Issues in Pediatric NeuroimagingIssues in Pediatric Neuroimaging
Kathleen M. Thomas, Ph.D.
Institute of Child Development
University of Minnesota
AcknowledgementsAcknowledgements
B. J. CaseyB. J. CaseyRonald E. DahlRonald E. DahlCharles A. NelsonCharles A. NelsonNeal D. RyanNeal D. RyanPaul J. WhalenPaul J. Whalen
Jane W. CouperusJane W. CouperusJennifer J. Grubba-DerhamJennifer J. Grubba-DerhamRuskin H. HuntRuskin H. HuntLisa RohrerLisa RohrerElise L. TownsendElise L. TownsendAngela TsengAngela TsengNathalie VizuetaNathalie Vizueta
Funding from: NIMH, NINDS, McKnight FoundationFunding from: NIMH, NINDS, McKnight Foundation
Face ProcessingFace Processing
IdentityIdentity EmotionEmotion Direction of gaze (intentions)Direction of gaze (intentions)
The “Specialness” of FacesThe “Specialness” of Faces
Newborn infants prefer to look at face-like objects over other objects.
Johnson & Morton, 1991
Holistic ProcessingHolistic Processing
Inversion EffectsInversion Effects
Inversion EffectsInversion Effects
Facial Emotion Processing in ChildrenFacial Emotion Processing in Children
Infants discriminate positive and negative facial expressions at least as young as 4 months of age
Discrimination of emotions within valence categories appears to take longer to develop, extending well into childhood and even early adolescence
0 45 87 129 171 213 255 300
* *
Time (seconds)
Stimuli
Thomas et al, 2001, Biological Psychiatry
fMRI of Facial Emotion in ChildrenfMRI of Facial Emotion in ChildrenfMRI of Facial Emotion in ChildrenfMRI of Facial Emotion in Children
(8-15 yrs) and adults
Y = -2 Y = -4
Left AmygdalaSubstantia Innominata/
Ventral Pallidum
Y = -2 Y = -4Fear vs. Fixation
Amygdala Response to FearAmygdala Response to Fear
Children Adults
Fear vs. NeutralP
erce
nt
MR
Sig
nal
Ch
ange
in L
eft
Am
ygd
ala
.2
.1
0
-.1
-.2
Group Differences in ActivityGroup Differences in ActivityGroup Differences in ActivityGroup Differences in Activity
Per
cen
t M
R S
ign
al C
han
geIn
Lef
t &
Rig
ht
Am
ygd
ala Fear 2 minus Fear 1 .3
.2
.1
0
-.1
-.2
-.3
Females Males
Gender Differences in HabituationGender Differences in Habituation
SummarySummary
The normal amygdala response to facial expressions differs across development
Ambiguity hypothesis
The amygdala response may differ between males and females
These data do not address whether these changes are due to differences in bottom-up or top-down processing streams
Amaral et al., (2003)
Adult lesions to the amygdala in non-human primates result in a lack of normal fear responses to threatening stimuli
Bilateral Amygdala LesionsBilateral Amygdala Lesions
Adolphs et al., 1999, Neuropsychologia
Face Emotion ProcessingFace Emotion Processing
Brain Damaged Controls
Bilateral Amygdala Lesions
Birbaumer et al. (1998)
Social phobics show an enhanced amygdala response to neutral faces compared to non-phobic controls
The amygdala response to odors (neutral, positive & negative) was no different for social phobics suggesting the effect is specific to face stimuli
The two groups did not differ in their ratings of the valence or aversiveness of the faces or odors
Social PhobiaSocial PhobiaSocial PhobiaSocial Phobia
Rauch et al, 2000, Biological Psychiatry
Masked Fear vs Masked HappyGroup Difference
Post-Traumatic Stress Disorder (PTSD)Post-Traumatic Stress Disorder (PTSD)Post-Traumatic Stress Disorder (PTSD)Post-Traumatic Stress Disorder (PTSD)
DeBellis et al. (2000)
Amygdala Size in Childhood AnxietyAmygdala Size in Childhood AnxietyAmygdala Size in Childhood AnxietyAmygdala Size in Childhood Anxiety
Right Amygdala
0.01
0.05
p value
Thomas et al, 2001, Arch Gen PsychiatryFear vs. Neutral
Anxious ChildrenAnxious ChildrenAnxious ChildrenAnxious ChildrenAnxious children showed differential right amygdala activation compared to non-anxious children
% S
igna
l Ch
ange
in R
. Am
ygda
la
Healthy Anxious Children Children (n=12) (n=12)
0.4
0.2
0
-0.2
-0.4
Relative Response to FearRelative Response to FearRelative Response to FearRelative Response to Fear
Anxious children activated the right amygdala more for fear faces than neutral faces, unlike non-anxious children who showed significantly more activity for neutral faces
Child Reported SCARED Score
% S
igna
l Ch
ange
in R
. Am
ygda
la
•••• ••
•
•
••
-1.0
-.5
0
.5
1.0
1.5
0 10 20 30 40 50 60
r = 0.624p < 0.005
Correlation with Everyday AnxietyCorrelation with Everyday AnxietyCorrelation with Everyday AnxietyCorrelation with Everyday Anxiety
% S
igna
l Ch
ange
in R
. Am
ygda
la
Healthy Anxious Depressed Children Children Children (n=5) (n=5) (n=5)
.6
.3
0
-.3
-.6
Relative Response Relative Response in Anxiety & Depressionin Anxiety & DepressionRelative Response Relative Response in Anxiety & Depressionin Anxiety & Depression
Depressed girls showed no change in the right amygdala for fear faces compared to neutral faces.
Considerations in Pediatric ImagingConsiderations in Pediatric ImagingConsiderations in Pediatric ImagingConsiderations in Pediatric Imaging
Hemodynamic response Spatial normalization Behavioral task Interpreting developmental effects
Does the hemodynamic response differ Does the hemodynamic response differ between adults and children?between adults and children?Does the hemodynamic response differ Does the hemodynamic response differ between adults and children?between adults and children?
Richter, 2003
Kang et al., 2003
The time course and amplitude of the hemodynamic response in a simple visual-motor response task is similar in adults and 7-8 year old children.
Can we use a common stereotaxic Can we use a common stereotaxic space for adult and child data?space for adult and child data?Can we use a common stereotaxic Can we use a common stereotaxic space for adult and child data?space for adult and child data?
Burgund et al, 2002
Individual sulci were located within 4 mm of one another between adults and 7-8 year old children.
This difference is within the resolution of most functional MRI measurements.
Burgund et al, 2002
Interpreting developmental differencesInterpreting developmental differencesInterpreting developmental differencesInterpreting developmental differences
Differences between age groups may include signal intensity (magnitude), extent of activity (volume), direction of effect, relation to behavior
What does the group difference reflect?
Children Adults
1800
1500
1200
900 600
300
0
Vol
ume
of A
ctiv
ity
*
* p < .05Casey et al. 1997
DLPFCChildren
Adults
Group Differences in Magnitude or Extent of ActivityGroup Differences in Magnitude or Extent of Activity
Correlation between Activity & Behavioral PerformanceCorrelation between Activity & Behavioral Performance
Casey et al., 1997
Group Differences in Activity-Behavior RelationshipGroup Differences in Activity-Behavior Relationship
0
5
10
15
20
25
30
0 500 1000 1500 2000
Volume of Orbitofrontal Activation in mm3
Nu
mb
er o
f F
alse
Ala
rms
500 1000 1500 2000 2500
35
0
Perinatal Insult Control40
r = -.28 r = -.41
Percent Change in MR Signal Intensity
Fal
se A
larm
s
Perinatal Insult
0
5
10
15
20
25
30
-1.0 -.60 -.40 -.20 .20 .40 .60 .80 1 2 3 40 0-.80
40
35
45
r = -.77 r = .12 r = .81p < .05 NS p < .0005
Control
Group Differences in Activity-Behavior RelationshipGroup Differences in Activity-Behavior Relationship
N-Back Spatial Working MemoryN-Back Spatial Working MemoryN-Back Spatial Working MemoryN-Back Spatial Working Memory
+
+
+
+
Trial 1
Trial 2
Trial 3
Trial 4
Time
Thomas et al, 1999
1 2 3 4 5
100 -
80 -
60 -
Runs
Per
cent
Acc
urac
y
Adults
Children
Equating Initial PerformanceEquating Initial PerformanceEquating Initial PerformanceEquating Initial Performance
Thomas et al, 1999
Brown et al, 2005
Teasing Apart Age Teasing Apart Age and Performanceand PerformanceTeasing Apart Age Teasing Apart Age and Performanceand Performance
Area is active regardless of age or performance
Activity differs as a function of performance rather than age
Activity differs as a function of age, regardless of performance
% M
R S
ign
al C
ha
ng
e
1 3 5 1 3 5
Ventral Prefrontal Activityduring Go/Nogo Task
Adults Children
number of go trials preceding a nogo trial1 3 5 1 3 5
Behavioral Performanceduring Go/Nogo Task
Adults Children
number of go trials preceding a nogo trial
nogo
nogo
nogo
go
go go go
go go go go go
Parametric ManipulationParametric Manipulation
Continuous Recognition Memory TaskContinuous Recognition Memory Task
New (Concrete)
New (Abstract)
Old (Abstract)
Lag 5
Lag 2
New (Concrete)
New (Abstract)
New (Concrete)
Old (Concrete)
New Old
Duration = 500 ms
ISI = 4000 ms
TR = 2000 ms
Lag 2
Lag 5
New (Concrete)
Old (Concrete)
Medial Temporal Lobe ActivityMedial Temporal Lobe Activity
R L R L
Adults 7-8 yr oldsNew > Old
Concrete > AbstractOld > New
Group Differences in ActivityGroup Differences in ActivityAdults
7-8 yr olds
AbstractConcrete
Adults ChildrenParietal Parietal
NewLag 2Lag 5
Adults ChildrenParietal Right Parietal
Electrophysiological DataElectrophysiological Data
Child
R
R
L
R
R
L
R
R
R
Source ModelSource ModelAdult
NewLag 2Lag 5
Adults ChildrenParietal Right Parietal
Adult Child
Source Timing AnalysesSource Timing Analyses