Nonverbal Cognitive Impairments
in Fragile X Syndrome: A
Neurocognitive Basis Shared
With Other Developmental
Disorders?
Tony J. Simon Ph.D.
tjsimon@ucdavis.edu
N TRI
The M.I.N.D. Institute
NeuroTherapeutics Research Institute
Dept. of Psychiatry and Behavioral Sciences,
University of California, Davis
Funding R01HD046159, R01HD042974, RL1NS062412
Overview of Syndromes
fragile X (Full Mutation)
Source - Xq27.3 mutation, > 200 CGG repeats
Prevalent VIQ>PIQ pattern
impaired space, time, numbers & “executive” cognition
Turner
Source - X monosomy (45,X)
Prevalent VIQ>PIQ pattern
impaired space, time, numbers & “executive” cognition
2
Overview of Syndromes
Chromosome 22q11.2 Deletion (22q11.2DS/VCFS)
Source - 1.5 - 3Mb chromosome 22q11.2 deletion
Prevalent VIQ>PIQ pattern
impaired space, time, numbers & “executive” cognition
Williams
Source - 7q11.23 deletion
Prevalent VIQ>PIQ pattern
impaired space, time, numbers & “executive” cognition
3
What is the Research Question?
Children from several different developmental disorder
populations have similar impairments in “non verbal
functioning”
visual-spatial, visual-motor, time, numbers, math,
attention
Often called “Nonverbal Learning Disorders” (NLD)
Reasonable(-ish) description, but what is explanation?
is there an underlying brain/mind account?
What follows is a possible explanation & supporting data
4
Population IQ Profiles
140
TD
DS22q11.2
Turner Syndrome
Fragile X
Collapsed WASI, WISC III, WISC IV IQ DATA
22q (N=57), TD (N=45), Turners (N=15), Fragile X (N=15)
120
MEAN Composite Values
100
80
60
114
76
88
72
116
80
98
84
111 78
90
76
112
81
82
40
20
0
FSIQ
VIQ/VCI
PIQ/PRI
INDICES
5
PSI
74
Reaction Time Task
Task: Press the
button as quickly
as you can when
you see the alien
Measures simple
psychomotor speed
6
Reaction Time Task
500
450
400
Average Median Manual RT
Control
DS22q11.2
FXS
TS
428.3
368.6
362.7
milliseconds
350
300
250
200
150
100
50
0
7
410.0
Objects, Space & Numbers
Space and Time are abstract concepts that have “scale” but no
actual values attached to them
we use mental “units” to break them up meaningfully
have to learn “how” much is a(n): inch/second, foot, hour
numbers were invented to describe “how many” units
What if your “mental units” don’t match parts of the real world
accurately? space/time estimates will be wrong, numbers won’t make
sense
this is what I now think explains most of the overlapping
cognitive impairments
If so, we’ll know what to fix and how. We’re about ready to try!
Methods
Experimental Cognitive performance tasks
Attention, Enumeration, Magnitude Judgment
MRI analysis
Structure, Connectivity, Function
Standardized neuropsychological testing
WISC IV, WIAT, Key Math, DKEFS, JLO …..
9
Prefrontal/Parietal Areas
PFC
PPL
Hypotheses
Spatiotemporal/Numerical/Executive impairments all
derive from impairments in “building block” functions
that typically depend on a “frontoparietal” brain network
Posterior parietal lobe (PPL): objects, space, number
PreFrontal Cortex (PFC): working memory, control
H1: lower resolution “mental pictures” increase error
e.g. compare detail in image from 8mP vs. 1mP digital camera
H2: neural network connection differences affect speed
e.g. Drive St. Louis - Chicago on I55 vs. via Indianapolis
All following experiments are designed to test these
11
“Crowding” & Attentional Resolution
From Cavanagh, 2004
If you fix your eyes (attention) on the cross, you cannot
accurately count the bars to the right
“Crowding” & Attentional Resolution
G
F R AGI L E
Reduced Space & Time Resolution
14
Measuring Parts of Space
Task: Press button to choose
who Kermit the Frog is closer to
(Miss Piggy or Fozzie Bear?)
Neither
Fozzie
Piggy
When Kermit is not close to one
end or at the center, error occurs
Measuring Space
90
80
70
60
E
r
r
o
r
50
DS22q11.2
Control
40
30
20
10
Kermit is nearer to?
Spatial resolution reduced by ~30% on this task
20
16
12
8
4
0
-4
-8
2
-1
-1
6
Fozzie
-2
0
0
Piggy
Measuring Time
Duration comparison:
Judge longer of two durations:
400ms vs +/- 10ms diff.
(staircase method)
Auditory & visual
From Debbané et al., 2005
Temporal resolution reduced by ~35% on this task
Quantity, Space & the Brain
Much evidence of spatial representations of magnitude
adults create analog repns of approximate quantity
activations show large network w/PPL focus
“near” quantities hard to distinguish than “far”
infants discriminate large sets w.r.t. Weber law
Predict: Hypergranularity effects largest with small ∂,
Dysconnectivity impairs speed
18
Distance Effect
Diffs = 1 to 7 (excl. 4), attempted balance for large/small n
61
19
Distance Effect Results: Blocks
1400
Distance Effect Task: BLOCKS
22q (N=58), Controls (N=60), TS (N=28), FXS (N=14)
1300
CON-BLX
22q-BLX
TS-BLX
FX-BLX
Adjusted MEDIAN Reaction TIme (ms)
1200
1100
1000
900
800
700
600
500
400
One
Two
Three
Five
Distance
20
Six
Seven
Distance Effect Results: Digits
1600
Distance Effect Task: NUMBERS
22q (N=58), Controls (N=60), TS (N=28), FXS (N=14)
1500
Adjusted MEDIAN Reaction Time (ms)
1400
CON-NUM
22q-NUM
TS-NUM
FX-NUM
1300
1200
1100
1000
900
800
700
600
One
Two
Three
Five
Distance
Six
Seven
Comparing Quantities
Median
Bars
Values
Median
Pitch
230
1800
226
220
1600
210
1574
200
1400
190
215
1574
pitch
Target Size
180
1200
170
167.5
1000
160
150
800
140
median
22q
median
22q
median
median TDTD
STANDARD
STANDARD
130
600
120
400
110
100
200
90
80
0
4
11 3
75 10
19 13
22 15
25 28
40 25
43 27
46 49
61 37
64 39
67 70
82 49
85 88
7 1391611
17 31
1934
213723
29 52
3155
335835
41 73
4376
457947
Step
Steps
Quantity
(space)
resolution
reduced
by ~30% on this task
Sound
(pitch)
resolution
appears
to be identical!
unsure
difference
is comparing
much greater
SoChildren
children with 22q not
justuntil
worse
overall at
values
22
Enumerating, Space & Attention
Quantifying small sets (subitizing) does not require
spatial “search” (shifts of attention) in adults
• PPL not involved (Sathian et al., 1999)
Predict: Hypergranularity effects minimal,
Dysconnectivity effects not expected
Quantifying larger sets requires spatial attention shifts
• PPL (& other areas) are involved
Predict: Hypergranularity effects large, predict undercounts
Dysconnectivity impairs speed also
23
Enumerating & Attention
Enumeration task requires subitizing/counting of
objects (or aliens)
within 2 degrees visual angle
24
Enumerating & Attention
8500
Enumeration Verbal Response: MEDIAN Reaction Time Values
22q (N=36), Control (N=43), TS (N=25), FXS (N=14)
8000
7500
7000
6500
Adjusted MEDIAN RT (ms)
6000
5500
5000
4500
4000
3500
3000
2500
22q RT MD
Con RT MD
TS RT MD
FXS RT MD
2000
1500
1000
500
0
One
Two
Three
Four
Five
Quantity
25
Six
Seven
Eight
Attentional Networks Test - ANT
Single
Congruent
Incongruent
Predict: Hypergranularity effects varying by network tested
Dysconnectivity impairs speed
26
ANT Orienting Results
1200
Attention Networks Task (ANT) Gratton Version: Overall ORIENTING
22q (N=48), ALL Con (N=52), FXS (N=12), TS (N=23)
Con RT
22q RT
FXS RT
TS RT
Median Adjusted RT (ms)
1100
1000
900
800
700
600
Valid
Neutral
Cue Type
27
Invalid
ANT Alerting Results
1200
22q
Control
TS
FXS
ANT Gratton ALERTING INDEX: NEUTRAL - NONE
22q (N=48), ALL Controls (N=52), TS (N=23), FXS (N=12)
MEDIAN Adjusted Reaction Time (ms)
1100
1000
900
800
700
600
NEUTRAL
NONE
CUE TYPE
28
1200
ANT Executive Results
Attention Networks Task (ANT) Gratton Version: Overall EXECUTIVE
MEDIAN Adjusted Reaction Time
22q (N=48), ALL Con (N=52), TS (N=23), FXS (N=12)
MEDIAN Adjusted Reaction Time (ms)
1100
1000
900
800
700
600
CONGRUENT
INCONGRUENT
29 Cue Type
22q
Control
TS
FXS
Diffusion Tensor Imaging
b=0
Tensor Matrix has three
primary diffusivities
• λ1, λ2 & λ3
anisotropy
trace ADC
anisotropy:
dominant
direction
In brain λ1 represents
diffusion parallel to axonal
fibers ... referred to as axial diffusivity
Diffusivities perpendicular to axonal fibers, λ2 & λ3, are
averaged & referred to as radial diffusivity
Fractional Anisotropy (FA) is the fraction of the magnitude of
[the tensor] we can ascribe to anisotropic diffusion
30
22q11.2DS Dysonnectivity
4 common clusters
• parietal L&R SLF
• frontal L&R FOF
FA 22q11.2DS > TD
RD TD > 22q11.2DS
Axial = x (primary)
Reduced cortical
Radial = average(y+z) connectivity in frontoparietal network?
FA = Axial/Total
FA/cognition
correlation
In all clusters:
(spatial attention task)
FA: 22q>TD, p<.0001 22q +ve in R. SLF
RD: TD>FA, p<.0001 TD -ve in R. FOF
31
Common Dysconnectivity Locus
FA: 22q>TD
FA: FXS>TD
FA: TD>TS
Text
R. Inferior Parietal Lobe
32
Common Dysconnectivity Locus!
FA: 22q>TD
FA: FXS>TD
FA: TD>TS
33
Object Tracking Task
34
Brain Activation Differences
22q: 1 > passive
TS: 1 > passive
TD: 1 > passive
Activations
1 target>passive view
FDR p<.01 corrected
t > 3.5
k=100 voxels
22q N=8 90% correct
TS N=7 90% correct
TD N=7 90% correct
FXS N=0 (high error)
NeuroTherapeutics
Now studying entire range of repeats and their “outcomes”
“dosage effect” is evident even in early findings
Suggests that interventions could reduce negative impact
bring cognitive profile back closer to unaffected range?
combination of planned video games & pharmacology?
Summary
Strikingly similar cognitive “endophenotype”
despite differences on standardized measures
Preliminary evidence of similar neural dysconnectivity
fiber tract integrity & “functional network”
Early evidence of common neurogenetic pathway?
developmental convergence on similar “end state” from
several different “start states”
37
Implications & Interventions
Advances in our research provide better understanding of:
“software” running in minds of those with several disorders
“data” that makes software less effective (hypergranularity)
“hardware” that changes typical software & data
If we can explain it, maybe we should be able to “fix” it
Evidence shows that typical spatiotemporal system is “plastic”
i.e. gets better, faster, “sharper” with practice/stimulation
We need to do 2 things (together) to test our explanation
1) carry out new studies to see if we are right
2) build and test intervention video games based on it
38
Collaborators
Jim Gee, Brian Avants, Gary Zhang - Penn
Elliott Beaton, Zhongle Wu,Yukari Takarae, Tracy
DeBoer, Margie Cabaral, Lyndsey Marcelino,
Marisol Mendoza - CABIL
Judy Ross - TJU, Michele Mazzocco - KKI
Paul & Randi Hagerman, Rob Berman, Cam
Carter, Susan Rivera & too many others to
mention in the “NTRI Team” - UC Davis &
beyond
Most of all, YOU! The families, our partners!
39