Transcript AGING PRESENTATION

AGING AND ITS EFFECTS
ON THE BRAIN
Prepared by:
Seniha Esen Yuksel
CVIP Lab
August 2004
Why Aging?
 Gray hair, wrinkled skin, changes in bone structure,
discomfort, suffering….
 Is it going go be possible to define when and where
the aging starts?
 Can we develop protective factors to slow down the
changes?
 Can we define the patterns of diseases such as
Alzheimer, Schizophrenia, Multiple Schlerosis,
Alcoholism and AIDS related dementia, and separate
these patterns from the aging process?
2
OUTLINE
 Brain basics
 Effects of aging
 Studies
 Diseases
3
As we get older, we encounter
 Decrease in the total brain weight & volume
 Cortical thinning
 Gyral atrophy
 Widening of sulci
 Expansion of ventricular volume
 Neurological disorders
4
BRAIN BASICS
CSF
 Gray matter is the cortex of the
brain which contains nerve cell
bodies.
 White matter contains lots of
nerve fibers that are sheathed in
a white fatty insulating protein
called myelin.
 Cerebrospinal fluid (CSF) is
the fluid that surrounds the
central nervous system, 100-140
ml in adults
5
Terms we should remember
 Gyrification is the folding of the brain.
 Gyrus is a bump on the cortex (pl: gyri)
 Sulcus is a groove (cut) (pl: sulci).
Lobes:
Frontal: thinking, planning
Parietal: pressure, pain, touch, taste
Occipital: Visual information
Temporal: Hearing, memory
6
Ventricles
If the brain is cut in cross section, there are 4
Cavities in it; the 4 ventricles :
Lateral ventricles, 3rd ventricle and 4th ventricle.
Third ventricle
Fourth ventricle
Images: http://www.epub.org.br/cm/n02/fundamentos/ventriculos_i.htm
7
Ventricles in MRI
8
OUTLINE
 Brain basics
 Effects of aging
 Studies
 Diseases
9
Evolution of Aging Problem
 Until 1984’s, aging is related to the loss of neurons in
the brain.
Ex: 100.000 neuron loss daily resulting in 19.7% loss
at the age of 80 [Brody et al.].
 With the advancements of neuron counting
technology, Terry et al. found out that there is not
much age related neural loss in cortex.
 The small decrease has been explained as the
cortical thinning or as the structural changes in
neurons as they lose their dendritic trees and spines
with age.
10
Evolution of Aging Problem II
 So the studies were concentrated on areas
that are more likely to get affected by the
aging process



Frontal Lobe: Thinking and planning
Temporal Lobe: Hearing and memory
Gray/white matter loss: as the reason for the
brain shrinkage (atrophy) & weight loss


Gray matter is where the information processing
is done.
White matter is the communication channel.
11
More about the gray matter….
 Cerebral cortex is made up of gray matter!
 Gray matter is composed of the neurons.
 Gray matter (GM) is the place where the
actual processing is done.
 Individuals over 60 years old have 17%
lighter brains than of young adults due to the
shrinking of the cerebral cortex, i.e. gray
matter loss!
12
Why does GM or WM losses occur?
 It is believed that the gray matter (GM)
loss occurs due to the decrease in the
size of large neurons (rather than a
notable decrease in the number of
neurons).
 White Matter (WM) loss occurs due to
damage of myelinated fibers with age.
13
What else happens?
As we get older, we encounter:
 Decrease in the total brain weight & volume
 Cortical thinning (Decrease in gray matter)
 Gyral atrophy (Thinning of gyri)
 Widening of sulci
 Expansion of ventricular volume
 Neurological disorders
Image: http://w3.ouhsc.edu/pathology/DeptLabs/pick.htm
14
OUTLINE
 Brain basics
 Effects of aging
 Studies
 Diseases
15
Types of studies
 On the specific parts of the brain such as the
cerebellum and brain stem.
 Studies on the segmentation and quantification of
gray matter and white matter

Gray matter studies:



Gyrification problem: Folding of the brain
Cortical thinning: Distance between gray & white matter
White matter studies
16
Regional Segmentation: Brain Stem &
Cerebellum Analysis (Luft et al.)
 First brainstem is segmented, then
cerebellum is segmented.
 Both segmentations are composed of
3 steps:
 Structural boundaries not defined
by different signal intensities are
manually traced.
 For the structural boundaries of the
brainstem anatomical landmarks
and planes are used to make the
process automatic.
 For the cerebellum, the boundary
was redrawn manually on every
image.
17
Regional Segmentation: Brain Stem &
Cerebellum Analysis (Luft et al.)
 Segmentation cont’d:
contrast-defined boundaries are
automatically segmented using a
region-growing algorithm in three
dimensions
 A 3D lattice is used to subdivide
the cerebellum into 11 regions.
 V1, V2, V3: Vermis 1,2,3
 IH: Lateral Hemisphere
 MH: Medial Hemisphere
 Results: Cerebellum shrinks with age,
brain stem stays stable.
 Automatic regional segmentation is
not easy.

18
Types of studies
 On the specific parts of the brain such as the
cerebellum and brain stem.
 Studies on the segmentation and quantification of
gray matter and white matter

Gray matter studies:



Gyrification problem: Folding of the brain
Cortical thinning: Distance between gray & white matter
White matter studies
19
Problems with GM&WM
Segmentation
 Partial Volume Effect
 Buried Cortex
 Veins and Nerve Fibers
 Inhomogeneities in the magnetic field in MR
images.
 Preservation of the form for surgical purposes
20
Segmentation Problems:
Partial volume effect (PVM)
 Multiple tissues contribute to a
single pixel  called PVM.
 Resulting intensity is the
weighted average of the
different tissues present.
 PVM occurs due to the finite
resolution of the scanner.
Illustration of partial volume effect  It
can be reduced by
a) Ideal Image b) Acquired image
decreasing the voxel size and
by application of sub-voxel
techniques
21
Segmentation Problems:
Buried Cortex
 Buried cortex is the main problem
in measuring the gyrification.
 The edges of gyral crowns touch
one another and the inside is filled
with CSF.
 Surface rendering programs can
distinguish a and b, but they will
fail to detect atrophy if


GM pixels touch one another (c)
there is a deep folding due to high
gyrification (d).
22
Gray Matter White Matter Analysis
Three types of studies on GM-WM analysis:
 Gray Matter Studies: Gyrification and
Thinning of Cerebral Cortex
 GM and WM Segmentation
 WM Segmentation With Tissue Examination
23
GM Studies: Gyrification
(Magnotta et al)
 Folding increases by age to increase surface
area and functional capacity.
 Quantitative measurements of gyrification can
provide important information on aging.
 Magnotta et al.(1999) has the first study to
examine the changes in sulcal and gyral
shape quantitatively.
24
GM Studies: Gyrification
 The main problem in measuring
the gyrification is the 'problem of
buried cortex'.
 Magnotta et al. solved this
problem by introducing erosion to
the entire brain surface so that the
gyral crowns don't touch each
other.
 Then GM and WM are segmented
by fuzzy and sharp classifiers.
Figure: An illustration of a normal brain (A) and an atrophic brain (B), with the sulci and gyri
outlined. Note that in the atrophic brain the gyri, shown in red, are more steeply curved and the
sulci, shown in blue, are more flattened.
25
GM Studies: Gyrification
 In the third step, a cortical
Fig: The surface visualization. The right
image illustrates how sulci are opened up
and ‘buried cortex’ is eliminated. Sulci
shown in blue and gyri in red.
Fig:The effects of retiling the cortical iso-surface.
isosurface is generated
for measurements.
 To preserve the topology,
new vertices are added to
the existing surface by
local retriangulations.
26
GM Studies: Gyrification
 In the fourth step the curvature measure is calculated.
 Curvature measure determines if the triangles are
convex of concave by the following formulation: The
curvature measure is the average
over all j of
2
| ji |cos .e | ji |
i and j: the centers of the neighbor triangles
θ: the angle between the normal to triangle i and the
vector from i to j.
 This formula gives the curvature index. The convex
(positive) values represent gyri, concave (negative)
values represent sulci.
 Finally, the surface area is calculated as the sum of the
areas of the triangles.
 The distance between each triangle and GM/WM
27
interface gives the cortical thickness.
GM Studies: Gyrification
Figure :
sulcal
curvature,
gyral
curvature,
cortical
thickness
vs. age
plots.
 Results:
 Sulcal
curvature index
becomes
increasingly
more negative, reflecting a
flattening and opening up
of the sulci
 Gyral
curvature
index
becomes
increasingly
more positive, reflecting a
narrowing of the gyral
crowns and a sharpening
of their curvature
 And the cortex becomes
progressively thinner.
28
Gray Matter White Matter Analysis
Three types of studies on GM-WM analysis:
 Gray Matter Studies: Gyrification and
Thinning of Cerebral Cortex
 GM and WM Segmentation
 WM Segmentation With Tissue Examination
29
GM Studies: Thinning of Cerebral
Cortex
 Dale et al. introduced the first




Fig: Intersection of the tesselated white
matter surface and pial surfaces with the
skull-stripped MRI volume.

complete, automated procedure to
make cortical analysis.
Sub-cortical regions (extensions of
GM) are cut out deforming an
ellipsoidal template into the shape of
the inner surface of the skull.
White matter is extracted by the use
of connected components
The resulting volume is deformed to
form the GM/WM surface.
But this surface deviates from the
spherical shape so these topological
defects are manually corrected.
Using this method, Salat et al. found
that the global cortical thickness and
global cortical volume declines by
increasing age.
30
GM Studies: Thinning of Cerebral
Cortex II
 Xu et al. , Kruggel et al., Zeng et al. also give automated
methods but they are not applied to the aging problem.
 Zeng applied a level set method with coupled surfaces to
segment both boundaries in a single step.
 Kruggel used marching tetrahedra algorithm to compute the
surface of the WM.
 Marching tetrahedra algorithm is a method where
triangulated surface meshes are used.
 Then he used deformable models to obtain an improved
model of WM.
 He calculated the cortical thickness by projecting a vertex
of GM mesh onto the triangles of WM mesh along its
plane normal.
 Kruggel’s and Zeng’s results are not consistent with
each other.
31
Gray Matter White Matter Analysis
Three types of studies on GM-WM analysis:
 Gray Matter Studies: Gyrification and
Thinning of Cerebral Cortex
 GM and WM Segmentation
 WM Segmentation With Tissue Examination
32
GM & WM Segmentation
 Ge et al. supports both the
GM and the WM contributes
to the brain atrophy.
 GM, WM and intracranial
space volumes were each
identified as individual 3D
fuzzy connected objects
according to their
 Affinity
 Fuzzy adjacency
 Hanging togetherness
Figure :
A) Original dual-echo FSE proton densityweighted image.
B) T2-weighted MR image.
C) Total intracranial volume image.
D) CSF volume image.
E) GM volume image.
33
F) WM volume image
GM & WM Segmentation
 Results:



To predict %WM and %GM as quadratic functions of age,
least-squares regression was implemented
%WM decreases in a quadratic fashion with a greater rate in
the adult midlife
%GM volume loss appears as linear function of age
throughout adult life.
Figure 10: Regression analysis of
fractional brain tissue volume estimates.
Models presented indicate the agerelated volume estimates throughout
adulthood in normal brains.
A) %GM B) %WM C) GM/WM ratio
34
Gray Matter White Matter Analysis
Three types of studies on GM-WM analysis:
 Gray Matter Studies: Gyrification and
Thinning of Cerebral Cortex
 GM and WM Segmentation
 WM Segmentation With Tissue Examination
35
Only WM matters in aging
 Peters et al. explains the conflicts in many of the papers as




segmentation errors & partial volume effect.
This study examines the tissue sections from the same brains.
Partial volume effect is reduced by decreasing the voxel size to 0.7mm
(previous studies used 1-1.5mm).
Results:
 No loss of gray matter (consistent with the fact that there is no or
little neuron loss with age)
 Significant loss of white matter especially in frontal lobes, and a
decrease in ventricular size.
Conclusion:
 Loss in the brain comes from the loss of myelinated nerve fibers.
 These changes in white matter could result in a disconnection
syndrome and contribute to the cognitive decline in normal aging.
 If alterations in myelin and myelinated nerve fibers could be
decreased, than some of the cognitive decline associated with
normal aging can be avoided.
36
Types of studies
 On the specific parts of the brain
 On quantification of gray matter and white matter
 Gray matter studies:
 Gyrification problem: Folding of the brain
 Cortical thinning: Distance between gray & white matter


GM and WM Segmentation
White Matter Segmentation With Tissue Examination
37
OUTLINE
 Brain basics
 Effects of aging
 Studies
 Diseases
38
Brain Disorders / Diseases
 Alzheimer
 Multiple Schlerosis
 Schizophrenia
 Alcoholism and AIDS related dementia
 Corticobasal Degeneration (CBD)
 Progressive Supranuclear Palsy (PSP)
(possible collaboration with medical school)
39
Brain disorders :
Progressive Supranuclear Palsy (PSP)
 PSP is an under-recognized brain disorder.
 Symptoms:
Slowing of movement and
reduced control of walking, balance,
swallowing, speaking and eye movement.
 Typically begin in one’s 60’s
 No effective medication.
 Causes are not known. A brain protein called
tau accumulates in brain cells in the
brainstem causing the cells to die.
40
Brain disorders :
Corticobasal Degeneration (CBD)
 CBD is a progressive neurological disorder.
 Initial symptoms begin around age 60.
 Symptoms:
Poor coordination, absence of
movements, impaired balance, abnormal muscle
postures, cognitive and visual-spatial impairments,
difficulty in speech, difficulty swallowing.
 Characterized by nerve cell loss and atrophy
(shrinkage) of multiple areas of the brain including
the cerebral cortex and the basal ganglia.
 No effective medication.
41
Possible collaboration:
Problem definition
 There is an overlap between PSP and CBD
(Feany et al., 1996), and it would be helpful to
distinguish these two pathologies.
 Are they the opposite ends of the same
disease?
 By comparison, CBD is less common, and
usually displays a more severe cortical
atrophy with fronto-parietal predominance
partly sparing the central area.
42
AGING STILL REMAINS AS AN
UNSOLVED PROBLEM!
Many Studies Have Shown Contradictory
Results Because Of The Difficulties Of
The Problem And Many Studies Have
Evolved In Time With The Advancements
Of The Technologies.
43
CONCLUSION
 Information from MRI images will not give the
answers to these questions…
 BUT IT WILL GIVE A CLUE!!!
44
THANK YOU!
QUESTIONS?
45