fMRI Part 2 Signal detection

1
fMRI - Part 2Signal detection
Magnus Borga
2
History of fMRI
-1990 Ogawa observes BOLD effect with T2
blood vessels became more visible as
blood oxygen decreased -1991 Belliveau
observes first functional images using a
contrast agent -1992 Ogawa et al. and Kwong et
al. publish first functional images using
BOLD signal
3
The BOLD SignalBlood Oxygen Level Dependant
signal
Rest
Activity
4
T2 Effect in fMRI
action
MR signal (S)
rest
TE
t
reception
excitation
5
Mechanism of BOLD Functional MRI
Brain activity
Oxygen consumption
Cerebral blood flow
Oxyhemoglobin Deoxyhemoglobin
Magnetic susceptibility
T2
MRI signal intesity
6
Hemoglobin
Hemoglogin (Hgb) - each globin chain
contains a heme group - each heme group can
attach an oxygen atom (O2) - oxy-Hgb (four
O2) is diamagnetic - deoxy-Hgb is
paramagnetic
Source http//wsrv.clas.virginia.edu/rjh9u/hemog
lob.html, Jorge Jovicich
7
Hemodynamics of the brain
Rest
Activity
Capillary
Artery
Capillary
Vein
Artery
Vein
Oxygen
Oxygen
Oxyhemoglobin
Deoxyhemoglobin
8
The BOLD signal

• The signal S is a sum of intravascular and
  extravascular signal
• S (1-V) Se V Si ,V is the venous blood
  volume fraction
• Make a linear approximation of the variation
  around the resting state (V0)
• DS (1-V0)Se DV Se V0 DSi DVSi
• The relative signal change is then
• DS/S V0 (k1(1-q) k2(1-q/v) k3(1-v))
• where q is the normalized deoxyhemoglobin content
  and v is the normalized volume
• (See appendix in Buxtons article)

9
Buxtons BALLOON model

• Assume an expandable venous compartment (a
  balloon)
• The flow in (Fin(t)) is controlloed by the blood
  flow regulation
• The flow out (Fout(t)) is determined by the
  preasure and resistance
• Fout(t) (P-Pd)/R

10
The BOLD response
11
Paradigm
Activity
Activity
Activity
Rest
Rest
Rest

• Example of activities
• Sensory
• Motor
• Higher cognitive functions, e. g. language,
  math,

12
128 x 128
lt 5
t
200 time frames
13
Locate time-series that are similar to the
paradigm
Paradigm y(t)
High correlation
x(t)
Low correlation
14
Threshold
15
Problem 1
?
Unknown delay between the paradigm and the
measured signal.
16
Solution
Expand the paradigm into a sum of sinusoidal
fuctions
Paradigm
y1(t) and y2(t)
y3(t) and y4(t)
y5(t) and y6(t)
17
Adaptive response-modelling
y(t) wy1y1(t) wy2y2(t) ... wy6y6(t)
?
18
Adaptive response-modelling
19
Problem 2
Low signal-to-noise ration in the time series
20
Solution
Spatial low-pass filtering
Local averaging of the time sequences
21
Local averaging of the time sequences
y
t
x
22
Local averaging of the time sequences
Average
23
Local averaging of the time sequences