Please enable JavaScript.
Coggle requires JavaScript to display documents.
Task 3 (Head motion (Why is head motion related signal variance important…
Task 3
Head motion
Why is head motion related signal variance important
because it produces substantial changes in the timecourses of resting state functional connectivity MRI
causing systematic but spurious correlation structures throughout the brain
head motion could potentially interact with artifacts
a single movement can produce changes in BOLD signal throughout the brain in a regionally specific manner, such that one portion of the brain displays increased signal, whereas another decreased signal
more variance, less power
why would you go through a lot of trouble getting rid of it
Is there a difference in psychiatric patients
probably more effect in older, developping and clinical populations
leads to more variance
more difficult to analyse because different variance between groups
How to decrease motion related effects
Scrubbing
DVARS
how much brain image intensity changes between frames
how much movement
possible regressor
big spike in specific volume
exclude it
2 approaches
excision of entire contaminated frames
parsing artifactual from real signal within frames?
adding motion as confounding
preporcessing
reallign data
6 df transformation
move it back to how it was before
use confound regressors
try to explain additional variance with motion variance
What is the effect of head motion
it shifts the position of brain matter in space
it fundamentally disrupts the establishment of magnetic gradients and BOLD
Often summarized in 1 statistic
can be misleading
increases short distance correlation, decreases long distance
if corrected opposite
Brain damage
Can you still compare functional results?
How to know if data reflects task-relevant differences
What could be confounding factors
Spatial normalization
establish one-to-one correspondence beetween the brains of different individuals
normalizing to standard tamplate
But what if they have brain injury?
can confound/bias normalization
CFM
cost function masking
removes leasions from normalization
How to assess quality
anatomical landmarks
face validity
Unified model gives most precise registration of normal brains
compare normalization of same brain with and without simulated leasion
construct validity
Unified models best
can it predict outcome of functional analyses
predictive validity
Conclusion
unified models are the best
improve quality of normalization for normal and lesion without causing distortions
affene method
linear
basic
Methods :
using brain volume to correct :star:
warping
non-linear method with more DF
study specific template
masking out the lesions
Schizophrenia (manoach)
Inconsistent results regarding hypo/hyperfrontality
Group averaging
may mask heterogeneity
Problems in MRI
Enhances signal to noise
transform brain images into common space
stretching and shrinking
may obscure differences
assumption
general principles of functional brain organization will transcend transformation
May be more variable than healthy subjects
neurodevelopmental abnormalities
Performance differences
Motivation may be lower
use money
Suggestions
modulate regressor by confounding factor
Buckner
Atlas based head size normailization
correcting for head size variation
automated
procedure
Measures
TIV
Manual Total Intracranial volume
reference
ASF
Atlas scaling factor
volume scaling factor required to match individual to atlas target
Should be proportional
used template of young mixed with old
Results
ASF was
Equivalent to manual TIV normalization
Reliable across sessions
able to correct between-gender head size
minimally biased by atrophy
can be used to detect dementia-associated hippocampal volume differences
gender was only factor influencing eTIV
Limitations
accuracy
sometimes you need super precise estimate
Morphometric analysis
but head size differs between people
Sources of error?
stroke
displaced area
loss of signal in area
more anatomical variance in patient group
in lesioned brain there is more difference with reference brain
Can you compare patients and participants in terms of mental effort needed for a task