Brain Imaging
Blok 4 | 2022-2023
,Inhoud
_________________________________________________________________________
Summary 3
fMRI terminology 3
1. Introduction 9
Lecture 9
Additional videos 14
Introduction to fMRI 14
Analysis of fMRI Data 15
fMRI Data Structure & Terminology 16
Literature 18
Progress and challenges in probing the human brain 18
2. Neuroscience Background 20
Lecture 20
3. Imaging Brain Activations 25
Lecture 25
Additional videos 29
4. BOLD analysis 33
Lecture 33
Basic fMRI analysis 33
Linear Time Invariance 34
General Linear Model (GLM) 36
5. MR Physics and Imaging 38
Lecture 38
6. GLM: Experimental design 41
Lecture 41
Deconvolution 41
GLM 42
Experimental design 42
Voodoo correlations 45
Additional videos 46
7. Patterns in time and space 48
Lecture 48
Pattern analysis 48
Multivariate pattern analysis 49
Representational similarity analysis (RSA) 51
8. Encoding models 54
Lecture 54
Receptive fields in fMRI 54
Receptive fields? 54
Extravisual 58
9. Comparison 61
Lecture 61
Multiple comparisons 61
Comparisons 64
1
, We’re all unique 65
10. Preprocessing 67
Lecture 67
First steps on the long road from raw data to results 67
Reproducible fMRI 71
11. Connectivity 73
Lecture 73
Anatomical connectivity 73
Functional connectivity (using time-series) 74
Inter-subject correlations 76
12. Research from the lab 78
Lecture 78
Quizzes 82
Intro Quiz 82
Neuro Quiz 82
BOLD Quiz 83
GLM Quiz 84
GLM Experimental Design Quiz 85
MR Physics Quiz 86
MVPA Quiz 86
Comparisons Quiz 87
Encoding models Quiz 88
Preprocessing Quiz 89
Connectivity Quiz 90
2
,Summary
_________________________________________________________________________
fMRI terminology
● Structural (T1) images
○ High spatial resolution
○ No temporal resolution (static image)
○ Can distinguish between different types of tissue
● Functional (T2) images
○ Lower spatial resolution
○ Higher temporal resolution (static image)
○ Can relate changes in signal to an experimental manipulation
● Different types of slices:
○ Sagittal: left to the right ear
○ Coronal: nose to the back of the head
○ Axial: top of the head to the bottom
● MRI images are typically acquired in axial slices (one at a time).
● This can performed:
○ Sequential: each adjacent slice consecutively.
○ Interleaved: every other slice, and then fills in the gaps on the second pass
● Together the slices make up a 3D brain volume.
● Field of view (FOV): the dimensions of the exact anatomic region included in a scan.
● Slice thickness
● Voxel size: cubic volume that span the 3D space of the brain
● Matrix size
● An experiment studies many subjects.
● Each subject may be scanned during multiple sessions.
● Each session consists of multiple runs.
● Each run consists of a series of brain volumes.
● Each volume is made up of multiple slices.
● Each slice contains many voxels.
● Each voxel has an intensity associated with it (corresponding time course).
3
, Brain imaging methods
Temporal resolution: the ability of an imaging Spatial resolution: the ability of an imaging
method to capture changes over time. In brain method to capture fine details and distinguish
imaging, temporal resolution is a measure of small structures in the brain. In brain imaging,
how quickly changes in brain activity can be spatial resolution is a measure of the smallest
detected and visualized. structures that can be resolved and visualized.
Low temporal High temporal Low spatial resolution High spatial resolution
resolution resolution
Study of slow Study of fast Visual Visualization of small
processes processes details
Example: Example: blood flow, Example: Example: neurons,
neural responses to blood vessels.
stimuli.
Methods: fMRI Methods: EEG Methods: EEG Methods: fMRI
MRI
● One image
● 3D {x,y,z}
● Structural measure: deals with brain structure.
fMRI
● MRI images over time.
● 4D {x,y,z,t}
● Measures the BOLD-contrast
● Functional measure: deals with cognitive/affective processes.
● Standard anatomy has a resolution of 1mm.
● Applications:
○ Localization (brain mapping): determine which regions of the brain are active during a
specific task.
○ Connectivity: determine how different brain regions are connected with one another.
○ Prediction: use a person’s brain activity to predict their perceptions, behavior, or
health status.
4
, fMRI
Pros Cons
● Non-invasive ● A strong magnet can be a dangerous
● There is generally no radioactive tracer thing.
injection involved. ● We are measuring a blood response
● Appealing balance between temporal (derivative of neural activity), not a
and spatial resolution. direct neural response.
● High-spatial resolution compared to ● The responses we record are slow, and
other imaging modalities. take multiple second to unfold
● We cannot research:
○ Moving objects
○ Levels of detail
○ Fast memory, eye movements,
etc. (activation is too short for
fMRI)
5
Blok 4 | 2022-2023
,Inhoud
_________________________________________________________________________
Summary 3
fMRI terminology 3
1. Introduction 9
Lecture 9
Additional videos 14
Introduction to fMRI 14
Analysis of fMRI Data 15
fMRI Data Structure & Terminology 16
Literature 18
Progress and challenges in probing the human brain 18
2. Neuroscience Background 20
Lecture 20
3. Imaging Brain Activations 25
Lecture 25
Additional videos 29
4. BOLD analysis 33
Lecture 33
Basic fMRI analysis 33
Linear Time Invariance 34
General Linear Model (GLM) 36
5. MR Physics and Imaging 38
Lecture 38
6. GLM: Experimental design 41
Lecture 41
Deconvolution 41
GLM 42
Experimental design 42
Voodoo correlations 45
Additional videos 46
7. Patterns in time and space 48
Lecture 48
Pattern analysis 48
Multivariate pattern analysis 49
Representational similarity analysis (RSA) 51
8. Encoding models 54
Lecture 54
Receptive fields in fMRI 54
Receptive fields? 54
Extravisual 58
9. Comparison 61
Lecture 61
Multiple comparisons 61
Comparisons 64
1
, We’re all unique 65
10. Preprocessing 67
Lecture 67
First steps on the long road from raw data to results 67
Reproducible fMRI 71
11. Connectivity 73
Lecture 73
Anatomical connectivity 73
Functional connectivity (using time-series) 74
Inter-subject correlations 76
12. Research from the lab 78
Lecture 78
Quizzes 82
Intro Quiz 82
Neuro Quiz 82
BOLD Quiz 83
GLM Quiz 84
GLM Experimental Design Quiz 85
MR Physics Quiz 86
MVPA Quiz 86
Comparisons Quiz 87
Encoding models Quiz 88
Preprocessing Quiz 89
Connectivity Quiz 90
2
,Summary
_________________________________________________________________________
fMRI terminology
● Structural (T1) images
○ High spatial resolution
○ No temporal resolution (static image)
○ Can distinguish between different types of tissue
● Functional (T2) images
○ Lower spatial resolution
○ Higher temporal resolution (static image)
○ Can relate changes in signal to an experimental manipulation
● Different types of slices:
○ Sagittal: left to the right ear
○ Coronal: nose to the back of the head
○ Axial: top of the head to the bottom
● MRI images are typically acquired in axial slices (one at a time).
● This can performed:
○ Sequential: each adjacent slice consecutively.
○ Interleaved: every other slice, and then fills in the gaps on the second pass
● Together the slices make up a 3D brain volume.
● Field of view (FOV): the dimensions of the exact anatomic region included in a scan.
● Slice thickness
● Voxel size: cubic volume that span the 3D space of the brain
● Matrix size
● An experiment studies many subjects.
● Each subject may be scanned during multiple sessions.
● Each session consists of multiple runs.
● Each run consists of a series of brain volumes.
● Each volume is made up of multiple slices.
● Each slice contains many voxels.
● Each voxel has an intensity associated with it (corresponding time course).
3
, Brain imaging methods
Temporal resolution: the ability of an imaging Spatial resolution: the ability of an imaging
method to capture changes over time. In brain method to capture fine details and distinguish
imaging, temporal resolution is a measure of small structures in the brain. In brain imaging,
how quickly changes in brain activity can be spatial resolution is a measure of the smallest
detected and visualized. structures that can be resolved and visualized.
Low temporal High temporal Low spatial resolution High spatial resolution
resolution resolution
Study of slow Study of fast Visual Visualization of small
processes processes details
Example: Example: blood flow, Example: Example: neurons,
neural responses to blood vessels.
stimuli.
Methods: fMRI Methods: EEG Methods: EEG Methods: fMRI
MRI
● One image
● 3D {x,y,z}
● Structural measure: deals with brain structure.
fMRI
● MRI images over time.
● 4D {x,y,z,t}
● Measures the BOLD-contrast
● Functional measure: deals with cognitive/affective processes.
● Standard anatomy has a resolution of 1mm.
● Applications:
○ Localization (brain mapping): determine which regions of the brain are active during a
specific task.
○ Connectivity: determine how different brain regions are connected with one another.
○ Prediction: use a person’s brain activity to predict their perceptions, behavior, or
health status.
4
, fMRI
Pros Cons
● Non-invasive ● A strong magnet can be a dangerous
● There is generally no radioactive tracer thing.
injection involved. ● We are measuring a blood response
● Appealing balance between temporal (derivative of neural activity), not a
and spatial resolution. direct neural response.
● High-spatial resolution compared to ● The responses we record are slow, and
other imaging modalities. take multiple second to unfold
● We cannot research:
○ Moving objects
○ Levels of detail
○ Fast memory, eye movements,
etc. (activation is too short for
fMRI)
5
