MIT HST 583 - FMRI Experimental Design (68 pages)

Previewing pages 1, 2, 3, 4, 5, 32, 33, 34, 35, 64, 65, 66, 67, 68 of 68 page document View the full content.
View Full Document

FMRI Experimental Design



Previewing pages 1, 2, 3, 4, 5, 32, 33, 34, 35, 64, 65, 66, 67, 68 of actual document.

View the full content.
View Full Document
View Full Document

FMRI Experimental Design

76 views


Pages:
68
School:
Massachusetts Institute of Technology
Course:
Hst 583 - Functional Magnetic Resonance Imaging: Data Acquisition and Analysis
Unformatted text preview:

FMRI Experimental Design Lila Davachi Department of Brain Cognitive Sciences M I T Because fMRI BOLD data is not an absolute measure of neuronal activity all study designs must provide the opportunity to statistically contrast the neuronal activity of interest with a suitable rest or background condition Thus study design is of paramount importance Why Hypothesis What Behavior Where Neuroanatomy How fMRI Study Design Thanks to Chantal Stern 43 7 Key Points What can fMRI tell you Always comparing across conditions Characteristics of the hemodynamic response HRF and how this affected the sequential development of fMRI paradigms and influences study design Sense of important design issues What good is fMRI What it can tell you Relative local neural activity LFP s NOT absolute neural activity NOT excitation vs inhibition NOT about necessity of a given region for a task NOT fine grained temporal information Key Points What can fMRI tell you Always comparing across conditions Characteristics of the hemodynamic response HRF and how this affected the sequential development of fMRI paradigms and influences study design Sense of important design issues Subtraction Paradigm Donder s method Ex How to measure time of a mental transformation A random series of A s and B s presented and the subject must 1 Respond whenever an event occurs RTi 2 Respond only to A not to B RTii 3 Respond X to A and Y to B RTiii RTi RT detect RT response RTii RT detect RT discrimination RT response RTiii RT detect RT discrimination RT choice RT response THUS RT discrimination RTii RTi RT choice RTiii RTii Criticisms of Subtraction Paradigm 1 That we already know what counts as a single mental process i e choice is a single mental process 2 Assume that adding components does not affect other processes i e assumption of pure insertion THUS one should pick tasks that differ along ONE dimension either change the task OR the stimuli but not BOTH And a resting baseline is good to include however the interpretation should be taken lightly more later The loose task comparison Does not hold all variables constant BUT 1 Uses a low level reference task 2 Allows the data to be examined for predictable stimulus or response driven activations 3 Allows the more extensive activation pattern to be observed The loose Task Comparison TASK 1 TASK 2 STI STI PRE PRE DOG DOG STI STI PRE PRE PAINT GRE PAINT GRE STI STI PRE PRE EGG STR EGG STR The tight task comparison Try to hold all variables constant including Stimulus display nominally or statistically Response and response selection characteristics Performance level especially if comparing cohorts Eye movements Emotional state minimize anxiety and boredom The tight Task Comparison TASK 1 TASK 3 STI STI PRE PRE DOG DOG STI STI PRE PRE PAINT GRE PAINT GRE STI STI PRE PRE EGG STR EGG STR TASK 2 BRAIN AREAS THAT DIFFER TASK 1 TASK 2 ALL ACTIVE BRAIN AREAS Thanks to Randy Buckner 2 minus 1 Example Interested in semantic processing and how it affects memory Parameters to specify in any experiment 1 Subjects normal vs special populations 2 What part of brain look at How many slices can you have for your TR 3 Choosing your TR How often can you take a full set of pictures 4 What coil will you use surface coils higher SNR only partial coverage head coils lower SNR complete coverage 5 Toggle many times between conditions within a scan 6 Run as many scans as possible within a subject Key Points What can fMRI tell you Always comparing across conditions Characteristics of the hemodynamic response HRF and how this affected the sequential development of fMRI paradigms and influences study design Sense of important design issues Visual Stimulation 2 sec Flashes Percent Signal Change 4 0 3 0 2 0 1 0 0 0 1 0 0 10 20 30 40 50 60 70 80 TIME seconds Blamire Ogawa et al PNAS 1992 90 100 110 120 Visual Cortex During Brief Visual Stimulation 2 0 100 msec 1000 msec 1 0 34 msec 0 0 5 0 5 TIME seconds Courtesy of Robert Savoy Kathleen O Craven MGH NMR Center 10 15 Blocked design fMRI BLOCKED HORSE abstract or concrete love upper or lowercase Blocked fMRI Memory Paradigm STI PRE HORSE 0 STI PRE LOVE GRE 8 52 STI PRE HUMOR STR 74 118 140 TIME SEC Wagner et al OHBM 1998 STI PRE CHAIR STR 184 206 250 Typical Blocked Design Response Percent Signal Change 0 50 0 25 0 0 25 HORSE 0 20 40 LOVE 60 HUMOR CHAIR 80 100 120 140 160 180 200 220 240 Time 13 Slices Per Brain Image Thanks to Robert Savoy 80 brain images per 4 minute run Thanks to Robert Savoy For purposes of illustration Thanks to Robert Savoy Examine the data from one slice of the brain as a function of time Thanks to Robert Savoy Thanks to Robert Savoy Combine these Combine these Are these voxel levels statistically different from each other Thanks to Robert Savoy Combine these Combine these Consider EACH voxel across all time points Thanks to Robert Savoy Typical Blocked Design Response Percent Signal Change 0 50 0 25 0 0 25 HORSE 0 20 40 LOVE 60 HUMOR CHAIR 80 100 120 140 160 180 200 220 240 Time Event Related fMRI BLOCKED SPACED EVENT RELATED 16 sec Spaced Event Related fMRI Language Paradigm COU 0 PRE TRA 32 AFT STA 64 TIME seconds Buckner Bandettini et al PNAS 1996 PEL DRI 96 Single Trial Response Across a Run MEAN MR SIGNAL 381 380 379 378 377 376 375 0 32 64 96 128 TIME SEC 160 192 228 256 Event Related Selectively Averaged Response 380 379 378 377 376 375 PERCENT SIGNAL CHANGE MEAN MR SIGNAL 381 5 0 0 1 2 3 4 5 6 7 8 9 101112131415 TIME seconds Broca s Area During Language Paradigm 379 378 377 376 375 374 0 2 4 6 8 10 12 14 TIME sec Thanks to Randy Buckner Rapid Event Related fMRI BLOCKED SPACED EVENT RELATED 16 sec RAPID EVENT RELATED 2 sec Assessing the Linearity Hypothesis 1 sec on 0 sec Dale and Buckner Hum Brain Map 1997 20 sec Response to Averaged Single Trials PERCENT MR SIGNAL 4 3 Split half reliability 2 1 0 1 0 1 2 Thanks to Randy Buckner 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 TIME sec Assessing the Linearity Hypothesis 5 Second ITI 0 sec 20 sec 0 sec 5 sec 20 sec Thanks to Randy Buckner Response to Averaged Double Trials 4 PERCENT MR SIGNAL 3 2 1 0 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 TIME sec Thanks to Randy Buckner Assessing the Linearity Hypothesis Separation of Responses 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 TIME SEC Thanks to Randy Buckner 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 …


View Full Document

Access the best Study Guides, Lecture Notes and Practice Exams

Loading Unlocking...
Login

Join to view FMRI Experimental Design and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view FMRI Experimental Design and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?