Ch8 - Signal, Noise, and Preprocessing of fMRI Data Flashcards
blood-oxygenation-level-dependent (BOLD) contrast
differences in signal on t2*-weighted images as a function of the amount of deoxygenated hemoglobin
preprocessing
computational procedures that are applied to fMRI data following image reconstruction but before statistical analysis
- intended to reduce variability in data not associated with experimental task
- preparation for statistical testing
signal
meaningful changes in some quantity
- fMRI: incl. changes in intensity associated with the BOLD response across a series of t2* images
noise
nonmeaningful changes in some quantity
Signal-to-Noise ratio (SNR)
relative strength of a signal compared to other sources of variability in data
raw SNR
ratio between MR signal intensity associated with the sample (brain) and thermal noise that is measured outside the sample
contrast-to-noise ratio (CNR)
magnitude of the intensity difference between different quantities divided by variability of their measurement
voxel-wise analysis
evaluation of statistical test at the level of individual voxels
region-of-interest (ROI) analysis
evaluations of hypotheses about the functional properties of brain regions
- often chosen to reflect a priori distinctions within the brain
functional signal-to-noise ratio (functional SNR)
ratio between the intensity of signal associated with changes in brain function and the variability in the data due to all sources of noise
functional resolution
ability to map measured physiological variation to underlying mental processes
partial volume effect
combination, within voxel, of signal contributions from 2+ distinct tissue types or functional regions
spatial extent
number of active voxels within a cluser of activation
susceptibility artifacts
signal losses on T2* -dependent images due to magnetic field inhomogeneities in regions with air/air adjacent tissue
sinuses
- air filled cavaties in the skull
- long venous channels that form the primary draining system for the brain
power spectrum
representation of the strenght of different frequency components within a signal
- Fourier transform converts a signal into its power spectrum
thermal noise
fluctuations in MR signal intensity over space or time that are caused by the thermal motion of electrons within the sample or scanner hardware
system noise
fluctuations in MR signal intensity over space or time that are caused by imperfect functioning of the scanner hardware
scanner drift
slow changes in voxel intensity over time
(part of system noise)
physiological noise
fluctuations in MR signal intensity over space and time due to physiological activity of the human body
- motion, respiration, cardiac activity, metabolic reactions
aliasing
the sampling of a signal at a rate insufficent to resolve the highest frequencies that are present
- energy at those frequencies become artifically expressed at lower frequencies, distorting the measured signal
quality assurance (QA)
set of procedures designed to identify problems with fMRI data so that they don’t compromise experimental analyses
phantom
object used for testing MR systems
- mostly filled with liquids or gel with known properties so problems with the scanner system can be readily identified
ascending/descending slice acquisition
collection of data in consecutive order such that slices are acquired sequentially from one end of the imaging volume to the other
interleaved slice acquisition
collection of data in alternating order
- first acquired from the odd-numbered slices, then even numbered slices
- minimizes influence of excitation pulses on adjacent slices
temporal interpolation
estimation of the value of a signal at a time point that was not originally collected
- using data from nearby time point
coregistration
spatial alihnment of 2 images or image volumes
reference volume
target image volume to which other image volumes are aligned
rigid-body transformation
spatial transformation that doesn’t change the size or shape of object
- 3 translational parameters & 3 rotational parameters
translation
movement of an object along an axis in space
rotation
turning of an object around an axis
cost function
quantity that determines the amount of residual error in a comparison
mutual information
in context of MRI:
amount of information about 1 image that is provided by knowledge of another
spatial interpolation
estimation of the intesity of an image at a spatial location that was not originally sampled, using data from nearby locations
shimming coils
electromagnatic coils that compensate for inhomogeneities in the static magnetic field
magnetic field mapping
the collection of explicit information about the strength of the magnetic field across space
field map
an image of the intensity of the magnetic field across space
bias field estimation
technique for estimating inhomogeneities in the magnetic field based on intensity variations in collected images
segmentation
the process of partitioning an image into constituent parts
- typically types of tissue or topographical divisions
normalization
transformation of MRI data from an individual subject to match the spatial properties of standardized image
E.g.:
averaged brain derived from a sample of many individuals
stereotaxic space
precise mapping of system using 3D coordinates
Talairach space
commonly used space for normalization of fMRI data
- coordinates based on measurements from a single postmortem human brain
MNI space
commonly used space for normalization of fMRI data
- coordinates derived from an average of MRI structural images from more than 100 individuals
cytoarchitecture
the organization of the brain based on cell structure
filter
within MRI: algorithm for removing temporal or spatial frequency components of data
Nyquist frequency
highest frequency that can be identified in a digitally sampled signal
-defiend as one-half of the sampling rate
task frequency
rate of presentation of a periodic experimental task
prewhitening
application of filters to remove autocorrelation in a time series of data
- in fMRI: applied before removing task-unrelated noise
precoloring
introduction of autocorrelation to a time series of fata so that the time series will have known statistical properties
(less common)
spatial smoothing
blurring og fMRI data across adjacent voxel to improve the validty of statistical testing and maximizing functional SNR
- at cost of spatial resolution
matched filters
principle that a filter of the same frequency as the signal of interest provides the max. signal-to-noise ratio
brain extraction
step during preprocessing that removes unwanted parts of imaging volume
E.g.:
Skull
multiple comparison problem
increase in number of false-positive results with an increasing number of statistical tests
- particular consequence of voxel-wise fMRI analysis
