Estimation of FMRI response delays.

Area of Science:

• Neuroimaging
• Signal Processing
• Biomedical Engineering

Background:

• Accurate estimation of the Blood-Oxygen-Level-Dependent (BOLD) response delay is crucial for understanding neural activity.
• Current methods, like cross-correlation, can be computationally intensive and require complex interpolation.
• Assessing the BOLD response delay helps in mapping brain function and connectivity.

Purpose of the Study:

• To present an efficient Hilbert Transform-based algorithm for BOLD response delay estimation.
• To compare the performance of the Hilbert method with existing cross-correlation and onset methods.
• To investigate and minimize errors in delay estimation due to discrete signal processing.

Main Methods:

• Developed an algorithm utilizing the Hilbert Transform for BOLD response delay estimation.
• Integrated parameter estimation from cross-correlation methods with simplified interpolation.
• Analyzed errors from Discrete Fourier Transform (DFT) on short signals and proposed mitigation strategies.
• Compared Hilbert-based delay estimates against BOLD response onset detection.

Main Results:

• The Hilbert Transform method provides an efficient estimation of BOLD response delay.
• The algorithm simplifies interpolation steps inherent in cross-correlation analysis.
• The Hilbert method demonstrates reduced variance in delay estimates compared to onset methods.
• The proposed method shows decreased sensitivity to noise contamination.

Conclusions:

• The Hilbert Transform offers an efficient and robust approach for BOLD response delay estimation.
• This method improves upon traditional techniques by reducing computational load and enhancing accuracy.
• The algorithm's reduced noise sensitivity makes it valuable for analyzing fMRI data.