For the complete documentation index, see llms.txt. This page is also available as Markdown.

Imaging Glossary

General MRI

  • Artifact – An image distortion caused by motion, hardware limitations, susceptibility, or other acquisition errors.

  • Contrast – Differences in signal intensity that allow tissues to be distinguished from one another.

  • Echo Time (TE) – The time between excitation and signal measurement. TE influences image contrast.

  • Field of View (FOV) – The physical dimensions of the region being imaged.

  • File formats - The structure of data

    • DICOM (Digital Imaging and Communications in Medicine) – The standard format produced by MRI scanners. A single scan is typically stored as hundreds or thousands of individual files containing both image data and acquisition metadata.

    • NIfTI (Neuroimaging Informatics Technology Initiative) – A neuroimaging file format that stores an entire image volume in a single file and is commonly used for image processing and analysis.

  • Hyperintensity/hypointensity – a cluster of voxels at a markedly higher/lower intensity (appearing brighter) than those around it

  • Intensity – The numerical signal value assigned to a voxel and displayed as brightness or darkness in an image.

  • Repetition Time (TR) – The time between successive excitation pulses. TR influences image contrast and scan duration.

  • Resolution – A measurement of the size of the voxels in an image. Smaller voxels = higher resolution = greater image detail.

  • Sequence – The specific type of scan that was performed when the patient/participant was in the MRI scanner; eg. Diffusion weighted (DKI, DTI, FBI, etc), T1MPRAGE, T2 FLAIR, proton density (PD), MR spectroscopy (MRS), etc.

  • Signal-to-Noise Ratio (SNR) – The ratio of image signal to background noise. Higher SNR generally improves image quality.

  • Voxel – A three-dimensional image element representing a volume of tissue. The three dimensional version of a pixel.

Anatomical MRI

  • FLAIR (Fluid-Attenuated Inversion Recovery) – A T2-weighted sequence that suppresses CSF signal, which is good for the visualization of lesions and white matter abnormalities.

  • T1-Weighted Imaging – A structural MRI sequence that provides strong gray-white matter contrast and detailed anatomical information.

  • T2-Weighted Imaging – A structural MRI sequence sensitive to tissue water content and commonly used for detecting pathology.

Diffusion MRI

MRI techniques that characterize the movement of water molecules within tissue.

  • b-value – A parameter that determines the strength of diffusion weighting. Higher b-values capture more information about diffusion at the cost of a higher signal-to-noise ratio; two b-vals (b1000 and b2000) are required for DKI

  • Diffusion Metrics

    • Axial Diffusivity (AD) – Diffusion measured along the principal fiber direction, typically calculated as the first eigenvalue (λ₁).

    • Axial Kurtosis (AK) – A DKI-derived measure of non-Gaussian diffusion along the principal diffusion direction.

    • Fractional Anisotropy (FA) – A scalar measure of diffusion directionality ranging from 0 (isotropic diffusion) to 1 (highly anisotropic diffusion).

    • Kurtosis Fractional Anisotropy (KFA) – A DKI-derived measure describing the directional variation of kurtosis.

    • Mean Diffusivity (MD) – The average diffusion magnitude across all measured directions.

    • Mean Kurtosis (MK) – The average kurtosis measured across all diffusion directions.

    • Radial Diffusivity (RD) – Diffusion measured perpendicular to the principal fiber direction, typically calculated as (Ξ»β‚‚ + λ₃) / 2.

    • Radial Kurtosis (RK) – A DKI-derived measure of non-Gaussian diffusion perpendicular to the principal diffusion direction.

  • Diffusion Tensor Imaging (DTI) – A diffusion MRI technique that models diffusion using a tensor representation and assumes Gaussian diffusion.

  • Diffusion Kurtosis Imaging (DKI) – An extension of DTI that models non-Gaussian diffusion and provides information about tissue complexity.

    • Kurtosis – A statistical measure describing deviation from Gaussian diffusion.

    • Gaussian Diffusion – Diffusion behavior that follows a normal distribution and is assumed by the DTI model.

    • Non-Gaussian Diffusion – Diffusion behavior that deviates from the assumptions of the DTI model due to tissue complexity and structural barriers.

  • Eigenvalue (λ₁, Ξ»β‚‚, λ₃) – Values derived from the diffusion tensor that describe diffusion magnitude along the principal diffusion axes.

  • Eigenvector – A vector that describes diffusion orientation and fiber direction.

  • Tensor – A mathematical object used to describe complex properties such as the magnitude and direction of water diffusion within tissue.

    • Diffusion Tensor – A mathematical model used in DTI to describe the three-dimensional magnitude and directionality of water diffusion within a voxel. The diffusion tensor forms the basis for metrics such as Fractional Anisotropy (FA), Mean Diffusivity (MD), Axial Diffusivity (AD), and Radial Diffusivity (RD).

    • Kurtosis Tensor – A mathematical model used in DKI to describe deviations from Gaussian diffusion within a voxel. The kurtosis tensor characterizes the complexity and heterogeneity of tissue microstructure and forms the basis for metrics such as Mean Kurtosis (MK), Axial Kurtosis (AK), Radial Kurtosis (RK), and Kurtosis Fractional Anisotropy (KFA)

Functional MRI (fMRI)

  • Blood Oxygenation Level-Dependent (BOLD) Signal – The signal contrast used in most fMRI studies, reflecting changes in blood oxygenation associated with neural activity.

  • Functional Connectivity – Statistical relationships between activity patterns in different brain regions.

  • Resting-State fMRI – fMRI acquired while participants are not performing a specific task.

  • Task-Based fMRI – fMRI acquired while participants perform cognitive, sensory, or motor tasks.

  • Temporal Resolution – The frequency at which brain activity measurements are sampled, typically determined by the repetition time (TR).

Positron Emission Tomography (PET)

A molecular imaging technique that detects gamma rays generated by positron-emitting radioactive tracers.

  • Amyloid Burden – The amount of amyloid-Ξ² deposition estimated from PET tracer uptake.

  • Attenuation Correction – A correction procedure that compensates for signal loss caused by photon absorption in tissue.

  • Centiloid Scale – A standardized scale used to compare amyloid PET measurements across tracers and studies.

  • Dynamic PET – PET acquisition performed over time to assess tracer kinetics.

  • Off-Target Binding – Tracer uptake occurring outside the intended biological target.

  • Radiotracer – A biologically active molecule labeled with a radioactive isotope for imaging purposes; eg. Florbetaben, Florbetapir

  • Reference Region – A brain region assumed to have negligible specific tracer binding and used for normalization.

  • Standardized Uptake Value (SUV) – A semi-quantitative measure of tracer uptake normalized by injected dose and body characteristics.

  • Standardized Uptake Value Ratio (SUVR) – A ratio of tracer uptake in a target region relative to a reference region.

Image Analysis

  • Atlas – A standardized brain parcellation used to define anatomical or functional regions.

  • Image Space - The physical space within a coordinate system in which any given MR image is set

    • MNI Space – A standardized stereotaxic coordinate system commonly used in neuroimaging analyses.

    • Native Space – The original space in which an image was acquired prior to registration or normalization.

  • Mask – A binary image used to include or exclude specific voxels from an analysis.

  • Normalization – The process of warping an image or many images into a common space.

    • This may involve warping all images into an average space or, more often, warping all images into a standard space (such as MNI space)

  • Parcellation – The division of the brain into distinct regions according to an atlas or segmentation scheme.

  • Preprocessing – the process of removing image artifacts and performing necessary corrections to raw MRI data; types of corrections performed during preprocessing

    • Motion Correction – A preprocessing step used to compensate for participant movement during image acquisition.

  • Quality Control (QC) – The process of visually and quantitatively assessing image quality and preprocessing outputs.

  • Registration – The process of spatially aligning 2+ images

    • Linear Registration – Registration using translations, rotations, scaling, and shearing.

    • Nonlinear Registration – Registration using local deformations that allow anatomical structures to be more precisely aligned.

  • Resampling – The process of interpolating voxel values when changing image resolution or coordinate space.

  • ROI - Region of interest; a specific part of the brain of interest to a specific analysis pipeline

  • Segmentation – The classification of voxels into tissue classes such as gray matter, white matter, and CSF.

  • Skull Stripping (Brain Extraction) – Removal of non-brain tissues from an MRI volume during preprocessing.

  • Transform – A mathematical mapping used to move an image between coordinate spaces.

Last updated