fMRI Lab class Spring 2006

Study Questions

 

1.       Introduction and course overview

1.1.      What are the relative strengths of fMRI, ERP/fMRI, DTI and habituation assessment in brain imaging and what characteristics of brain computation do they track?

1.2.      How might you use these methods to test a cognitive model?

1.3.      What are representational and control areas of the brain?

2.       E-Prime: basics computerized experimentation and timing

2.1.      What are the functions of E-Studio, E-Run, E-Merge and E-DataAid

2.2.      What are properties in E-Prime

2.3.      How do you export behavioral data from E-Prime to a program like SPSS?

2.4.      What are package functions and how are they used?

2.5.      What in a list object is the Weight, Procedure, and attributes?

2.6.      How do you reference an Attribute in a display object presentation text?

2.7.      What are nested lists and know how to use them.

2.8.      How would you implement a simple experiment in E-Prime like a Stroop and calculation task? 

2.9.      What typically causes timing errors in experiments?

2.10.   How does a monitor refresh rate effect timing?

2.11.    What does Block’s law imply about relating refresh based presentation to constant stimulus (as in slide presentation)

2.12.    What is cumulative versus event timing and whey should you use each?

2.13.    What is pre-release?  How doest it improve timing? How does it complicate response scoring/ proving response related feedback.

2.14.   What is .OnsetTime .RT, .RTOnsetTime and how can it be logged and review. 

2.15.   If the refresh is slightly slower then the intended refresh rate (e.g, 60.1 Hz how will timing drift?)

2.16.   How would you use a spread sheet to calculate the duration of displays?

3.       E-Prime: code programming experiments

3.1.      What is script?

3.2.      What are the variable types in E-Basic

3.3.      What is the experimental context object?

3.4.      How is information on a list object used in an experiment?

3.5.      How do you set a variable to be used on a display object?

3.6.      What is a package call?

3.7.      How do you get and set properties of objects?

3.8.      Define and be able to use the following operations: IF, FOR, LOOP, DIM, LONG, STRING, debug.print, “&”, clock.read, random(), global/local variables

3.9.      Be able to write script that would:

3.9.1.           output the time a display object occurred, the response collected and RT to the debug log

3.9.2.           get an attribute, check it and either put up an error message if unexpected or set duration of display object

3.9.3.           provide  cumulative accuracy for across multiple blocks in an experiment

4.       E-Prime: fMRI synchronization, time audit and running in magnet

4.1.      What are special considerations for fMRI experiments rather than standard behavioral experiments

4.2.      What is the BEST package file and what does it do to facilitate fMRI experiments

4.3.      How to you use BESTLogEvent and BESTLogResponseEvent in terms of where to place the call and set the parameters

4.4.      Know the function of BEST SessionInit, MenuBegin/End, RunBegin/End and where to place them in an experiment

4.5.      Know how to “jitter” the display.

4.6.      Know how to collect a response, indicate to the subject that the response was collected but not alter the presentation times of the experiment

4.7.      What is a pdat file?

4.8.      Know how to get timing data out of a pdat file?

4.9.      What should be the timing mode and prerelease settings of display objects? What are the symptoms of these parameters not being set properly?

4.10.   What is the Session, Run, Block, and Run level of an experiment?

5.       fMRI scanning, BOLD response, and experimental design

5.1.      How does MRI able to measure the signal from brain tissue?

5.2.      What is T1, T2, T2* of the tissue?

5.3.      What is the quality of hemoglobin that allows functional imaging?

5.4.      Be able to describe what BOLD stands for and the steps by which a stimulus produces and increase in the MR signal?

5.5.      What is the hemodynamic lag?

5.6.      What are major contributions of noise to the MR signal?

5.7.      What is the signal correlated with in terms of neural events?

5.8.      Can the lag across locations be used to estimate when a brain areas was activated relative to other areas? 

5.9.      What is the difference between Echo Planar and Spiral scan and when is one preferred?

5.10.   What are the advantages and disadvantage of higher field strength magnets?

5.11.   What are the advantages and disadvantages of surface and phased array coils?

5.12.   What are functional versus anatomical imaging and what is each used for?

5.13.   How doe the MR response intensity change with stimulus intensity? Can you detect a very short stimulus? 

5.14.   As stimuli get closer in time how does the MR response change?

5.15.   What are block, event, and parametric designs and when is each best to use?

5.16.   When might you use a single pedestal, oscillatory, block, parametric and event design?

5.17.   What are parametric, non-parametric, and exploratory analysis methods and why use each?

5.18.   What are the four steps in Platt’s Strong inference and how would you apply them to fMRI experiments. 

6.       fMRI data processing, Introduction to Brain Voyager – Preprocessing

6.1.      What are the files types that come from the scanner and what are they converted to in BV?

6.2.      What is a run, volume, slice, and voxel?

6.3.      Why do we alter the data in preprocessing?

6.4.      How does 3D motion correction work ? What does it correct for? What does it not correct for?

6.5.      What are the six degrees of movement and what are acceptable ranges?  Which are the dimensions of the greatest problems?

6.6.      What is slice time correction? What artifacts does it remove?

6.7.      What is scanning order and why it important to specify this?

6.8.      What is spatial filtering and FWHM specification?

6.9.      What artifacts are removed by low pass filtering, high pass filtering, and band pass filtering? 

6.10.   Given an sample voxel time course be ready to draw what the result would look like after low and high pass filtering. 

6.11.   What are “Spikes” in your data? 

6.12.   What is mean intensity correction and why must you be careful in applying it. 

6.13.   What quality control checks should you do on your data to make sure the parameters for the files are reasonable and the subject collected data is reasonable? 

7.       Spatial normalization – alignment in Talairach, Cross reference to atlases

7.1.      When must there be spatial normalization?

7.2.      What is Talairach normalization and the benefits and deficits

7.3.      What are ROI Based alignment and benefits and deficits

7.4.      What is cortex based alignment and benefits and deficits

7.5.      What does Talairach normalize for and what does it not?

7.6.      What is the Talairach coordinate system, what defines the center, edges, and orientation of the box?

7.7.      What is the Digital Anatomist?

7.8.      How do you relate Talairach locations to cortical gyrus/sulcus and Brodmann areas

7.9.      If you use a single subject data versus group average what do you do if the activation is not in grey matter?

7.10.   How do you sue BrainTutor and Talairach Daemon to identify locations?  What if they disagree?

7.11.   When should a brain Not be put into a merged Talairach data set?

7.12.   What do you do with brains with tumors?

7.13.   Is Talairach good enough to distinguish V1 from V2?  DLPFC form ACC? 

8.       fMRI General Linear Model (GLM)

8.1.       Be able to identify main and interaction effects in a 2 way ANOVA

8.2.       What are three things that are  different in a fMRI use of ANOVA/CLM?

8.3.      What is the advantage of GLM versus t-Test, and correlation?

8.4.      As you increase P what happens to activation maps?

8.5.      What assumptions of parametric analysis is typically violated?

8.6.      What are beta weights, residuals, design matrices?

8.7.      How is GLM different for one subject one run, multiple subjects multiple runs, and calculation from a GLM

8.8.      What is the meaning of the +, 0, - and blank specification for Overlay GLM contrasts and relative contribution maps?

8.9.      What is and when do you use a relative contribution analysis?

8.10.   What is a conjunction analysis and what is it good for and how do you do it in BV?

8.11.   How are Event related designs different from block design analysis?

8.12.   What is a fixed and random effect and when should you run each.  How many subjects does it typically require?

8.13.   What does separate subject and study predictors mean in the analysis specification? 

8.14.   What happens when you create an interaction design? 

9.        fMRI Safety

9.1.      How frequent and how serious are scanner safety incidents?

9.2.      What are the four classes of risk?

9.3.      What are the risk from Magnetic field, gradient switching, RF, acoustic noise, coolant, closed bore?

9.4.      When is the magnet off?

9.5.      What do you do if subject says had metal in eye?

9.6.      When do exclude a subject for running?

9.7.       What are eddy currents and what harm can they pose?

9.8.      What is the 5 Gauss line and what does it mean for hazards?

9.9.      What produces peripheral nerve stimulation?

9.10.   How is warming controlled for?

9.11.   How loud is the magnet?  How do we protect the subject?

9.12.   What are risks of putting someone in/out of magnet?

9.13.   Is claustrophobia predictable? 

9.14.   What is the emergency squeeze ball, how to instruct someone to use it?

9.15.   What are the emergency shut down switches when should you versus the technologist use them?

9.16.   How would you get someone out of the magnet if power failed in the building?

9.17.   What are the risks of slow and fast helium gas leaks?

9.18.   What needs to be screened for in a magnet. 

9.19.   How administers and discloses the pregnancy tests?

10.    fMRI Ethics

10.1.   What is an Incidental Finding?

10.2.   What does it mean to operating under “good Samaritan rules? 

10.3.   What are the four types of HARM that can be done by reporting an incidental finding to the subject?

10.4.   How much should you check for incidental findings?

10.5.   What are the odds of finding something and likely hood of needing to be reported, urgently reported (in a week), or require immediate followup?

10.6.   When and how should the subject be told?

10.7.   Describe some local incidental findings that have occurred and example of impact.

10.8.   Do facilities use a common method of handling IFs and review?

10.9.   Do subjects want to know of IFs when they are benign, malignant but curable, malignant and not curable, life threatening?

10.10.                     What is the role for fMRI in imaging of children?  What are the risks? 

11.    Effective fMRI Experimental design

11.1.   What are the relative strengths and weakness of given designs?

11.2.   What is the shape of the hemodynamic response and how does it act as a filter? What are implications as conditions Hz increases from 0.05 to 0.2 Hz?

11.3.   How is fMRI noise distributed?

11.4.   For block designs where is the sweet spot duration of a block?

11.5.   What are the advantages/disadvantages of blocked designs?

11.6.   What are advantages/disadvantabes of tightly matched versus “low level” controls?

11.7.   Why is a massed design bad?

11.8.   When use a pedestal design versus an alternating design?

11.9.   What are castle, fixed ordering and random/counterbalanced designs?

11.10.                     What are factorial designs?

11.11.                     How might you use a cognitive conjunction design?

11.12.                     What are the advantages/disadvantages of event related designs?

11.13.                     Contrast and give when to use slow-ER, rapid-ER, and Oddball designs?

11.14.                     How do the results look for the two methods?

12.    Statistical power in fMRI

12.1.   How does the standard error of the mean reduce with number of observations and why is this important in fMRI experiment design?

12.2.   What is d’, what are typical d’, mean effect sizes, and SDwithin and SDbetween

12.3.   How do within and between subject variance contribute to shared variance?

12.4.   What are the benefits and problems of doing within subject designs and not doing random effects analysis?

12.5.   How do spatial and temporal filtering improve statistical power?  Are they significant factors?

12.6.   Be able to read a power curve and be able to select number of subjects and volumes for a study.

12.7.   How do you use results from the first few subjects to guide future experiments in terms of deciding on number of subjects, runs, and interpretation?

13.    Threshold for significance correction & False Discovery Rate

13.1.   What is the multiple comparison problem?

13.2.   Why use Volume thresholds?

13.3.   What is familywise error rate and Bonferroni correction.

13.4.   What is False Discovery Rate correcting for? Is it the method of choice?

13.5.   How does FDR and Family wise error rate effect an ROI form?

13.6.   What spatial smoothing should you do with FDR?

14.    Practical guide to subject running fMRI experiment

14.1.   Subject selection

14.2.   Special populations children & patients

14.3.   Simulator running

14.4.   Head movement control

14.5.   Subject communication

14.6.   Feedback

14.7.   Ethics & incidental findings

14.8.   Human subject approval

14.9.   Dealing with claustrophobia

14.10.                     How do pregnancy test and inform patient?

14.11.                     What about body piercing or metal in eye?

14.12.                     How do you put a person in the magnet to minimize movement

14.13.                     How do you adjust the mirror? 

14.14.                     How do you communicate between runs? 

15.     (Midterm Exam)

16.    Convolution analysis methods (Delayed till after break) M. Cole

16.1.   What is the standard hemodynamic response (time course and shape)?

16.2.   How well does GLM analysis handle widely spaced effects, closely spaced (2s) without jitter, closely spaced with jitter.?

16.3.   What is effect of wrong HRF shape used in analysis?

16.4.   How does deconvolution method work? What assumptions doe it make about shape?

16.5.   Describe how deconvolution approach differs from standard GLM approach?

16.6.   What are sustained and transient effects in fMRI?  How do you analyze for them?

16.7.   What is a mixed sustained and transient analysis ?How does it help interpretation?

16.8.   What is a deconvolutional GLM and how is it different from standard GLMs and what are the advantages and disadvantages of doing them?

16.9.   How do you separate block and trial effects?