Significance of White Matter Hyperintensities in MCI

Charles DeCarliUniversity of California at Davis

Alzheimer’s Disease CenterImaging of Dementia and Aging

(IDeA) Laboratory

MCI is Early ADMCI is Early AD

Petersen et al, Arch Neurol, 2001Petersen et al, Arch Neurol, 2001

AD ProcessAD Process

HippocampalInjury

HippocampalInjury MCIMCI ADAD

White Matter Hyperintensities

T1 PD T2

Vascular Risk Factors Predict WMH

Risk Factor P-value

Age <0.0001

Diabetes 0.0113

Heart Disease <0.0001

Systolic Blood Pressure <0.0001

Left Ventricular Hypertrophy

0.0001

Treated Hypertension <0.0001

Jeerakathil, et al, Stroke, 2004Jeerakathil, et al, Stroke, 2004

WMH Quantification

White Matter Hyperintensities and MCI

White Matter Hyperintensities and MCI

Predictor VariablePredictor Variable Adjusted RR andAdjusted RR and95% CL95% CL

AgeAge 1.18 [1.03,1.35]1.18 [1.03,1.35]

WMHWMH 5.34 [1.80,15.9]5.34 [1.80,15.9]

ApoE4ApoE4 3.54 [1.43,8.78]3.54 [1.43,8.78]

Diastolic BPDiastolic BP 1.70 [1.07,2.71]1.70 [1.07,2.71]

DeCarli et al, Arch Neurol, 2001DeCarli et al, Arch Neurol, 2001

• DeCarli et al, Arch Neurol, 2001

• Wu et al, Neurology, 2002

• Lopez et al, Arch Neurol, 2003

• Petkov et al, J Int Neuropsychol Soc, 2004

WMH are Associated with Episodic Memory DeficitsWMH are Associated with Episodic Memory Deficits

AD ProcessAD Process

HippocampalInjury

HippocampalInjury

MCIMCI

CVDCVD

White MatterInjury

White MatterInjury

ExecutiveControlDeficit

ExecutiveControlDeficit

Role of AD and WMHIn MCI

Role of AD and WMHIn MCI

Different Mechanisms of Episodic Memory Failure in MCI

IDeA Lab

Christine Wu Nordahl, Charan Ranganath, Andrew Yonelinas, Charles DeCarli,Bruce Reed,William J. Jagust

Neuropsychologia 43(11) 1688-1697; 2005

         

0

5

10

15

20

25

30

35

40

Cou

nt

.04 .06 .08 .1 .12 .14 .16 .18 .2 .22

Normalized hippocampal volume

.126 = 25th percentile

0

10

20

30

40

50

60

70

80

Cou

nt-20 0 20 40 60 80 100

WMH volume

19.375 = 75th percentile

White matter hyperintensity volume

aMCI SubtypesaMCI Subtypes

Based on SALSA dataset, n=122

MCI-HA MCI-WMH

Controls MCI-HA MCI-WMH

n 20 11 11

Age 78.65 (6.34) 74.55 (5.65) 77.64 (3.56)

Education 15.6 (2.79) 15.8 (3.46) 13.5 (1.51)

Gender (M/F) 5/15 4/7 5/6

MMSE 29.63 (0.49) 27.46 (1.81)* 27.27(2.45)*

Left HC .150 (.03) .102 (.03)* .148 (.02)

Right HC .152 (.02) .107 (.03)* .147 (.01)

WMH load 13.15 (15.185) 7.66 (2.93) 34.38 (12.2)*

Hypertension 47% 45% 82%

Type II Diabetes 12% 0% 27%

Subject Characteristics

* differs from other groups, p < .05

Behavioral Tasks

1. Episodic Memory Task: Object-Color Association

study18 red/18 green

delay~60m

testred or green?

MCI-HA and MCI-WMH are equally impaired on the color association episodic memory task

.4

.5

.6

.7

.8

.9

Acc

ura

cy

episodic memory

* *

* Differs from controls and MCI-HA p < .01

MCI-WMHMCI-HAcontrol

A D + B R

+ b

study delay2 sec

probe

Behavioral Tasks

2. Working Memory Task: Item Recognition2, 4, 6 - item verbal

4 - item spatial

.4

.5

.6

.7

.8

.9

1A

ccu

racy

2-item verbal 4-item verbal 6-item verbal 4-item spatial

MCI-WMHMCI-HAcontrol

MCI-WMH are impaired on the item recognition task

* * ***

* Differs from controls and MCI-HA p < .01** Differs from controls p < .01, Differs from MCI-HA p = .08

Behavioral Tasks

3. Working Memory Task: N-back- 1-back- 2-back

SS

XB

x t

T

…

1-back target

1-back target

2-back target

MCI-WMH are also impaired on the n-back working memory tasks

SS

XB

x t

T

…

* Differs from controls and MCI-HA p < .01

.4

.5

.6

.7

.8

.9

1

Acc

ura

cy

1-back 2-back

MCI-WMMCI-HCcontrol

* *

MCI-WMHMCI-HAcontrol

Summary of Behavioral Data

• MCI-HC and MCI-WMH are equally impaired on the episodic memory task

• MCI-WMH are impaired on working memory, both in simple maintenance as well as a more complicated task involving manipulation and maintenance

White Matter Changes Compromise Prefrontal Cortex Function in Healthy Elderly

IDeA Lab

Christine Wu Nordahl, Charan Ranganath, Andrew Yonelinas, Charles DeCarli,Evan Fletcher,William J. Jagust

In Press: Journal Cognitive Neuroscience

Behavioral Tasks

Episodic Retrieval Study (prior to scanning)

18 red/18 green

delay~60m

Test (during scanning)red or green?

A D + B R

b

study delay2 sec

probe

+ Working memory:

Low and high loads

Working memory activations

• Bilateral dorsal prefrontal cortex activations

• Bilateral ventral prefrontal cortex activations

Left Dorsal PFC Right Dorsal PFC

-2

0

2

4

6

8

10

12

-2 -1.5 -1 -.5 0 .5 1 1.5

WMH severity

r = -. 718p = .004

-10123456789

10

-2 -1.5 -1 -.5 0 .5 1 1.5

WMH severity

WMH severity

-10123456789

-2 -1.5 -1 -.5 0 .5 1 1.5

Anterior Cingulate

Working memory task:Dorsal prefrontal cortex function is negatively

correlated with WMH severity

Consistent with evidence suggesting that dorsal PFC is more affected in aging

Left Dorsal PFC Right Dorsal PFC

-2

0

2

4

6

8

10

12

-2 -1.5 -1 -.5 0 .5 1 1.5

WMH severity

r = -. 718p = .004

-10123456789

10

-2 -1.5 -1 -.5 0 .5 1 1.5

WMH severity

WMH severity

-10123456789

-2 -1.5 -1 -.5 0 .5 1 1.5

Anterior Cingulate

Anterior cingulate activity is associated with cognitive control (ie. ability to guide thought and action in accordance with internal intentions)

Functional connectivity between anterior cingulate and PFC mediates successful task performance (Kondo et al 2004)

Working memory task:Anterior cingulate cortex function is negatively

correlated with WMH severity

Summary of fMRI Data

• WMH a presumed indicator of cerebrovascular brain injury is associated with working memory impairment

• Executive control processes are likely involved and related to amount of WMH

Disconnection of Working Memory Processes by White Matter Hyperintensities

Adriane Mayda

Graduate Student, IDeA Lab

University of California at Davis

Disconnection Hypothesis

VascularFactorsVascularFactors

White MatterInjury

White MatterInjury

WMHWMH Cognitive Deficits

Cognitive Deficits??

FA FA

Anisotropy Mapping

Location of WMH in MCI

Normal Controls (n=30)Normal Controls (n=30) MCI (n=29)MCI (n=29)

Location of WMH and ROI

A

A

P

PR

A P

Correlations

R2 = 0.20

0.82

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.98

1

1.02

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8

Log WMH (%TCV)

Ac

cc

ura

cy

y

R2 = 0.62

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.98

1

1.02

200 250 300 350 400

Fractional Anisotropy

Ac

cu

rac

y

Total WMHTotal WMH FA within WMHFA within WMH

Summary

• aMCI has at least 2 subtypes– Hi WMH– Small Hippocampi

• Amnesia in MCI-WMH results from– Impairments in executive control– Reduced prefrontal activation– Disconnection from posterior targets

Conclusions

• WMH are common to the elderly

• WMH alone can be associated with episodic memory impairments

• WMH likely contribute to susceptibility to late life cognitive impairment and dementia

• WMH are potentially modifiable

• Adriane Mayda• Christine Nordahl• Evan Fletcher• Owen Carmichael• IDeA Lab

Supported by NIH: P30 AG10129, R01 AG09029,R01 AG021028, R01 NS 29993,P01 AG12435, P01 AG0027232,R01 AG111101, R01 AG08122,R01 AG16495, U01 AG024904

http://neuroscience.ucdavis.edu/idealab/