• Course Description
• Course Prerequisites
• General Information
• Syllabus
• Lectures
• Laboratory projects
• Bibliography
• Contact

Credits:
2 - 2 - 8 (lectures - lab - out of class time)

Contents:
This team taught, multidisciplinary course covers the fundamentals of magnetic resonance imaging (MRI) relevant to the conduct and interpretation of human brain mapping studies.
The course provides in depth coverage of the physics of image formation, mechanisms of image contrast, and the physiological basis for image signals.  Parenchymal and cerebrovascular neuroanatomy and application of sophisticated structural analysis algorithms for segmentation and registration of functional data are discussed.  Additional topics include functional MRI (fMRI) experimental design including block design, event related and exploratory data analysis methods, building and applying statistical models for fMRI data. Human subjects issues including informed consent, institutional review board requirements and safety in the high field environment are presented.

Format:
Weekly 2 hour lecture followed by weekly 2 hour laboratory/discussion session.  Laboratory will include fMRI sessions and data analysis workshops.  Assignments include reading of both textbook chapters and primary literatures as well as solving problem sets and analysing fMRI data  in the laboratory.

TOP

This HST course will be open for registration to MIT, Harvard, and affiliated graduate students who have the following prerequisites:
probability theory, linear algebra, differential equations, introductory or college level courses in neurobiology, physiology and physics.
Or with special permission of instructor.

TOP

TOP

SYLLABUS

    A. MRI Physics

• Introduction to course (R. Gollub). Historical background, fMRI in perspective with other brain imaging modalities (R. Savoy). Begin physics of MRI (J. Jovicich). Safety and human subject issues (MRI safety video).
• Overview of MRI physics. Radio-frequency pulses. Spatial encoding of MR signal. Conventional and fast MRI sequences and parameters. Thermal noise. (L. Wald)
• BOLD contrast, how brain tissue components influence contrast and resolution. Magnetic field strength effects. Contrast limitations. Spatial and temporal resolution. Variations based on pulse sequences. Physics of perfusion, difussion and difusion tensor imaging. (J. Jovicich, L. Wald)
• Effects of CBV and CBF in BOLD signal. Alternatives to BOLD contrast. Physiological sources of noise in fMRI data: sources and correction (J. Jovicich and R. Hoge)

• Applied respiratory and cardiovascular physiology. (R. Banzett)
• Physiologic basis of BOLD signal. Neurovascular coupling. Global vs regionally specific changes in CBF and metabolism. (R. Hoge)
• Neural regulation of CNS vasculature (R. Gollub). Clustered volume acquisition and cardiac gating techniques for detection of auditory activation and brainstem signals. (J. Melcher)

• Characterization of structural MR data. Structural-to-functional registration. Spatial information. Individual variability. Difussion tensor imaging uses for white matter tract tracing. (D. Kennedy)
• A brief introduction to the retinotopic structure of early visual cortex, the construction and use of surface models of the cortex, automatic segmentation of subcortical structures. (B. Fischl)

• Functional neuroanatomic systems (R. Gollub). Experimental design. Block, spaced single trial, rapid single trial, periodic, other mixed modes. (L. Davachi)

• Paradigm for statistical inference. (E. Brown)
• Building statistical models for fMRI data. (E. Brown)
• Application of statistical models for fMRI data. Statistical considerations in experimental design. (A. Dale)
• MEG-fMRI statistical analysis modeling: optimizing spatio-temporal resolution. (A. Dale)

LABORATORY PROJECTS

 
1. Introduction to fMRI data acquisition

Part 1:  The lab includes familiarizing students with the MRI scanning environment as well as acquiring image data from both a phantom and a human subject during a conventional fMRI experiment. (Outline)
(2 hours - Rick Hodge, Larry Wald and Jorge Jovicich, data will be acquired at NMR Center MGH)

Part 2:  The second part of the lab involves analysing the data acquired during the first part and answering a set of questions. The goal of this analysis is to demonstrate the impact of noise and image quality effects on fMRI signal detection. (Lab1 Manual)
(2 hours- Rick Hodge, Larry Wald and Jorge Jovicich, analysis will be done at MIT)
 

2. Biophysical basis of fMRI signals

Part 1: The goal of this lab is to demonstrate the use of both physiological measurement equipment and MRI techniques to monitor the physiological status of subjects in the scanner during a fMRI study. (Outline)
(2 hours - Robert Banzett and Rick Hodge, data will be acquired at NMR Center MGH)

Part 2:  The goal of the second part of the lab is to analyse the data acquired during the first part to isolate effects of different physiological processes (flow, metabolism) from the BOLD signal. (Lab2 Manual)
(2 hours - Robert Banzett and Rick Hodge, analysis will be done at MIT)
 

3. Improving fMRI signal detection using physiological data

The goal of this lab is to use examples from auditory cortex and brainstem to illustrate how fMRI signal detection can be improved:
1) using physiological signals measured during imaging.
    Example: cardiac gating, a technique that avoids cardiac-related signal fluctuations (particularly a problem in brainstem structures)
2) by tailoring experimental design to the neurophysiological properties of the system under study.
    Example: clustered volume acquisition, a technique that reduces the impact of scanner acoustic noise on auditory fMRI activation
(Lab3 Manual)
(2  hours - Irina Sigalovsky and Jennifer Melcher)
 

4.  Characterization of structural MRI data

This Lab examines ways in which different brain anatomical structures can classified based on the signal intensity of high spatial resolution anatomical MR images acquired with different contrast weightings. (Lab4 Manual)
(2 hours- David Kennedy and Bruce Fischl)
 

5. Statistical analysis of fMRI data

Part 1: Use of concepts introduced in the statistics lectures: design matrix, design efficiency, Finite Impulse Response and Gamma event-response models. (Lab5a Manual)
(2 hours-Doug Greve)

Part 2: Use of the generalized linear model as implemented in the Statistical Parametric Mapping (SPM) software. (Lab5b Manual)
 (2 hours - Russell Poldrack)

TOP

BIBLIOGRAPHY

• Haacke, Brown, Thompson and Venkatesan (1999) Magnetic Resonance Imaging: Physical Principles and Sequence Design, Wiley-Liss.

• Hall et al., "Sparse" Temporal Sampling in Auditory fMRI, Hum. Brain Mapp. 7:213-223 (1999)
• Guimaraes et al., Imaging Subcortical Auditory Activity in Humans, Hum. Brain Mapp. 6:33-41 (1998)

Statistics