Denoising physiological data collected during multi-band, multi-echo EPI sequences

Area of Science:

• Neuroimaging
• Physiological monitoring
• Biomedical engineering

Background:

• Physiological data collection during fMRI aids data cleaning and psychophysiological research.
• MRI pulse sequences often introduce artifacts into physiological recordings (ECG, EDA).
• Existing artifact filtering methods need adaptation for advanced fMRI sequences.

Purpose of the Study:

• To evaluate BIOPAC Systems' artifact filtering recommendations for fMRI.
• To extend these recommendations for multiband, multi-echo fMRI sequences.
• To provide practical tools for artifact mitigation in physiological fMRI data.

Main Methods:

• Assessed BIOPAC physiological data (ECG, EDA) during fMRI.
• Evaluated artifact filtering for single-band, single-echo sequences.
• Developed and tested digital filters for multiband, multi-echo sequences, incorporating slice timing, multiband factor, and repetition time.

Main Results:

• Artifact magnitude and frequency vary with fMRI pulse sequence parameters.
• Digital filters effectively mitigate MRI pulse sequence artifacts in physiological data.
• The proposed filtering approach is applicable to various fMRI sequence types.

Conclusions:

• Artifacts in physiological fMRI data can be successfully mitigated using tailored digital filters.
• The developed methods improve the quality of physiological recordings for fMRI.
• Open-source tools and notebooks facilitate the implementation of these denoising techniques.