🧬 biology

A Distinct Second Spectral Peak in White Matter BOLD Signals: Mathematical Modeling, Spatial Distribution and Age-Related Evolution

This study identifies and characterizes a robust, age-dependent "second peak" at approximately 0.06 Hz in white matter BOLD signals, proposing it as a novel, organized physiological phenotype linked to perivascular clearance mechanisms that offers new insights into white matter functional integrity and brain aging.

Original authors: sijing wu, Muwei Li, Yurui Gao, Lyuan Xu, Yu Zhao, John C Gore, Zhaohua Ding

Published 2026-08-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: sijing wu, Muwei Li, Yurui Gao, Lyuan Xu, Yu Zhao, John C Gore, Zhaohua Ding

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

For decades, the brain's white matter was treated as the silent wiring of the human mind. In the popular imagination of how we think and feel, the focus has always been on the gray matter, the outer layer where neurons fire and conversations happen. The white matter, a dense network of insulated cables connecting these neurons, was viewed merely as the passive infrastructure that carries the signals. When scientists used functional magnetic resonance imaging to watch the brain at rest, they often treated the signals coming from this white wiring as background noise, a nuisance to be filtered out so they could study the more active gray matter. However, a growing body of evidence suggests that this quiet tissue is actually humming with its own meaningful activity, carrying physiological information that has been overlooked. The question is no longer whether this tissue is active, but what kind of activity it holds and what it tells us about the health of the brain.

A team of researchers at Vanderbilt University Medical Center and other institutions has now uncovered a specific, rhythmic pattern hidden within the white matter signals that had previously gone unnoticed. By analyzing high-quality brain scans from 120 healthy young adults, they discovered a distinct "second peak" in the frequency of the blood flow signals within the white matter. While the signals in the gray matter typically fade away smoothly as the frequency increases, the white matter signals show a clear, organized bump or rise around a specific speed of fluctuation, roughly 0.06 times per second. This bump sits on top of the usual background signal, creating a unique signature that is stable and consistent across different people. The researchers found that this feature is not random noise; it is a robust, measurable trait that appears in the vast majority of the white matter in the brain, particularly in the deep, central pathways that connect the two halves of the brain and link the top and bottom.

To understand what this bump meant, the scientists built a mathematical model to describe the shape of the signal. They found that the white matter signal is composed of two parts: a general background that slowly fades out, and this specific, rhythmic peak that stands out clearly. When they looked at the data from individual people, they saw that while the exact height and width of this peak varied from person to person, the location of the peak remained steady at that same frequency. This consistency suggests that the peak is a fundamental feature of how the brain's white matter functions, acting like a reliable fingerprint for each individual. The researchers also mapped where this peak was strongest in the brain. They found that it was most prominent in the deep, central tracts of the white matter, such as the corpus callosum, and became less distinct as they moved toward the outer edges of the brain. This spatial pattern mirrors the major highways of the brain's wiring, suggesting that the phenomenon is tied to the large-scale structure of the tissue.

The study also looked at the fluid that surrounds the brain, known as cerebrospinal fluid, and found a striking similarity. The signals from this fluid, which fills the hollow spaces in the center of the brain, showed the same distinct peak at the same frequency. This connection between the white matter and the surrounding fluid hints that the rhythm might be driven by a shared physical process, possibly related to how fluids move and clear waste from the brain. The researchers noted that this frequency range is similar to rhythms observed during sleep when the brain clears out metabolic waste, suggesting that this signal might be a window into the brain's cleaning and maintenance systems.

Perhaps the most revealing part of the study came from looking at how this signal changes as people get older. The researchers analyzed brain scans from a separate group of people ranging from their thirties to their nineties. They found that the distinct peak in the white matter signals was strong in younger adults but began to fade as people aged. In the group of people aged 81 to 90, the peak had virtually vanished. This gradual decline suggests that the mechanism driving this rhythm is sensitive to the aging process. Unlike the gray matter, which did not show this clear peak at any age, the white matter's signal appears to be a unique indicator of the brain's physiological state over time.

These findings offer a new way to look at the brain's white matter, not just as a passive cable, but as an active tissue that reflects the body's underlying physiological rhythms. The presence of this specific signal, its connection to the fluid surrounding the brain, and its disappearance with age suggest that it could serve as a new tool for understanding brain health. While the exact biological cause of this rhythm is still being investigated, the researchers propose that it may be linked to the movement of fluids and blood vessels that help keep the brain clean. By identifying this distinct pattern, scientists may have found a new, non-invasive way to monitor the functional integrity of the brain's wiring and track how it changes as we age, opening a door to understanding brain health that was previously closed.

Technical Summary: A Distinct Second Spectral Peak in White Matter BOLD Signals

Problem Statement
While brain white matter (WM) is anatomically critical, it has historically been overlooked in resting-state functional magnetic resonance imaging (rs-fMRI), often treated as nuisance background. Although recent work has established that WM blood-oxygenation-level-dependent (BOLD) signals contain physiologically meaningful information, primarily through functional connectivity analyses, the intrinsic spectral architecture of these signals remains poorly understood. A specific, puzzling observation noted in prior work is a distinct spectral peak around 0.05–0.06 Hz in WM power spectra, which deviates from the monotonic exponential decay typically observed in gray matter (GM). Furthermore, a similar peak was recently identified in cerebrospinal fluid (CSF) signals. The precise nature, spatial organization, and physiological drivers of this "second spectral peak" in WM require comprehensive characterization.

Methodology
The study utilized high-quality rs-fMRI data from the Human Connectome Project (HCP) S1200 release (120 healthy young adults, 22–36 years) and the HCP-Aging cohort (participants aged 31–90 years).

Preprocessing: Data underwent standard HCP pipeline processing, including motion correction and physiological noise regression. Individualized tissue masks for GM, WM, and CSF were generated using SPM12 probability maps, with CSF masks eroded to prevent overlap.
Spectral Estimation: Power spectral density (PSD) was estimated for each voxel using Welch's method within the 0.01–0.20 Hz frequency range.
Mathematical Modeling: To characterize the second peak, the amplitude spectrum was fitted using a combined exponential-Lorentzian model: $S(f) = c + ae^{-bf} + \frac{A}{\pi}\frac{\gamma}{(f-f_0)^2+\gamma^2}$ . This model superimposes a Lorentzian function (representing the narrow peak) onto an exponentially decaying background.
Validation: Model adequacy was assessed against a pure exponential model using the Akaike Information Criterion (AIC) and the coefficient of determination ( $R^2$ ). Valid peaks were defined by specific constraints on center frequency ( $0.03–0.15$ Hz), width, height, and fit stability.
Analysis: The study examined the peak at group, subject, and voxel levels, analyzing spatial distributions, energy contributions, and age-related evolution across six decade-based groups in the HCP-Aging cohort.

Key Results

Existence and Modeling: A robust second spectral peak was identified in WM BOLD signals, centered at approximately 0.06 Hz. The combined exponential-Lorentzian model provided a significantly better fit ( $R^2 = 0.9998$ ) than a pure exponential model ( $R^2 = 0.9274$ ), with a substantial improvement in AIC ( $\Delta AIC = 171.98$ ).
Quantitative Characterization: In the 0.03–0.09 Hz band, the Lorentzian component contributed an average of 11.8% of the total spectral energy in WM, significantly higher than in GM (6.0%). CSF signals exhibited a comparable energy contribution and peak frequency, suggesting a shared underlying mechanism.
Individual-Level Phenotype: The peak was observable in the vast majority of subjects. While the center frequency was stable (concentrated around 0.06 Hz), parameters such as peak height, width, and energy contribution showed inter-subject variability, supporting its use as an individual-level phenotype.
Spatial Organization: The prevalence of valid peaks was not uniform; it was highest in deep WM regions (corpus callosum, internal capsule, corona radiata) and decreased toward peripheral regions. The center frequency displayed an anterior-posterior gradient, mirroring patterns previously found in the cortex.
Age-Related Evolution: In the HCP-Aging cohort, the height of the second spectral peak in WM decreased systematically with age. The peak was prominent in the 31–60 age range, showed significant reduction in the 61–80 range, and was virtually absent in the 81–90 age group. CSF signals showed a similar aging-driven decline, whereas GM signals did not exhibit a pronounced second peak at any age.

Significance and Claims
The authors propose that the white matter second spectral peak is a genuine, robust, and physiologically meaningful feature.

Physiological Interpretation: The peak's presence in both WM and CSF, its specific frequency range (~0.06 Hz), and its preference for deep WM tracts proximal to ventricles suggest it may reflect low-frequency fluid-related dynamics. The authors hypothesize a link to glymphatic or perivascular clearance mechanisms, potentially involving vascular expansion and fluid displacement, similar to dynamics observed during sleep. They note that the dichotomy between WM (showing the peak) and GM (lacking a clear peak) may be due to WM's lower vascular density and reduced susceptibility to the rapid cognitive updates that overwhelm global fluctuations in GM during wakefulness.
Methodological Contribution: The study establishes a quantitative framework for modeling this spectral phenomenon, moving beyond simple power summation.
Clinical Potential: The authors suggest this spectral signature could serve as a novel, noninvasive biomarker for investigating white matter functional integrity and brain aging. Unlike current biomarkers derived almost exclusively from gray matter, this feature offers a new dimension for probing ubiquitous low-frequency physiological dynamics in the brain.

The paper maintains a modest stance, acknowledging that while the data supports the existence of the peak and its correlation with aging and CSF dynamics, the specific biological mechanisms remain speculative and require further investigation involving higher temporal resolution fMRI and physiological recordings.

More like this