1

Designing a behavioral experiment

Chris Rorden Designing fMRI studies

– fMRI signal is sluggish and additive.– Efficient designs maximize predictable changes in HRF.– Efficient designs are often very predictable

Participant may anticipate events. Techniques for balancing efficiency and psychological validity.

2

Finding effects

Statistics are based on the ratio of explained predictable versus unexplained variability:

We can improve statistical efficiency by– Increasing the task related variance (signal)

Designing Experiments (today’s lecture)

– Decreasing unrelated variance (noise) Spatial and temporal processing lectures.

– Good signal in our fMRI data Physics lectures

Signal+NoiseNoise

F=SignalNoise

t=

3

fMRI Signal

There are two crucial aspects of the BOLD effect:– The HRF is very sluggish

Delay between brain activity and changes in fMRI images (~5s).

– The HRF is additiveDoing a task twice causes about twice as much change as

doing it once.

4

The BOLD timecourse

Visual cortex shows peak response ~5s after visual stimuli.

Indirect measure

0 6 12 18 24

% Signal Change

2

1

0

Time (seconds)

5

Temporal Properties of fMRI Signal

We predict the HRF by convolving the neural signal by the HRF. We want to maximize the amount of predictable variability.

Convolved Response

=

Neural Signal HRF

6

BOLD effects are additive

Three stimuli presented rapidly result in almost 3 times the signal of a single stimuli (e.g. Dale & Buckner, 1997).

Crucial finding for experimental design. Note there are limits to this additivity effect, but the basic point

is that more stimuli generate more signal (see Birn et al. 2001)

7

Comparing predictable HRF

Consider 3 paradigms:1. Fixed ISI: one stimuli every

16 seconds.– Inefficient

2. Fixed ISI: one stimuli every 4 seconds.

– Insanely inefficient: virtually no task-related variability

3. Block design: cluster five stimuli in 8 seconds, pause 12 seconds, repeat.

– Very efficient.– Cluster of events is additive.

Note peak amplitude is x3 the 16s design.

8

Block Designs

aka ‘Box Car’, or ‘Epoch’ designs. Different cognitive processes occur in distinct

time periods1. Press left index finger when you see 2. Press right index finger when you see 3. Do nothing when you see

9

Optimal Design

Block designs are optimal.– Present trials as rapidly as possible for ~12 sec– Summation maximizes additive effect of HRF.– Consider experiment:

Three conditions, each condition repeated 14 times (once every 900ms)1. Press left index finger when you see 2. Press right index finger when you see 3. Do nothing when you see

Note huge predictable variability in signal.

10

Block designs

While efficient, block designs are often predictable.

May not be experimentally valid.Optimal block length around 12s, followed by

around 12s until condition is repeated.– Avoid long blocks:

Reduced signal variabilityLow frequency signal will be hard to distinguish from low

frequency signals such as drift in MRI signal.

11

Block designs good for detection, poor for estimating HDR.

Block design limitations

Detection: which areas are active?

Estimation: what is the timecourse of activity?

-10 0 10 20 30 40

R_Tap L_Tap right left

-0.005

0.000

0.005

0.010

12

Block design limitations

While block designs offer statistical power, they are very predictable.– E.G. our participants will know they will

press the same finger 14 times in a row.

Many tasks not suitable for block design– E.G. Novelty detection, memory, etc.– Your can not post-hoc sort data from

block designs, e.g. Konishi, et al., 2000 examine correct rejection vs hits on episodic memory task.

13

Event related designs

Much less power than block designs.– Simply randomizing trial order of our block design, the typical event related design has one quarter the efficiency.– Here, we ran 50 iterations and selected the most efficient event related design.

Still half as efficient as the block design. Note this design is not very random: runs of same condition make it efficient.

14

Permuted Blocks

Permuted block designs (Liu, 2004) offer possible some unpredictability…

Permuted Design:

1. Start with a block design

2. Randomly swap stimuli

3. Repeat step 2 for n iterations

More iterations = less predictable, less power

15

Permuted Blocks

Below you can see our study after 10 permutations during the first minute of scanning.

Permuted block designs can offer a balance of power and predictability.

16

Jittered Inter-Stimulus Interval

Dale et al. suggest using exponential distribution for inter-trial intervals.

Exponential Distribution:– Many trials have short duration– A few trials have long duration– Efficient because jittering makes events block-like

1 condition, fixed ISI = little variability 1 condition, exponential ISI = more variability

17

Interstimulus Intervals and Power

Fixed ISI: low statistical powerFixed ISI have most power if >12sec between stimuliAt that rate, only a few dozen trials in a 10 minute scan.

In theory, variable ISI can offer much more efficiency than fixed ISI.

Exponential Distribution

18

Should you use variable ISIs?

In practice, variable ISIs often reduce power. Most experiments have more than one condition, so

fixed ISI designs also have temporal variability. Unless you are looking at low-level processes (e.g. early

vision), trials must be separated by a couple seconds. For multi-condition studies, the minimum time between

trials is crucial.– People are faster to respond to fixed ISI than variable ISI– Therefore, fixed ISI are often more powerful– However, variable ISI may help us reconstruct the true shape

of the HRF measured.

19

Tips

For event related designs – helpful if TR is either variable or a not evenly divisible by the interstimulus interval.

Allows you to accurately estimate whether conditions influence the latency of response.

TR divisible by ISI TR not divisible by ISI

20

Generate your own experiments…

Set the TR (time per volume) Set the number of volumes Set minimum ISI – this will be time between trials for block

designs. Set the mean ISI – this will be the average time between

trials for event related designs. Set the number of conditions. Iterations – you can compute hundreds of event related

designs and choose the most efficientHigh iterations will lead to efficient but predictable designs.

Permutations – select the number of permutations for the permuted block design.Fewer permutations lead to efficient but predictable designs. Press the type of study

you want to generate1. Block2. Permuted Block3. Fixed ISI Event4. Exponential ISI Event

21

Experiment generator

Software reports variance. Higher variance corresponds with more power.

– Power relative – do not directly compare studies with different TR or volumes.

– Only approximate estimate of power: does not ensure conditions have uncorrelated responses.

– Press ‘i’ button to see text file of condition onset times (you can paste into e-prime).

22

General guidelines (Nichols et al)

1. If possible, use block design– Keep blocks <40s (temporal processing lecture describes why)

2. Limit number of conditions– Pairwise comparisons far apart in time may be confounded by low

frequency noise.

3. Randomize order of events that are close to each other in time.

4. Randomize SOA between events that need to be distinguished.

5. Run as many people as possible for as long as possible.

6. Have testable anatomical prediction

23

Increasing power

Increasing the sample size (more people, more scans per person) is a fantastic way to increase statistical power.

However, long sessions can lead to problems:– Increased head motion– Poor task compliance (bored = fall asleep)– Learning effects (make sure the different

conditions balanced throughout session).