Cognitive Reserve
Yaakov Stern
Columbia University
Qu’est-ce que la réserve?
Altérations
cérébrales
Réserve
Manifestations
cliniques
La réserve peut expliquer la disjonction entre
le degré d’altération cérébrale et les
manifestations cliniques de ces lésions.
Active vs. Passive Models
One convenient, although not entirely
accurate, subdivision of reserve models is
whether they envision reserve as a passive
process, or see the brain as actively
attempting to cope with or compensate for
pathology.
Passive Models
• Brain Reserve
• Neuronal Reserve
• Threshold
Threshold Summarized by Satz (1993)
• Hypothetical construct: “Brain Reserve
Capacity” (BRC)
• Concrete examples of BRC: brain size,
neuronal count
• There is individual variation in BRC
• Education is one index of BRC
• There is a critical threshold of BRC. Once
depleted past this critical point, specific
clinical or functional deficits emerge.
Brain Reserve Capacity
Passive, Threshold Model
Lesion
Lesion
Functional
Impairment
Cutoff
Patient 1
Patient 2
Threshold: a passive model of
reserve
• Assumes that there is some fixed cutoff or
threshold at which functional impairment
will occur.
• Quantitative vs. Qualitative
– Assumes that a specific type of brain damage
will have the same effect in each person.
– Does not recognize individual differences in
processing cognitive or functional tasks in the
face of by brain damage.
Active Model of
Reserve
• Relies on the quality of the response -- not
the amount of BRC.
• Relies more on the “software” than the
“hardware”
• Reserve can vary when BRC is held
constant
Brain Reserve Capacity
Active Model
Lesion
Lesion
Functional
Impairment
Patient 1
Patient 2
Cognitive Reserve
• Resilience/plasticity of cognitive networks
in the face of disruption
– Neural Reserve: efficiency/capacity of existing
brain networks
– Neural Compensation: ability to adapt alternate
networks or brain areas
Research Approach
Outcome
Brain Damage
Reserve
Attempt to hold one or more components
constant and allow the others to vary
Brain Reserve: Association Between Head
Circumference and Alzheimer’s Disease
Schofield, et al, 1997
Brain Reserve is Not So Simple
Exercise and environmental stimulation can
activate brain plasticity mechanisms and remodel
neuronal circuitry in the brain.
They can increase vascularization, neuronal
survival and resistance to brain insult, BDNF,
serotonin, dopamine, IGF-1, neurogenesis in the
dentate.
Isn’t more flexible thinking mediated by neurons
and synapses anyway?
Advancing AD Pathology
Initiation
Factors
Promoting
Factors
Diagnosis
Clinical
Symptoms
Appear
Death
Incident Dementia in The Washington
Heights Study
Group
N
Incident
Cases
Relative
Risk
Low Education
264
69
2.02
High Education
318
37
1
Low Occupation
327
71
2.25
High Occupation
201
17
1
95% CI
1.3-3.1
1.3-3.8
Stern et al, JAMA 1994
.
Valenzuela &
Sachdev,
Psychological
Medicine, 2005
Father’s
occupation
0.11
0.25
0.38
Cognition
at 8 years
0.20
0.45
Education
by 26 years
0.36
0.24
Own occupation
at 43 years
0.50
0.13
NART at 53
Richards, JCEN 2003
High Reserve
Low Reserve
Age-related Neural Changes
Literacy and memory decline
in non-demented elders
Manly et al, JCEN 2003
High Reserve (Education)
Low Reserve (Education)
Score at initial visit
AD Neuropathology
Stern et al Neurology 1999
More rapid memory decline in AD patients
with higher educational attainment
SRT Total Recall
20
15
Education Group
10
Low Predicted
High Predicted
Low Actual
5
-1
High Actual
0
1
2
3
4
5
6
Time
Stern et al Neurology 1999
Progression of composite cognitive score
before and after incidence of AD
Scarmeas et al, JNNP, 2005
Bronx Aging Study
Blue indicates less than 7 years education (32 Ss), red indicates 8 to 11 years (64 Ss), and green
indicates 12 or more years education (21 Ss).
Hall, C. B. et al. Neurology 2007;69:1657-1664
Low reserve
MCI
mild AD
High reserve
Diagnostic Threshold
normal
Clinical Severity
Reserve, AD Pathology, and
Clinical Diagnosis
Mild
Moderate
AD Pathology
Stern, JINS 2002
Education and rCBF
Controlling for clinical
disease severity, there is an
inverse relationship
between education and a
functional imaging proxy
for AD pathology
Similar findings have been
noted for occupational
attainment and leisure
activities
Stern et al, Ann Neurol 1992
Education Occupation and rCBF
Stepwise multiple regression
Education:
Predictors of P3 detector flow:
R squared
mMMS, BDRS, age, age at onset, duration .190
+ education
.304
Occupation:
Predictors of P3 detector flow:
R squared
age, mMMS, BDRS, duration, education
.293
+ interpersonal skills
.437
+ physical demands
.515
Stern et al: Ann Neurol 1992; Neurology 1995
Interaction of AD Pathology and Education
Global Cognitive Function
Education * AD path = 0.088, p<.01
22 years
18 years
15 years
Summary Measure of AD Pathology
Bennett DA et al, Neurology 2003
Influence of CR in the association
between WMH severity and cognition
Presented path coefficients:
visual-spatial ability/memory/executive-speed/language.
*p < .05. **p<.01. ***p< .001
The direct relation between
cognitive reserve and WMH in
women was -0.04 (ns).
By including the executive/speed
or the language cognitive variable
in the model, the path coefficient
between cognitive reserve and
WMH volume changed from
negative to significantly positive
(0.23 and 0.46). When cognitive
function is statistically controlled,
women with higher measures of
cognitive reserve had more severe
WMH pathology.
Brickman et al, submitted
Cognitive Reserve, Aging and AD
• Two individuals who appear the same clinically,
whether demented of non-demented, can have
widely divergent levels of underlying age-related
neural changes or AD pathology.
• Thus, the clinical diagnosis of normal aging, MCI
or AD may be accompanied by very minimal
pathology or more than enough to meet
pathological criteria for AD.
• Measuring CR therefore becomes an important
component of diagnosing and characterizing aging
and dementia.
Cognitive Reserve, Aging and AD
• Optimal clinical evaluation of age-related
cognitive change or AD should include:
– A measure of pathology
• age-related atrophy, amyloid imaging
– A measure of an individual’s CR, that is, the ability the
ability to cope with this pathology:
• Proxies for CR such as education or IQ
• fMRI measured expression of “CR networks”
• This type of evaluation is important for
–
–
–
–
early diagnosis and characterization
prognosis
measuring progression over time
assessing of the effect of interventions
Using Functional Imaging to Study CR
• Goal: To understand how cognitive reserve may
be neurally implemented.
– Emphasis on networks mediating CR, not task
performance
• Working hypothesis: CR operates through
individual differences in how tasks are processed
in the brain.
• Basic approach: Challenge participants with a
demanding task and investigate differences in
task-related activation between individuals with
high and low CR.
• Assumption: Because CR modulates most aspects
of cognitive performance in the presence of
pathology, this approach should work with most
demanding tasks.
Outcome
Brain Damage
Reserve
Task-related
network
expression
Task or NP
performance,
Clinical Outcome
Age- or AD-related
pathology
Measured CR or
CR-specific
Network
Scheme for evaluating task-related network
expression in young and old
Neural Reserve
Neural Compensation
Stern, Cognitive Reserve, Neuorpsychologia, 2009
Modified Sternberg Task
1700
1600
1500
RT (ms)
1400
1300
1200
Y
E
1100
1000
900
800
1
3
6
set size
”Load-related” activation: the change in activation as set size
increases
We focus on load-related activation because CR might be more
related to the coping with increases in task demand than to taskspecific features.
Load-dependent Activation During Retention:
Neural Reserve and Neural Compensation
Primary Network
Greater Network Expression
Compensatory Network
Elder
Young
• 2 spatial patterns were expressed
• The primary network was
expressed by both young and old
• Higher expression of the primary
network was associated with
poorer performance,  Neural
Reserve
• The 2nd, “compensatory”
network was expressed primarily
by the elders; higher expression
was associated with poorer
performance
Less Efficient Processing (RT slope)
Zarahn et al., Neurobiol Aging 2007
Network 1
+
Network 2
smaller is better (more efficiency)
-
Local
Density
Slope RT
larger is better (less atrophy)
-
Age
+
Cognitive
Reserve
smaller is better
(faster speed of
processing)
Steffener at al., Brain Imaging and Behavior 2009
A Generalized Neural
Representation Of CR
• CR allows people to better maintain
function in multiple activities and cognitive
domains in the face of brain pathology.
• If a particular brain network subserves CR,
it should be active across tasks with varying
processing demands.
• In other words: CR-related activation may
not be task specific
• Goal: Are there patterns of CR-related brain
activity common to two different tasks?
Generalized Representation Of CR
Strategy
2 activation tasks with
different demands:
- Letter task: Stimuli are
1, 3, or 6 letters
- Shape Task: Stimuli are
1, 2 or 3 shapes
Goal
Can we find common CRrelated activation?
Generalized Representation Of CR
• In younger subjects, a brain
network was identified during the
encoding phase of two different
tasks whose load-related increase
in expression correlated with CR
• Older subjects expressed the
network only in the letter task
• In the context of these tasks, this
network represents a neural
instantiation of CR
• Areas in this CR-related network
have been associated with
executive and control processes
Stern et al., Cerebral Cortex, 2007
Education
WRAT
Occupation
BNT
VLT-Total
VLT- Delayed
VLT- Recognition
Pegboard, dominant
Pegboard, non-dom
Odd-man-out Total
Category Fluency
Speed
u1
u2
u3
u4
u5
u6
u7
u8
u9
u10
u11
u12 u13
Letter Number
Memory
Trails difference
Cognitive
Reserve
Letter Fluency
Picture Vocab
Cognitive Reserve and
Executive Function
.84*
.-.34*
.44*
Executive
Function
u14
u15
Siedlecki et al, JINS, in press
Using interventions studies to test
theories of reserve
Aging/AD Pathology
Clinical Disease
?
Brain
Reserve
Cognitive
Reserve
Conclusions
• Epidemiologic and imaging evidence support the
concept of cognitive reserve
• Reserve is malleable: it is influenced by aspects of
experience in every stage of life
• Cognitive reserve may be mediated by
efficiency/capacity of existing brain networks,
ability to enlist new, compensatory networks, or
“pure” CR-related networks
• The concept of cognitive reserve is applicable to a
wide range of conditions that impact on brain
function at all ages
• Influencing cognitive reserve may delay or reverse
the effects of aging or brain pathology.