Resting State fMRI Functional Connectivity Analysis Using Dynamic Time Warping.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Data Analysis

Background:

• Traditional functional connectivity analysis relies on static, zero-lag correlations of brain signals.
• This approach fails to capture dynamic changes and time-lag structures in brain networks.
• Existing methods are limited in their ability to detect subtle group differences.

Purpose of the Study:

• To introduce Dynamic Time Warping (DTW) as a novel method for assessing functional connectivity.
• To account for signal non-stationarity and time-lags in resting-state fMRI data.
• To improve the stability and sensitivity of functional connectivity analysis.

Main Methods:

• Applied Dynamic Time Warping (DTW) distance to evaluate functional connectivity strength.
• Utilized simulated fMRI data to test DTW's performance against traditional correlation analysis.
• Validated the method on resting-state fMRI data from repeated measurements and a public dataset for classification tasks.

Main Results:

• DTW effectively captures dynamic interactions and is less sensitive to global noise than correlation analysis.
• DTW analysis demonstrated increased stability and reduced within-subject variability in connectivity patterns.
• DTW-based classifiers significantly outperformed correlation-based classifiers in detecting group differences.

Conclusions:

• Dynamic Time Warping (DTW) provides a robust and sensitive approach for analyzing resting-state functional connectivity.
• DTW enhances the characterization of functional brain networks by accounting for temporal dynamics.
• This method offers a promising alternative for neuroimaging data analysis and group comparisons.