T
Apparatus, methods, and other embodiments associated with the spectral analysis of T 2 spectral data are described. One example magnetic resonance imaging (MRI) method includes accessing a data set comprising T 2 spectral data associated with a magnetic resonance imaging (MRI) signal received from an object. The T 2 spectral data is decomposed with multi-exponential functions to determine T 2 spectra. The T 2 spectral amplitude is regularized with a regularized constant. The regularized constant is sufficiently large to smooth the T 2 spectra so adjacent pixels of the T 2 spectral data have similar characteristics. The T 2 spectral data is weighted so that the regularization is uniformly weighted for the spectral amplitudes. The T 2 spectra is partitioned into intervals corresponding to myelin water fraction (MWF), tissue water fraction (TWF), Long T2 water fraction (LWF), and cerebrospinal fluid fraction (CSF). Parametric maps are generated based, at least in part, on the T 2 spectral data.
1. A method, comprising:
acquiring T2 spectral data from an object prepared by a pulse sequence applied by an MRI apparatus; and
controlling a computer to perform a spectral analysis of the T2 spectral data, the spectral analysis comprising:
decomposing the T2 spectral data with an integral using multi-exponential functions to determine T2 spectra;
regularizing the T2 spectral data by applying a regularized constant, where the regularized constant is sufficiently large to smooth the T2 spectra so adjacent pixels of the T2 spectral data have characteristics that differ by less than a threshold amount;
partitioning the T2 spectra into intervals by determining thresholds, where a threshold is defined by the T2 spectral data acquired in a predetermined time period, and where an interval corresponds to an individual spectrum; and
generating at least one parametric map to correspond to at least one interval.
2. The method of claim 1 , the spectral analysis further comprising weighting the multi-exponential function.
3. The method of claim 2 , where the weighting is configured to be performed so that the regularization is uniformly weighted for the spectral amplitudes.
4. The method of claim 2 , where the weighting is an inverse of log-spaced T2 time interval matrix.
5. The method of claim 1 , where the pulse sequence comprises a refocusing slice and an excitation slice, and where the refocusing slice thickness is three times larger than the excitation slice thickness.
6. The method of claim 1 , where the T2 spectral data was acquired over two interleaved scans, and where an interleaved scan acquires five slices.
7. The method of claim 5 , where a slice gap between the five slices of a scan is based, at least in part, on an increased refocusing slice thickness.
8. The method of claim 1 , where the intervals correspond to myelin water fraction (MWF), tissue water fraction (TWF), Long T2 water fraction (LWF), and cerebrospinal fluid fraction (CSF).
9. The method of claim 1 , where regularizing the multi-exponential functions is performed according to a regularized non-negative least-squares approach.
10. The method of claim 1 , where the regularized constant is in a range of 1.6 to 2.0.
11. The method of claim 7 , where the time periods are based, at least in part, on T2 spectral data acquired at 3.0 Tesla, and correspond to the intervals: MWF having a time period of 0 ms-40 ms, TWF having a time period of 40 ms-200 ms, LWF having a time period of 200 ms-800 ms, and CSF having a time period of 800 ms-2000 ms.
12. The method of claim 1 , where the spectral amplitudes are weighted equally.
13. The method of claim 1 , where the filter is a nonlocal mean filter.
14. A non-transitory computer-readable medium storing computer-executable instruction that when executed by a computer control the computer to perform a method, the method comprising:
decomposing T2 spectral data with an integral using multi-exponential functions to determine T2 spectra;
regularizing the multi-exponential functions by applying a regularized constant, where the regularized constant is sufficiently large to smooth the T2 spectra so adjacent pixels of the T2 spectral data are similar to within a threshold amount;
partitioning the T2 spectra into intervals by determining thresholds, where a threshold is defined by the T2 spectral data acquired in a predetermined time period, and where an interval corresponds to an individual spectrum; and
generating a parametric map to correspond to at least one interval.
15. The non-transitory computer readable medium of claim 14 , the method further comprising:
weighting the T2 spectral data so that the regularization is weighted to within a desired uniformity for the spectral amplitudes.
16. The non-transitory computer readable medium of claim 14 , where an inverse of log-spaced T2 time interval matrix is applied to the T2 spectral data.
17. The non-transitory computer readable medium of claim 14 , where the at least one interval corresponds to myelin water fraction (MWF), tissue water fraction (TWF), Long T2 water fraction (LWF), and cerebrospinal fluid fraction (CSF).