete stillness

Dipole Orientation Sensitive to tangential and Sensitive only to tangential

radial dipoles dipoles
 ERP Mapping
• ERP – event related potentials
• Traditional analysis: ERP waveforms at certain electrode
positions
• ERP mapping attempts to determine points in time when map
configurations change and/or when they differ between
experimental conditions
• Relies on the fact that, whenever the spatial configuration of
the electric field on the scalp differs, different neuronal
populations are active in the brain, reflecting an alteration of
the functional state of the brain

Christoph M. Michel, Margitta Seeck and Theodor Landis, Spatiotemporal

Dynamics of Human Cognition News Physiol. Sci 1999 Oct, 14:206-214
 Neuroimaging Based on Cerebral
Blood Flow
• Besides emitting electric potentials or
magnetic fields, neuronal activity also requires
energy
• Energy is supplied by metabolizing glucose
• This requires oxygen which is supplied by the
vascular system
• So tracking blood flow and molecules involved
with oxygen transport may give us clues about
the level of neuronal activity
Neural Activity & Blood Flow
• Neurons consume energy (glucose)
when activated
• Oxygen is required to metabolize the
glucose
• As clusters of neurons are activated,
there is an increased need for oxygen in
that area
• Oxygen is transported to neural tissue
via oxy-hemoglobin in the blood
• The oxygen exchange occurs in the
capillary beds
• As oxy-hemoglobin gives up oxygen to
the neural tissue, it is transformed into
deoxygenated hemoglobin
Oxy-Hb and deoxy-Hb are correlates of brain activity through
oxygen consumption by neurons
David J. Heeger & David Ress. (2002)What does fMRI tell us about neuronal activity? Nature Reviews Neuroscience 3, 142-151
Hemodynamic Response
 Vascular Density

Why bother with arteries, capillaries?

Vascular density in the brain is proportional to synaptic density
• Imaging techniques monitoring markers
associated with blood flow in the brain
– fMRI
– fNIRS
– PET
Positron Emission Tomography
(PET)
 PET Machine

http://www.radiologyinfo.org/content/petomography.htm
http://en.wikipedia.org/wiki/Positron_Emission_Tomography
 Positron Emission Tomography (PET)

• Measures the distribution of particular organic molecules and

compounds (e.g., water, glucose, neurotransmitters) in the
brain
• The organic molecules and compounds are not detectable
because they do not emit electromagnetic signals
• Positron-emitting isotopes of these organic molecules and
compounds are introduced into the blood invasively (e.g.
H2O15)
• After a short time period, the isotopes are dispersed
throughout the brain
 Positron Emission Tomography (PET)

• These isotopes, along with the blood, flow to the areas of the
brain with the highest metabolic needs
• These areas are assumed to be the most active at the given
point in time
• The positrons in the isotopes interact with electrons, and
produce photons which can be detected at the surface of the
head
• The greater the activation of an area, the more photons
originate from that area
 Positron Emission Tomography (PET)
• Tomography is accomplished by computer using
sophisticated algorithms
• The final PET images show areas of different hues, each
hue representing a different degree of activation of the
underlying brain structures
• The final PET images are superimposed on a structural
image of the brain (MRI or CT scan)
 PET Images

http://encarta.msn.com/media_461519549_761555359_-1_1/Positron_Emission_Tomography.html
 Some properties of PET
• Spatial resolution: 5-10 mm
• How good is that?
– Under one sq mm of cortical surface
• 130,000 neurons
• 1400 minicolumns (at est. avg. 93 neurons/col)
• Temporal resolution: “…on the order of
minutes…” (Papanicolaou, 1998, p. 14)
Functional Magnetic Resonance
Imaging (fMRI)
 MRI Machine

http://www.radiologyinfo.org/photocat/photos.cfm?Image=philips5.jpg&&subcategory=Brain
 Another MRI Machine

http://www.radiologyinfo.org/photocat/photos.cfm?Image=hitachi.jpg&&subcategory=Brain
 It all starts with hydrogen…

magnetic dipole moment (MDM)

MDMs align with / against uniform

magnetic field Bo

small % more align with Bo  net

magnetization
So what goes on inside the scanner?
• Place hydrogen nuclei (brain) in a uniform magnetic field
(scanner)

• Apply radiofrequency (RF) pulse, normally at 90° to

magnetic field. This ‘tips’ MDMs of hydrogen nuclei

• MDMs of hydrogen nuclei are now

at 90° to main field (z), i.e. in x,y plane

• Terminate RF pulse and let nuclei relax: MDMs

return to original (z) orientation; energy released
during relaxation is measured by receiver coil
 Phase
• BEFORE the RF pulse:
MDMs are not in phase

• DURING the RF pulse:

MDMs in phase  strong signal in x,y plane

• AFTER the RF pulse:

MDMs begin to dephase  signal decays

• Two reasons for this dephasing:

• inhomogeneities in the magnetic field
• ‘spin-spin’ interactions between neighbouring nuclei
 Time to relax…
• time course for MDMs to return to original
(z) orientation

• time course of breakdown of magnetization

in the x,y plane due to spin-spin interactions

• Different tissues have different T1 and T2

relaxation rates

• time course of breakdown of magnetization

T1 relaxation in the x,y plane due to inhomogeneities in
T2 relaxation the magnetic field
T2* relaxation
BOLD Blood Oxygenation Level Dependent
 We measure O2 ratios in blood

• BOLD (Blood Oxygenation Level Dependent) contrast: measures

inhomogeneities in magnetic field due to changes in the level of
O2 in the blood

• deoxyhaemoglobin = magnetic! oxyhaemoglobin = not magnetic!

So we can use the change in fMRI signal

to infer the relative oxygenation of the blood!

High ratio deoxy:oxygenated blood Low ratio deoxy:oxygenated blood

 fast decrease in MRI signal  slow decrease in MRI signal
 Summary of the physics
• MDMs of hydrogen nuclei align to magnetic field in scanner
• RF pulse causes them to spin, in phase, in x,y plane
• Once pulse has stopped they fall back to direction of magnetic
field, dephasing as they do so
• Dephasing takes various amount of time, depending in part on
inhomogeneities in magnetic field
• Inhomogeneities are caused by variable ratio of deoxygenated
: oxygenated blood
• Assumption: activity in brain area lowers this ratio and
thereby decreases speed of decay of MRI signal
 Functional Magnetic Resonance
Imaging (fMRI)
• Temporal resolution: not very specific
– Image reflects the increase in oxygenated blood 5 to
8 seconds after the neurons fire
• Spatial resolution: good (order of mm)
– However, it is unclear whether the imaged area is
precisely the area involved in the activity
– The flow of oxygenated blood into the depleted
area may also flow into neighboring vessels in areas
where neural firing did not occur
BOLD and electro-physiology: correlation
Local field potential (LFP), measured with an implanted electrode

Same area in V1 of cat, Kim DS (2004)

BOLD and electro-physiology: correlation

Single unit recording, cat V1, Kim 2004

Active area
 Area that “lights up”
(hypothetical example)
fMRI Example

http://www.fmrib.ox.ac.uk/fmri_intro/fusion.gif
Functional Near-Infrared Spectroscopy
(fNIRS)
 fNIR: An Emerging Modality
• Emerging technology
• Monitors hemodynamic changes
• Relatively good spatial resolution (compared to EEG)
• Good temporal resolution (compared to PET, fMRI)
• Safe, affordable, portable, non-invasive, minimally
intrusive, rugged
• Relatively easy to integrate with other modalities

Optical
Imaging
 Principles of fNIRS
• Photons that enter the tissue undergoes two types of
interaction:
– Scattering
• mainly causes the photons to change their direction of motion
• due to cell membranes, tissue boundaries
• assumed to be constant
– Absorption
• causes the photons to lose their energy to the medium
• due to chromophores in the tissue i.e. Hb,HbO2,water
• can be changing depending on the changes in the concentrations of
the chromophores

Light Propagation in Tissue

Courtesy of Christopher H. Contag, Stanford University
 Principles of fNIRS (cont.)
Optical Window in Tissue • Tissue is transparent to light within the near
Infrared (NIR) range
– hence NIR can penetrate through human tissue.

• Main chromophores (absorbers) within the NIR light

range are oxygenated and deoxygenated
hemoglobin as compared to other tissue
Near IR components such as water, lipid, melanin, etc.

• Oxy and deoxyHb have distinct absorption

characteristics within the NIR range which allows
spectroscopic measurements
• Prefrontal cortex is activated during cognitive tasks

• By monitoring the attenuation changes due to the

changes in absorption, concentration changes of
oxy-Hb and deoxy-Hb can be obtained, hence it is
possible to obtain information on:
– Oxygen consumption
– Blood volume changes
 Principles of fNIRS (cont.)

H Obrig , R Wenzel, M Kohl, S Horst, P Wobst, J Steinbrink, F Thomas, A Villringer. Near-infrared spectroscopy: does it function in functional activation studies
of the adult brain? International Journal of Psychophysiology 35(2-3):125-142, 2000
The fNIR System
• The fNIR device provides relative
change in hemoglobin levels,
calculated using modified Beer-
Lambert law.
– Oxygenated hemoglobin change:
delta O2Hb (µmol/L)
– Deoxygenated hemoglobin change:
delta HHb (µmol/L)
– Total hemoglobin change:
delta cHb (µmol/L)
Modified Beer-Lambert Law
I in
OD  log( )    .c.d .DPF  G
I out

Having the same Iin at two different instances

I rest
OD  log( )   HB .c HB .d .DPF   HBO2 .c HBO2 .d .DPF
I test
Measuring at two different wavelengths

 OD1   1 .d .DPF  1 2 .d .DPF   c HB 

HB HBO

 OD    HB   HBO2 
 2  
  2 .d . DPF  HBO2
2 .d . DPF   c 
F

Can be solved for Concentrations for non-singular F matrix

Drexel Continuous Wave fNIRS Sytem

Cognitive Optical Brain Imager (COBI) Studio

 Monitoring Areas
Dorsolateral Prefrontal Cortex
Brodman’s Areas 10, 9, 45,46

Right Left
 Experiment Setup
Data
Acquisition
Computer

Real-time
16 Channel Data

Control Signals
fNIR data
(transmitted through TCP/IP)

fNIR
Sensor

Protocol
Computer
Data Collection – Raw Signals
Data Collection - Oxygenation
 Signal Processing
• The physiologically irrelevant data (such as respiration and heart pulsation effects)
and equipment noise, etc. is first eliminated from the raw fNIR measurements by
using a low-pass filter with a cut-off frequency of 0.14Hz
 Summary of fNIRS
• fNIR provides brain imaging that is:
– Safe
– Portable and deployable in the field
– Wearable & wireless
– Rugged
– Integration with EEG and other cognitive state
measures

• fNIR results are consistent with fMRI results

A practical consideration: Cost

• Most expensive: MEG

– About $2 million for the machine
– $1 million for magnetically shielded room
• Next most expensive: PET
• Next: fMRI
• Cheapest: EEG and fNIRS
 Supplementary Materials
• Neural Correlates of Reasoning
– Cambridge Handbook Chapter 20 by Goel
• Neuroimaging basics
– Oxford Handbook Chapter 6 by Morrison &
Knowlton
– Supplementary: Glimcher Chap 5,6