Mbirn: Diffusion MRI calibration data dissemination

From NAMIC Wiki
Revision as of 13:26, 18 December 2006 by Andy (talk | contribs) (Update from Wiki)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
Home < Mbirn: Diffusion MRI calibration data dissemination
  • Short Description
    • This database contains structural and diffusion MRI data from two healthy volunteers that were scanned multiple times on a 1.5T Philips system to investigate reproducibility of diffusion-derived metrics, such as fractional anisotropy maps.
    • Keywords: diffusion MRI, human brain, reproducibility


  • Access to and potential value of this data
    • This database as well as processed data will be open to all BIRN researchers. While we work investigating how the reproducibility of diffusion derived metrics like fractional anisotropy depend on signal-to-noise ratio (SNR) and the number of gradient orientations, the database could be used for many other purposes. For example, we can study reproducibility of tractography. We can test effect of different types of fitting routine and measure their stability. We can map regional differences of reproducibility due to physiological noise such as pulsation. The data could also be a great resource as a working dataset for processing and/or visualization tool developments.


  • Longer Overview
    • Diffusion MRI data can be used to generate an array of quantitative maps such as fractional anisotropy (FA), eigenvectors, eigenvalues, and apparent diffusion coefficient (ADC) maps, which are based on physical properties of water molecules. This, in theory, means that values obtained with different imaging parameters are comparable. However, it is an important research effort to confirm that these values are reproducible and insensitive to various imaging parameters and thus values from different studies are comparable. DTI has a large degree of freedom in their parameter setting. This research aims to characterize impact of imaging parameters on DTI results.
    • In the first phase of this research, we chose to investigate the effects of the following parameters on the reproducibility of DTI derived data: signal-to-noise ratio (SNR), number of gradient orientation, b-value, and echo time. This first database is designed to assess the impact of SNR and gradient orientation.
    • The data consists of DTI datasets with b = 1,000 and 30 gradient orientations. With 5 b0 images, one complete dataset has 35 images. We repeated the measurement 7 - 15 times. From this dataset, we can generate DTI-based contrast using from 1 datasets (lowest SNR) up to 15 datasets (highest SNR). Also, by using only a subset of the 30 orientation, we can study impacts of the number of gradient orientation. For example, we can extract data with 10 orientations and signal average 3 times (total 30 diffusion-weighted images). We can compare this data with 1 set of 30 orientation data. We can also measure the effect of the number of b0 images. To measure the reproducibility of various analysis results, the entire study was repeated 3 times at different occasions using the same subject.


  • Data description
    • The DTI data can be divided into two categories:
Whole Brain Coverage for Volunteer A (Male, age 33)
 * Three sessions, each on a different day
 * In each session:
 ** 56 transverse slices (2.5 mm slice thickness, no slice gap) extending from the base of the cerebellum to the top of the skull
 ** Scan time per complete 56 slice acquisition = 5 min:9 sec
 ** N = 7 repetitions (except in session 3, which had eight repetitions)
 
Partial Brain Coverage for Volunteer B (Male, age 24)
 * Three sessions, each on a different day
 * In each session:
 ** 25 transverse slices (2.5 mm slice thickness, no slice gap) extending from below the corpus callosum  to the central semiovale
 ** Scan time per complete 25 slice acquisition = 2 min:18 sec
 ** N = 15 repetitions
 


  • ** A single DTI dataset including 30 diffusion weighted volumes (according to the Jones30 diffusion encoding scheme) and 1 minimally weighted (b0) volume (composed of 5 scanner averaged volumes). This acquisition had a ratio of acquired diffusion weighted volumes to minimally weighted volumes of 6:1.


  • ** Acquisition parameters
      • Scanner: 1.5T MR unit (Gyroscan NT, Philips Medical Systems, Best, The Netherlands)
      • SENSE Head Coil - Single shot, multi slice, spin echo, EPI readout
      • FOV = 240 mm x 240 mm in plane
      • Scan Matrix = 96 x 96
      • Reconstructed Matrix = 256 x 256
      • Sense P reduction (AP) = 2
      • Foldover / Fat shift = AP / P
      • TE / TR = 100 ms / (8135ms for 56 slices, 3632 ms for 25 slices)
      • Max Diffusion Gradient = 19.5 mT / m (gradient duration = maximum)
      • b factor = 1000 s/mm^2
      • Diffusion Encoding Scheme = Jones30
      • SPIR fat suppression = yes
      • Cardiac gating = no



  • References


  • Technical Contact
    • Susumu Mori (susumu@mri.jhu.edu)


  • Acknowledgements