View
216
Download
0
Category
Tags:
Preview:
Citation preview
fMRI Analysiswith emphasis on the General Linear Model
http://www.fmri4newbies.com/
Last Update: January 18, 2012Last Course: Psychology 9223, W2010, University of Western Ontario
Jody CulhamBrain and Mind Institute
Department of PsychologyUniversity of Western Ontario
What data do we start with
• 12 slices * 64 voxels x 64 voxels = 49,152 voxels• Each voxel has 136 time points (volumes)• Therefore, for each run, we have 6.7 million data points• We often have several runs for each experiment
…
These #s are from an obsolete scanner. With a modern 3T, we can get 3X the slices
Why do we need stats?• We could, in principle, analyze data by voxel surfing: move the cursor
over different areas and see if any of the time courses look interesting
Slice 9, Voxel 1, 0Slice 9, Voxel 0, 0
Even where there’s no brain, there’s noise
Slice 9, Voxel 9, 27Here’s a voxel that responds well whenever there’s visual stimulation
Slice 9, Voxel 13, 41Here’s one that responds well whenever there’s intact objects
Slice 9, Voxel 14, 42
Here’s a couple that sort of show the right pattern but is it “real”?
Slice 9, Voxel 18, 36
Slice 9, Voxel 22, 7
The signal is much higher where there is brain, but there’s still noise
Why do we need stats?
• Clearly voxel surfing isn’t a viable option. We’d have to do it 49,152 times in this data set and it would require a lot of subjective decisions about whether activation was real
• This is why we need statistics
• Statistics:– tell us where to look for activation that is related to our
paradigm– help us decide how likely it is that activation is “real”
The lies and damned lies come in when you write the manuscript
Predicted Responses• fMRI is based on the Blood Oxygenation Level Dependent (BOLD) response• It takes about 5 sec for the blood to catch up with the brain• We can model the predicted activation in one of two ways:
1. shift the boxcar by approximately 5 seconds (2 images x 2 seconds/image = 4 sec, close enough)
2. convolve the boxcar with the hemodynamic response to model the shape of the true function as well as the delay
PREDICTED ACTIVATION IN VISUAL AREA
BOXCAR
SHIFTED
CONVOLVEDWITH HRF
PREDICTED ACTIVATION IN OBJECT AREA
Types of Errors
Slide modified from Duke course
Is the region truly active?
Doe
s ou
r st
at t
est
indi
cate
th
at t
he r
egio
n is
act
ive?
Yes
No
Yes No
HIT Type I Error
Type II Error
Correct Rejection
p value:probability of a Type I error
e.g., p <.05
“There is less than a 5% probability that a voxel our stats have declared as “active” is in reality NOT active
Statistical Approaches in a Nutshellt-tests
• compare activation levels between two conditions
• use a time-shift to account for hemodynamic lag
correlations
• model activation and see whether any areas show a similar pattern
Fourier analysis
• Do a Fourier analysis to see if there is energy at your paradigm frequency
Fourier analysis images from Huettel, Song & McCarthy, 2004, Functional Magnetic Resonance Imaging
Effect of Thresholds
r = 00% of variance
p < 1
r = .246% of variance
p < .05
r = .5025% of variance
p < .000001
r = .4016% of variance
p < .000001
r = .8064% of variance
p < 10-33
Complications
• Not only is it hard to determine what’s real, but there are all sorts of statistical problems
r = .246% of variance
p < .05
What’s wrong with these data?
Potential problems
1. data may be contaminated by artifacts (e.g., head motion, breathing artifacts)
2. .05 * 49,152 = 2457 “significant” voxels by chance alone
3. many assumptions of statistics (adjacent voxels uncorrelated with each other; adjacent time points uncorrelated with one another) are false
The General Linear Model (GLM)GLM definition from Huettel et al.:• a class of statistical tests that assume that the
experimental data are composed of the linear combination of different model factors, along with uncorrelated noise
• Model– statistical model
• Linear– things add up sensibly (1+1 = 2)
• note that linearity refers to the predictors in the model and not necessarily the BOLD signal
• General– many simpler statistical procedures such as
correlations, t-tests and ANOVAs are subsumed by the GLM
Benefits of the GLM• GLM is an overarching tool that can do anything that the
simpler tests do• allows any combination of contrasts (e.g., intact -
scrambled, scrambled - baseline), unlike simpler methods (correlations, t-tests, Fourier analyses)
• allows more complex designs (e.g., factorial designs)• allows much greater flexibility for combining data within
subjects and between subjects• allows comparisons between groups• allows counterbalancing orders within and between
subjects• allows modelling of known sources of noise in the data
(e.g., error trials, head motion)
A Simple Experiment
IntactObjects
ScrambledObjects
BlankScreen
TIME
One volume (12 slices) every 2 seconds for 272 seconds (4 minutes, 32 seconds)
Condition changes every 16 seconds (8 volumes)
Lateral Occipital Complex• responds when subject views objects
What’s real?• I created each of those time courses based by taking
the predictor function and adding a variable amount of random noise
= +
signal
noise
Formal Statistics
• Formal statistics are just doing what your eyeball test of significance did– Estimate how likely it is that the signal is real given how noisy the data is
• confidence: how likely is it that the results could occur purely due to chance?
• “p value” = probability value
– If “p = .03”, that means there is a .03/1 or 3% chance that the results are bogus
• By convention, if the probability that a result could be due to chance is less than 5% (p < .05), we say that result is statistically significant
• Significance depends on– signal (differences between conditions)– noise (other variability)– sample size (more time points are more convincing)
We’ll begin with activation
Response to Intact Objects is 4X greater than Scrambled Objects
Then we’ll assume that our modelled activation is off because a transient component
Our modelled activation could be off for other reasons
All of the following could lead to inaccurate models• different shape of function• different width of function• different latency of function
Reminder: Variability of HRF
Intersubject variability of HRF in M1Handwerker et al., 2004, NeuroImage
…though really motion is more complex
• Head motion can be quantified with 6 parameters given in any motion correction algorithm– x translation– y translation– z translation– xy rotation– xz rotation– yz rotation
• For simplicity, I’ve only included parameter one in our model
• Head motion can lead to other problems not predictable by these parameters
Now let’s throw in a pinch of linear drift
• linear drift could arise from magnet noise (e.g., parts warm up) or physiological noise (e.g., subject’s head sinks)
and then we’ll add a dash of low frequency noise
• low frequency noise can arise from magnet noise or physiological noise (e.g., subject’s cycles of alertness/drowsiness)• low frequency noise would occur over a range of frequencies but for simplicity, I’ve only included one frequency (1 cycle per
run) here– Linear drift is really just very low frequency noise
and our last ingredient… some high frequency noise
• high frequency noise can arise from magnet noise or physiological noise (e.g., subject’s breathing rate and heartrate)
Now let’s be the experimenter• First, we take our time course and normalize it using z scores
• z = (x - mean)/SD
• normalization leads to data where– mean = zero– SD = 1
Alternative: You can transform the data into % BOLD signal change.This is usually a better approach because it’s not dependent on variance
If you only pay attention to one slide in this lecture, it should be the next one!!!
We create a GLM with 2 predictors
fMRI Signal
× 1
× 2
=
ResidualsDesign Matrix
++
“what we CAN
explain”
“what we CANNOT explain”
= +Betasx
“how much of it we CAN explain”
“our data” = +x
Statistical significance is basically a ratio of explained to unexplained variance
Implementation of GLM in SPM
• SPM represents time as going down• SPM represents predictors within the design matrix as grayscale plots (where black = low, white = high)
over time• GLM includes a constant to take care of the average activation level throughout each run
– SPM shows this explicity (BV may not)
T
ime
Many thanks to Øystein Bech Gadmar for creating this figure in SPM
IntactPredictor
ScrambledPredictor
Effect of Beta Weights
• Adjustments to the beta weights have the effect of raising or lowering the height of the predictor while keeping the shape constant
The beta weight is NOT a correlation
• correlations measure goodness of fit regardless of scale• beta weights are a measure of scale
small ßlarge r
large ßlarge r
small ßsmall r
large ßsmall r
We create a GLM with 2 predictors
fMRI Signal
“what we CAN
explain”
“what we CANNOT explain”
“how much of it we CAN explain”
“our data”
when 1=2
when 2=0.5
=
= BetasxDesign Matrix
+
+ Residuals
+
= +x
Statistical significance is basically a ratio of explained to unexplained variance
Correlated Predictors
• Where possible, avoid predictors that are highly correlated with one another
• This is why we NEVER include a baseline predictor– baseline predictor is almost completely correlated with the sum of
existing predictors
Two stimulus predictors Baseline predictor
+
=
r = -.95
r = -.53
r = -.53
Which model accounts for this data?
x β = 1
• Because the predictors are highly correlated, the model is overdetermined and you can’t tell which beta combo is best
OR
x β = 1
x β = 0
+
+
x β = 0
x β = 0
x β = -1
+
+
Contrasts in the GLM• We can examine whether a single predictor is significant
(compared to the baseline)
• We can also examine whether a single predictor is significantly greater than another predictor
Contrasts
“balanced”
Conjunction of contrasts• e.g., (+1 -1 0) AND (+1 0 -1)• (Bio motion - Nonbio motion) AND (Bio motion > control)• more rigorous than balanced contrast• hypothetical (but not actual) conjunction p value = multiple of independent p values
• e.g., .01 x .01 = .001
A Real Voxel• Here’s the time course from a voxel that was significant in the +Intact -
Scrambled comparison
Maximizing Your Power
As we saw earlier, the GLM is basically comparing the amount of signal to the amount of noise
How can we improve our stats?• increase signal• decrease noise• increase sample size (keep subject in longer)
= +
signal
noise
How to Reduce Noise
• If you can’t get rid of an artifact, you can include it as a “predictor of no interest” to soak up variance
Example: Some people include predictors from the outcome of motion correction algorithms
Corollary: Never leave out predictors for conditions that will affect your data (e.g., error trials)
This works best when the motion is uncorrelated with your paradigm (predictors of interest)
Convolution of Single Trials
Neuronal Activity
Haemodynamic Function
BOLD Signal
Time
Time
Slide from Matt Brown
Fast fMRI DetectionA) BOLD Signal
B) Individual Haemodynamic Components
C) 2 Predictor Curves for use with GLM (summation of B)
Slide from Matt Brown
DEconvolution of Single Trials
Neuronal Activity
Haemodynamic Function
BOLD Signal
Time
Time
Slide from Matt Brown
Deconvolution Example• time course from 4 trials of two types (pink, blue) in a “jittered” design
Single Stick Predictor
• single predictor for first volume of pink trial type
Predictors for Pink Trial Type
• set of 12 predictors for subsequent volumes of pink trial type
• need enough predictors to cover unfolding of HRF (depends on TR)
Predictors for the Blue Trial Type
• set of 12 predictors for subsequent volumes of blue trial type
Predictor x Beta Weights for Pink Trial Type• sequence of beta weights for one trial type yields an estimate of
the average activation (including HRF)
Predictor x Beta Weights for Blue Trial Type• height of beta weights indicates amplitude of response (higher
betas = larger response)
Linear Deconvolution
• Jittering ITI also preserves linear independence among the hemodynamic components comprising the BOLD signal.
Miezen et al.2000
Fast fMRI: Estimation
Pros:• Produces time course• Does not assume specific shape for hemodynamic function• Robust against trial history biases (though not immune to it)• Compound trial types possible
Cons:• Complicated• Unrealistic assumptions about linearity if trials are too close in time
– BOLD is non-linear with inter-event intervals < 6 sec.– Nonlinearity becomes severe under 2 sec.
• Sensitive to noise
What’s this #*%&ing reviewer complaining about?!
• Particularly if you do voxelwise stats, you have to be careful to follow the accepted standards of the field. In the past few years the following approaches have been recommended by the stats mavens:
1. Correction for multiple comparisons
2. Random effects analyses
3. Correction for serial correlations
Dead Salmon
• 130,000 voxels• no correction for
multiple comparisons
poster at Human Brain Mapping conference, 2009
Correction for Multiple Comparisons
1) Bonferroni correction• divide desired p value by number of comparisons
Example:desired p value: p < .05number of voxels: 50,000required p value: p < .05 / 50,000 p < .000001
• quite conservative• can use less stringent values
• e.g., Brain Voyager can use the number of voxels in the cortical surface
• small volume correction: use more liberal thresholds in areas of the brain which you expected to be active
With conventional probability levels (e.g., p < .05) and a huge number of comparisons (e.g., 64 x 64 x 12 = 49,152), a lot of voxels will be significant purely by chance
e.g., .05 * 49,152 = 2458 voxels significant due to chance
How can we avoid this?
Correction for Multiple Comparisons
2) Gaussian random field theory• Fundamental to SPM• If data are very smooth, then the chance of noise points passing
threshold is reduced• Can correct for the number of “resolvable elements” (“resels”)
rather than number of voxels
Slide modified from Duke course
3) Cluster correction
• falsely activated voxels should be randomly dispersed • set minimum cluster size to be large enough to make it unlikely that a
cluster of that size would occur by chance• some algorithms assume that data from adjacent voxels are
uncorrelated (not true)• some algorithms (e.g., Brain Voyager) estimate and factor in spatial
smoothness of maps• cluster threshold may differ for different contrasts
4) Test-retest reliability
• Perform statistical tests on each half of the data• The probability of a given voxel appearing in both purely by chance is
the square of the p value used in each halfe.g., .001 x .001 = .000001
• Alternatively, use the first half to select an ROI and evaluate your hypothesis in the second half.
5) False discovery rate (Genovese et al, 2002, NeuroImage)
• “controls the proportion of rejected hypotheses that are falsely rejected”• standard p value (e.g., p < .01) means that a certain proportion of all
voxels will be significant by chance (1%)• FDR uses q value (e.g., q < .01), meaning that a certain proportion of the
“activated” (colored) voxels will be significant by chance (1%)• works in theory, though in practice, my lab hasn’t been that satisfied
Is the region truly active?
Doe
s ou
r st
at t
est
indi
cate
th
at t
he r
egio
n is
act
ive?
Yes
No
Yes No
HIT Type I Error
Type II Error
Correct Rejection
6) Poor man’s Bonferroni
• Jack up the threshold till you get rid of the schmutz (especially in air, ventricles, white matter)
• If you have a comparison where one condition is expected to produce much more activity than the other, turn on both tails of the comparison
• Jody’s rule of thumb: “If ya can’t trust the negatives, can ya trust the positives?”
Example: MT localizer data
Moving rings > stationary rings (orange)Stationary rings > moving rings (blue)
Correction for Temporal Correlations
Statistical methods assume that each of our time points is independent.
In the case of fMRI, this assumption is false.
Even in a “screen saver scan”, activation in a voxel at one time is correlated with it’s activation within ~6 sec
This fact can artificially inflate your statistical significance.
Autocorrelation function
To calculate the magnitude of the problem, we can compute the autocorrelation function
For a voxel or ROI, correlate its time course with itself shifted in time
Plot these correlations by the degree of shift
original
shift by 1 volume
shift by 2 volumes
If there’s no autocorrelation, function should drop from 1 to 0 abruptly – pink line
The points circled in yellow suggest there is some autocorrelation, especially at a shift of 1, called AR(1)
time
BV can correct for the autocorrelation to yield revised (usually lower) p values
BEFORE
AFTER
Temporal Smoothing of Data
• We have the option in our software to temporally smooth our data (i.e., remove high temporal frequencies)
• However, I recommended that you not use this option
• Now do you understand why?
Clarification
• correction for temporal correlations is NOT necessary with random effects analyses, only for fixed effects and individual subjects analysis
Collapsed Fixed Effects Models• assume that the experimental manipulation has same effect in each
subject• treats all data as one concatenated set with one beta per predictor
(collapsed across all subjects)• e.g., Intact = 2
Scrambled = .5
• strong effect in one subject can lead to significance even when others show weak or no effects
• you can say that effect was significant in your group of subjects but cannot generalize to other subjects that you didn’t test
Separate Subjects Models• one beta per predictor per subject
• e.g., JC: Intact = 2.1JC: Scrambled = 0.2DQ: Intact = 1.5DQ: Scrambled = 1.0KV: Intact = 1.2KV: Scrambled = 1.3
• weights each subject equally• makes data less susceptible to effects of one rogue subject
Random Effects Analysis• Typical fMRI stats test whether the differences between conditions are significant
in the sample of subjects we have tested
• Often, we want to be able to generalize to the population as a whole including all potential subjects, not just the ones we tested
• Random effects analyses allow you to generalize to the population you tested
• Brain Voyager recommends you don’t even toy with random effects unless you’ve got 10 or more subjects (and 50+ is best)
• Random effects analyses can really squash your data, especially if you don’t have many subjects. Sometimes we refer to the random effects button as the “make my activation go away” button.
• Though standards were lower in the early days of fMRI, today it’s virtually impossible to publish any group voxelwise data without random effects analysis
• You don’t have to worry about it if you’re using the ROI approach because (1) presumably the ROI has already been well-established across multiple labs; and (2) posthoc analyses of results in an ROI approach allow you to generalize to the population (assuming you include individual variance)
underpaid graduate students in need of a few bucks!
Fixed vs. Random Effects GLM
• Fixed Effects GLM cannot tell the difference between these data sets because (Intact sum - Scram sum) is the same in both cases
• In Random Effects GLM, Data set #1 would be more likely to be significant because all 3 subjects show a trend in the same direction (intact > scrambled), whereas in data set #2, only 2 of 3 subjects show a difference in that direction
Subject Intact beta
Scram beta
Diff
1 4 3 1
2 2 3 -1
3 4 1 3
SUM 10 7 3
Subject Intact beta
Scram beta
Diff
1 4 3 1
2 2 1 1
3 4 3 1
SUM 10 7 3
Sample Data #1 Sample Data #2
Strategies for Exploration vs. Publication
• Deductive approach– Have a specific hypothesis/contrast planned– Run all your subjects– Run the stats as planned– Publish
• Inductive approach– Run a few subjects to see if you’re on the right track– Spend a lot of time exploring the pilot data for
interesting patterns– “Find the story” in the data– You may even change the experiment, run additional
subjects, or run a follow-up experiment to chase the story
• While you need to use rigorous corrections for publication, do not be overly conservative when exploring pilot data or you might miss interesting trends• Random effects analyses can be quite conservative so you may want to do exploratory analyses with fixed effects (and then run more subjects if needed so you can publish random effects)
To Localize or Not to Localise?
Neuroimagers can’t even agree how to
SPELL localiser/localizer!
Approach #1: Voxelwise Statistics
1. You don’t necessarily need a priori hypotheses (though sometimes you can use less conservative stats if you have them)
2. Average all of your data together in Talairach space3. Compare two (or more) conditions using precise statistical
procedures within every voxel of the brain. Any area that passes a carefully determined threshold is considered real.
4. Make a list of these areas and publish it.This is the tricky part!
Voxelwise Approach: Example• Malach et al., 1995, PNAS
• Question: Are there areas of the human brain that are more responsive to objects than scrambled objects
• You will recognize this as what we now call an LO localizer, but Malach was the first to identify LO
LO activation is shown in red, behind MT+ activation in green
LO (red) responds more to objects, abstract sculptures and faces than to textures, unlike visual cortex (blue) which responds well to all stimuli
The Danger of Voxelwise Approaches
Source: Decety et al., 1994, Nature
• This is one of two tables from a paper• Some papers publish tables of activation two pages long• How can anyone make sense of so many areas?
Approach #2: Region of interest (ROI) analysis
• If you are looking at a well-established area (such as visual cortex, motor cortex, or the lateral occipital complex), it’s fairly easy to activate and identify the area
1. Do the stats and play with the threshold till you get something believable in the right vicinity based on anatomical location (e.g., sulcal landmarks) or functional location (e.g., Talairach coordinates from prior studies)
2. Once you have found the ROI, do independent experiments, extract the time course information and determine whether activation differences between conditions are significant– Because the runs that are used to generate the area are
independent from those used to test the hypothesis, liberal statistics (p < .05) can be used
Example of ROI ApproachCulham et al., 2003, Experimental Brain Research
Does the Lateral Occipital Complex compute object shape for grasping?
Step 1: Localize LOC
IntactObjects
Scrambled Objects
Example of ROI ApproachCulham et al., 2003, Experimental Brain Research
Does the Lateral Occipital Complex compute object shape for grasping?
Step 2: Extract LOC data from experimental runs
Grasping
Reaching
NSp = .35
NSp = .31
Example of ROI ApproachVery Simple Stats
NSp = .35
NSp = .31
% BOLD Signal Change
Left Hem. LOC
Subject Grasping Reaching
1 0.02 0.03
2 0.19 0.08
3 0.04 0.01
4 0.10 0.32
5 1.01 -0.27
6 0.16 0.09
7 0.19 0.12
Extract average peak from each subject for
each condition
Then simply do a paired t-test to see
whether the peaks are significantly different between conditions
• Instead of using % BOLD Signal Change, you can use beta weights• You can also do a planned contrast in Brain Voyager using a module
called the ROI GLM
Utility of Doing Both Approaches
• We also verified the result with a voxelwise approach
Verification of no LOC activation for grasping > reaching even at moderate threshold (p < .001, uncorrected)
Example: The Danger of ROI Approaches
• Example 1: LOC may be a heterogeneous area with subdivisions; ROI analyses gloss over this
• Example 2: Some experiments miss important areas (e.g., Kanwisher et al., 1997 identified one important face processing area -- the fusiform face area, FFA -- but did not report a second area that is a very important part of the face processing network -- the occipital face area, OFA -- because it was less robust and consistent than the FFA.
Comparing the two approaches
Voxelwise Analyses• Require no prior hypotheses about areas involved • Include entire brain • Often neglect individual differences• Can lose spatial resolution with intersubject averaging• Can produce meaningless “laundry lists of areas” that are difficult
to interpret • You have to be fairly stats-savvy and include all the appropriate
statistical corrections to be certain your activation is really significant
• Popular in Europe
Comparing the two approaches
Region of Interest (ROI) Analyses• Extraction of ROI data can be subjected to simple stats (no need
for multiple comparisons, autocorrelation or random effects corrections)
• Gives you more statistical power (e.g., p < .05)• Hypothesis-driven
• Useful when hypotheses are motivated by other techniques (e.g., electrophysiology) in specific brain regions
• ROI is not smeared due to intersubject averaging• Important for discriminating abutting areas (e.g., V1/V2)
• Easy to analyze and interpret • Neglects other areas which may play a fundamental role• If multiple ROIs need to be considered, you can spend a lot of
scan time collecting localizer data (thus limiting the time available for experimental runs)
• Works best for reliable and robust areas with unambiguous definitions
• Popular in North America
A Proposed Resolution
• There is no reason not to do BOTH ROI analyses and voxelwise analyses– ROI analyses for well-defined key regions– Voxelwise analyses to see if other regions are also involved
• Ideally, the conclusions will not differ• If the conclusions do differ, there may be sensible reasons
– Effect in ROI but not voxelwise• perhaps region is highly variable in stereotaxic location between subjects
• perhaps voxelwise approach is not powerful enough
– Effect in voxelwise but not ROI• perhaps ROI is not homogenous or is context-specific
Finding the Obvious
Non-independence error• occurs when statistical tests performed are not independent
from the means used to select the brain region
Arguments from Vul & Kanwisher, book chapter in press
A priori probability of getting JQKA sequence = (1/13)4 = 1/28,561
A posteriori probability of getting JQKA sequence = 1/1 = 100%
Non-independence ErrorEgregious example• Identify Area X with contrast of A > B• Do post hoc stats showing that A is statistically higher than B• Act surprised
More subtle example of selection bias• Identify Area X with contrast of A > B• Do post hoc stats showing that A is statistically higher than C and C is
statistically greater than B
Arguments from Vul & Kanwisher, book chapter in press
Figure from Kriegeskorte et al., 2009, Nature Neuroscience
Double Dipping & How to Avoid It• Kriegeskorte et al.,
2009, Nature Neuroscience
• surveyed 134 papers in prestiguous journals
• 42% showed at least one example of non-independence error
Correlations Between Individual Subjects’ Brain Activity and Behavioral Measures
Sample of Critiqued Papers:Eisenberg, Lieberman & Williams, 2003, Science• measured fMRI activity during social rejection• correlated self-reported distress with brain activity• found r = .88 in anterior cingulate cortex, an area implicated in physical pain perception• concluded “rejection hurts”
social exclusion > inclusion
“Voodoo Correlations”
• reliability of personality and emotion measures: r ~ .7• reliability of activation in a given voxel: r ~ .7• highest expected behavior: fMRI correlation is ~.74• so how can we have behavior: fMRI correlations of r ~.9?!
2009
Voodoo
The original title of the paper was not well-received by reviewers so it was changed even though some people still use the term
“Voodoo Correlations”
"Notably, 53% of the surveyed studies selected voxels based on a correlation with the behavioral individual-differences measure and then used those same data to compute a correlation within that subset of voxels."
Vul et al., 2009, Perspectives on Psychological Science
Avoiding “Voodoo”
• Use independent means to select region and then evaluate correlation
• Do split-half reliability test– WARNING: This is reassuring that the
result can be replicated in your sample but does not demonstrate that result generalizes to the population
Is the “voodoo” problem all that bad?
• High correlations can occur in legitimately analyzed data• Did voxelwise analyses use appropriate correction for multiple comparisons?
– then result is statistically significant regardless of specific correlation
• Is additional data being used for1. inference purposes?– if they pretend to provide independent support, that’s bad2. presentation purposes?– alternative formats can be useful in demonstrating that data is clean (e.g., time
courses look sensible; correlations are not driven by outliers)
Recommended