Use of Multimodal Neuroimaging Techniques to Examine Age ... · Age, Sex, and Alcohol-Related...

Use of Multimodal Neuroimaging Techniques to Examine Age, Sex, and Alcohol-Related Changes in Brain

Structure Through Adolescence and Young Adulthood

Adolf Pfefferbaum, M.D.Edith V. Sullivan, Ph.D.

Center for Health Sciences, SRI InternationalDepartment of Psychiatry & Behavioral Sciences, Stanford University

School of Medicine

Supported by NIH/NIAAA

American Psychiatric AssociationSan Diego, CA24 May 2017

National Consortium on Alcohol and NeuroDevelopment in Adolescence

Prospective monitoring of brain development in 831 adolescents annually for 5 years to• determine the effects of early, heavy alcohol use on

brain structure and function before drinking onset• 647 no/low drinking• 134 exceeded criteria• Cohort sequential design

age 12-14, 15-17, 18-21 years

Oregon HSU

SRI

UC San Diego

UPitt

Duke

5 U.S. Recruitment Sites

FUNDINGNIAAA, NIDA, NIMH, NICHD

NeuroInformatics

Platform

SRI+Stanford

Oregon HSU

SRI

UC San Diego

UPitt

Duke

5 U.S. Recruitment Sites

FUNDINGNIAAA, NIDA, NIMH, NICHD

NeuroInformatics

Platform

SRI+Stanford

Extending Analysis of Imaging Data

Cortical Myelin

Subcortical Brain Iron

Effects of Initiation of Drinking

Structural MRIQuantitative Measures of Regional Brain Tissue

white matter

gray matter

CSF

Tissue Segmentation

Structural MRI

FreeSurfer Gray Matter Regions

FrontalParietal

TemporalOccipital

Cingulate

Subcorticalwhitematter

Corpus callosum

Pons

SRI24 White Matter Regions

NCANDA 647 No-to-Low Drinking Sample

Sex Ethnicity

Pfefferbaum et al. Cerebral Cortex2016

Source of Sex Difference in Cortical Measures

Age

Cortical Thickness

Cortical Volume

Age

Cortical Surface Area

Age

MaleFemale

Age

Cortical Thickness

Cortical Volume

Age

Cortical Surface Area

Age

MaleFemale

Supratentorial volume

Cortical Thickness

Cortical Volume

Supratentorial volume

Cortical Surface Area

Supratentorial volume

MaleFemale

Age (years)

Age (years)

Regional Cortical Thickness

NCANDA CohortBaseline MRI by Age and Sex

Frontal

Temporal

BoysGirls

Pfefferbaum et al. Cerebral Cortex2016

NeuroInformatics

Platform

Extending Analysis of Imaging DataCortical Myelin

Stiles & Jernigan Neuropsychology Review 2010

Dongjin KwonKilian Pohl

Measuring Cortical Thicknessin No/Low Adolescents

at Baseline

Average Age 21.00

1.78 mm 3.63 mm

Average Age

12.22

Cortical Thickness

1.78 mm 3.63 mm

Cortical thickness (mm)

4

Medial frontal

Anterior temporal

Insular

Occipital

12 year olds

1.78 mm 3.63 mm

Cortical thickness (mm)

3

Medial frontal

Anterior temporal

Insular

Occipital

14 year olds

1.78 mm 3.63 mm

Cortical thickness (mm)

2

Medial frontal

Anterior temporal

Insular

Occipital

16 year olds

1.78 mm 3.63 mm

Cortical thickness (mm)

1

Medial frontal

Anterior temporal

Insular

Occipital

18 year olds

20 year olds

Medial frontal

Anterior temporal

Insular

Occipital

01.78 mm 3.63 mm

Cortical thickness (mm)

Compute Myelinin 226 Adolescents

at Baseline

0.04 0.96%

T1/T2 image

T1 image

T2 image

Register

motor

occipital

sensory

cingulate

Heschl

Age-related Difference in Myelin Content

Average

0.04 0.96%April 4 2017

Developmental Slope

-0.01 0.01% / Age

Average

0.04 0.96%

precentralgyrus

centralsulcus

SignificantAge Effect

D. Kwon et al. OHBM June 27, 2017 Vancouver, Canada

Dongjin KwonKilian Pohl

Age-related Difference in Myelin Content

motor central sulcus

posterior cingulate

NeuroInformatics

Platform

Extending Analysis of Imaging DataSubcortical Brain Iron

Michael De BellisEric Peterson

T2*18-21

Echo-planar Imaging (EPI) fMRI Sequence(T2* Weighted)

Globus pallidus Substantia nigra

Substantia nigraGlobus pallidus

Cerebellar dentate

Susceptibility Weighted Imaging (SWI)(T2* Weighted)

Michael De Bellis

Substantia nigraGlobus pallidus

Cerebellar dentate

Spin-echo Diffusion Tensor Imaging (DTI)(T2 Weighted)

Hallgren and Sourander, Journal of Neurochemistry, 1958

Iron in the Brain

• Non-heme iron in the brain– primary iron deposition not from

bleeding– necessary for dopamine

transmitter function

• Methods to image iron– Field-Dependent Imaging*– Susceptibility Weighted Imaging

(SWI)*– Quantitative Susceptibility Mapping

(QSM)†– R2’ mapping‡

*Pfefferbaum et al. NeuroImage, 2009; *Sullivan et al. Brain Imaging Behav. 2010; †Poynton et al. IEEE Trans Med Imaging,

2015; †Bilgic et al. NeuroImage, 2012; ‡Gelman et al. Radiology 1999; ‡Haacke, et al. MRM 2005

Non-heme iron susceptibility (T2*) signal losstransverse relaxivity (T2) signal loss

Iron effect greater T2 and T2* weighting greater iron effectless T1 weighting greater iron effect

T2* > T2

DTI sequence has higher spatial resolution and less B0 spatial distortion

Estimating Non-heme Iron Concentration fromStandard NCANDA Protocols

12-14 years old

Substantia nigra

15-17 years old

Substantia nigra

T2_1818-21 years old

Substantia nigra

Dentate nucleus

Substantia nigraPutamen

Pallidum Red nucleus

R-squared: Signal vs Age

Estimating Age-related Change in Non-heme Iron Concentration from

Standard NCANDA Protocols

Dentate nucleus

Substantia nigraPutamen

Pallidum Red nucleus

R-squared: Signal vs Age

Estimating Age-related Change in Non-heme Iron Concentration from

Standard NCANDA Protocols

Dentate nucleus

Substantia nigraPutamen

Pallidum Red nucleus

R-squared: Signal vs Age

Estimating Age-related Change in Non-heme Iron Concentration from

Standard NCANDA Protocols

Dentate nucleus

Substantia nigraPutamen

Pallidum Red nucleus

R-squared: Signal vs Age

Estimating Age-related Change in Non-heme Iron Concentration from

Standard NCANDA Protocols

Estimating Non-heme Iron Concentration fromStandard NCANDA Protocols

Dentate nucleus

Substantia nigraPutamen

Pallidum Red nucleus

Dentate nucleus

Red nucleus

Substantia nigra

Putamen

Pallidum

Estimating Non-heme Iron Concentration fromStandard NCANDA Protocols

SWI index (ppm)

R2

estim

ate

from

DTI

Dentate nucleus

Red nucleus

Substantia nigra

Putamen

Pallidum

Extending Analysis of Imaging DataSubcortical Brain Iron

Michael De BellisEric Peterson

It is possible to estimate iron deposition using longitudinal and group DTI or fMRI data

Results are more stable with diffusion-weighted DTI than fMRI

Data can be merged across GE and Siemens scanners

DTI iron estimates correlate well with in-vivo QSMsusceptibility measures

NeuroInformatics

Platform

Extending Analysis of Imaging DataEffects of Initiation of Drinking

NCANDA 2 Year Follow-upRegional cortical volume trajectories in 483 of 647 no-to-low drinking adolescents meeting imaging and drinking criteria followed longitudinally for 2 years

• 65 transitioned into moderate drinking• 62 transitioned into heavy drinking• 356 remained no-to-low drinker• 1423 MRI brain scans

Cahalan et al. CriteriaAverage drinks per

occasion (last 3months): 1-2 1-2 1-2 3-4 3-4 >4Largest # drinks in year: 1-2 3-4 >4 3-4 >4 >4

<1x/year<1x/month

Frequency 1-3x/month4-8x/month>8x/month

Daily

Moderate DrinkerControl

Heavy Drinker (N=62)

(N=356) (N=65)

Volu

me

Frontal Cortical Gray Matter

Age

No/low drinkers

Baseline

adapted from Pfefferbaum et al. Cerebral Cortex 2016

MaleFemale

Volu

me

Frontal Cortical Gray Matter

Age

No/low drinkers

adapted from Pfefferbaum et al. Cerebral Cortex 2016

MaleFemale

Baseline

Corrected for intracranial volume

Volu

me

Frontal Cortical Gray Matter

Age

No/low drinkers

Baseline

adapted from Pfefferbaum et al. Cerebral Cortex 2016

Volu

me

Frontal Cortical Gray Matter

AgePreliminary unpublished data

No/low drinkers

Longitudinal1 year

Volu

me

Frontal Cortical Gray Matter

AgePreliminary unpublished data

Longitudinal2 year

No/low drinkers 356

Regional Gray Matter Volume SlopesDecline with Age

Male (N=180)Female (N=176)

Frontal Parietal

Temporal Occipital

CingulateInsular

Pfefferbaum et al. in review

Volu

me

Central White Matter

Age

No/low drinkers

Baseline

adapted from Pfefferbaum et al. Cerebral Cortex 2016

Volu

me

AgePreliminary unpublished data

No/low drinkers

Longitudinal1 year

Central White Matter

Volu

me

AgePreliminary unpublished data

Longitudinal2 year

No/low drinkers 356

Central White Matter

Regional White Matter Volume SlopesDecelerating Growth with Age

Central white matter

Pons

Corpus callosum

Pfefferbaum et al. in review

Male (N=180)Female (N=176)

Volu

me

Frontal Cortical Gray Matter

AgePreliminary unpublished data

Longitudinal

No/low drinkers 356

Volu

me

Frontal Cortical Gray Matter

AgePreliminary unpublished data

No/low drinkers 356Transitioners 127

Longitudinal

Volu

me

Frontal Cortical Gray Matter

AgePreliminary unpublished data

No/low drinkers 356Moderate drinkers 65

Longitudinal

Volu

me

Frontal Cortical Gray Matter

AgePreliminary unpublished data

No/low drinkers 356Heavy drinkers 62

Longitudinal

Regions where heavy drinkers have significantly steeper reduction in gray matter volume than no-low drinkers

Adj

uste

dVo

lum

e

Occipital

Age

Adj

uste

dVo

lum

eFrontal

Age

Age

Adj

uste

dVo

lum

e

Cingulate

Adj

uste

dVo

lum

e

Total

Age

No/low

Moderate

Heavy

Regional Gray Matter VolumesAccelerated Decline with Initiation of Heavy Drinking

Pfefferbaum et al. in review

No/low

Moderate

Heavy

Frontal Parietal

Temporal Occipital

CingulateInsular

FreeSurfer Parcellated Cortical Regions34 Regions

FreeSurfer Parcellated Cortical Regions34 Regions

FreeSurfer Parcellated Cortical Regions34 Regions

pars opercularis

caudal middle frontal

rostral middle frontal

pars triangularis

superior frontal postcentral

precentral

paracentral

precuneus

cuneus

isthmus cingulate

posterior cingulate

Pfefferbaum et al. in review

Moderate drinkersHeavy drinkers

Slop

e

Maximum Drinks in Past Year

Superior parietal Inferior parietal

Slop

e

Maximum Drinks in Past Year

Precuneus Total parietal

Slop

e

Slop

e

Maximum Drinks in Past Year Maximum Drinks in Past Year

r=-0.208, p=0.0189Rho=-0.217, p=0.0141

r=-0.236, p=0.0076Rho=-0.239, p=0.0068

r=-0.196, p=0.0271Rho=-0.198, p=0.0256

r=-0.203, p=0.0222Rho=-0.204, p=0.0217

Steeper Regional Parietal TrajectoriesCorrelations with Greater Maximum Drinks in Past Year

Oregon HSU

SRI

UC San Diego

UPitt

Duke

5 U.S. Recruitment Sites

FUNDINGNIAAA, NIDA, NIMH, NICHD

NeuroInformatics

Platform

SRI+Stanford

Extending Analysis of Imaging DataCortical Myelin

There is a developmental trajectory of cortical myelin increase and decrease

Subcortical Brain IronThere is a developmental trajectory of subcortical non-heme iron deposition

Effects of Initiation of DrinkingInitiation of heavy drinking alters structural cortical developmental trajectory