Baseline Correction of the Human ¹H MRS(I) Spectrum Using T₂* Selective Differential Operators in the Frequency...

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Spectroscopy

Background:

• Baseline distortion from water and fat signals complicates ¹H Magnetic Resonance Spectroscopic Imaging (MRS(I)) of the human brain.
• Accurate quantification of metabolites in MRS(I) is essential for neurological disorder diagnosis and research.

Purpose of the Study:

• To develop and validate an effective preprocessing technique for calibrating baseline distortions in ¹H MRS(I) data.
• To improve the accuracy of metabolite quantification by mitigating interference from water and fat signals.

Main Methods:

• Designed a T₂* selective filter using differential filtering in the frequency domain.
• Determined filter parameters by matching T₂* selectivity profiles with metabolite T₂* profiles.
• Evaluated the filter using simulated MRS spectral signals and real human ¹H MRSI data.
• Quantitatively analyzed filtered data using LCModel software.

Main Results:

• The T₂* selective filter effectively removed water and fat signal-induced baseline distortions from simulated MRS(I) data.
• Analysis of real human ¹H MRSI data showed significant improvements in the accuracy of metabolite quantification.
• Reduced Cramer-Rao Lower Bound (CRLB) levels indicate enhanced precision in metabolite estimation.

Conclusions:

• The proposed T₂* selective filtering is a reliable method for preprocessing ¹H MRS(I) data.
• This technique enhances the reliability and accuracy of metabolite quantification in human brain MRS(I) studies.
• The method offers a valuable tool for improving the diagnostic and research applications of MRS(I).