Resampling Strategies to Estimate Mean Concentrations from Low SNR In Vivo MR Spectra
(2007) In Proceedings of the International Society for Magnetic Resoance in Medicine 15.- Abstract
- A resampling method for quantitative MRS is presented that retains absolute concentrations. Averaging the spectra from a subset improved the concentration
estimates via increased SNR. Using short-TE STEAM spectra from human thalamus, the influence of subset size was systematically studied and verified
against conventional bootstrap re-sampling. Even small subsets provided considerably better signal-to-noise ratio and reliability. The standard deviation or
confidence intervals can be derived by correcting the variance over all subset estimates for the inherent redundancy. Subsets of about half the group size are
recommended, because of their large number and the almost normal distribution of mean concentration estimates.
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/c14fd670-5dfd-402a-9413-e16d08e97234
- author
- Helms, Gunther LU ; Savic, Ivanka and Dechent, Peter
- organization
- publishing date
- 2007
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- keywords
- MRS quantification resampling
- host publication
- Proc. Intl. Soc. Mag. Reson. Med.
- series title
- Proceedings of the International Society for Magnetic Resoance in Medicine
- volume
- 15
- article number
- 1394
- publisher
- International Society for Magnetic Resonance in Medicine
- ISSN
- 1524-6965
- 1524-6965
- language
- English
- LU publication?
- yes
- id
- c14fd670-5dfd-402a-9413-e16d08e97234
- date added to LUP
- 2017-01-19 16:46:26
- date last changed
- 2020-08-05 02:19:31
@inproceedings{c14fd670-5dfd-402a-9413-e16d08e97234, abstract = {{A resampling method for quantitative MRS is presented that retains absolute concentrations. Averaging the spectra from a subset improved the concentration<br/>estimates via increased SNR. Using short-TE STEAM spectra from human thalamus, the influence of subset size was systematically studied and verified<br/>against conventional bootstrap re-sampling. Even small subsets provided considerably better signal-to-noise ratio and reliability. The standard deviation or<br/>confidence intervals can be derived by correcting the variance over all subset estimates for the inherent redundancy. Subsets of about half the group size are<br/>recommended, because of their large number and the almost normal distribution of mean concentration estimates.}}, author = {{Helms, Gunther and Savic, Ivanka and Dechent, Peter}}, booktitle = {{Proc. Intl. Soc. Mag. Reson. Med.}}, issn = {{1524-6965}}, keywords = {{MRS quantification resampling}}, language = {{eng}}, publisher = {{International Society for Magnetic Resonance in Medicine}}, series = {{Proceedings of the International Society for Magnetic Resoance in Medicine}}, title = {{Resampling Strategies to Estimate Mean Concentrations from Low SNR In Vivo MR Spectra}}, url = {{https://lup.lub.lu.se/search/files/20027804/ISMRM_2007_1394.pdf}}, volume = {{15}}, year = {{2007}}, }