A realistic model of resting-state BOLD spectra in white matter
Abstract Resting-state BOLD signals in white matter show a spectral elevation near 0.06 Hz on an exponentially decaying background, and white-matter hemodynamic response functions (HRFs) have a prominent initial dip. We model the resting signal as shot noise: randomly timed neural events are filtered by the HRF, so that the spectrum is the input spectrum multiplied by the squared HRF gain, plus instrumental noise. The initial dip and post-stimulus undershoot remove low-frequency power, making the white-matter HRF a band-pass filter with maximum gain near 0.05 Hz. One set of physiological parameters fits both a measured white-matter HRF (R² = 0.993) and the group-mean white-matter spectrum (R² = 0.987), with a resting transit time of about 4 s and a metabolic response weighted about twice as heavily as in gray matter. Deepening the dip raises the filter gain and moves its maximum toward 0.06 Hz, linking the dip to the spectral feature. A narrower line in individual spectra is reproduced either by an underdamped HRF that matches the measured response over its first 20 s or by a rhythmic input near 0.06 Hz. The model identifies which HRF properties could explain the decline of the peak with age.
Authors
- John C. Gore (ORCID: https://orcid.org/0000-0002-9480-0900)
- Tzu-Wei Li (ORCID: https://orcid.org/0000-0002-6506-369X)
- Yurui Gao (ORCID: https://orcid.org/0000-0002-4339-5351)
- Adam W. Anderson (ORCID: https://orcid.org/0000-0002-3788-6138)
- Kurt G. Schilling (ORCID: https://orcid.org/0000-0003-3686-7645)
- Zhaohua Ding (ORCID: https://orcid.org/0000-0002-1805-2955)
- Warren W. Boschen
- Muwei Li (ORCID: https://orcid.org/0000-0003-3596-9287)
Institutions
- Vanderbilt University (US)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-12
- DOI
- https://doi.org/10.21203/rs.3.rs-11473671/v1
- Primary Topic
- Advanced MRI Techniques and Applications
- Type
- preprint
Funders
- National Institutes of Health
- McDonnell Center for Systems Neuroscience