🧠 neuroscience

Brief visual oddball events evoke CSF flow in the spinal canal

This study introduces an automated, multi-echo fMRI method to measure spinal canal cerebrospinal fluid (CSF) flow and demonstrates that brief visual oddball stimuli can provoke measurable CSF movement, which correlates with decreased BOLD activity across a diverse adult population.

Original authors: Riley, E., Voss, H. U., De Rosa, E., Anderson, A. K.

Published 2026-11-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Riley, E., Voss, H. U., De Rosa, E., Anderson, A. K.

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 ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed.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

Inside the human skull, a rigid bony case holds the brain, a soft, jelly-like organ that cannot expand. Because the skull is unyielding, the brain has no room to swell when its activity increases. Instead, the brain relies on a delicate, rhythmic exchange of fluids. When neurons fire and demand more energy, blood vessels widen to deliver fresh oxygen. This sudden influx of blood pushes the cerebrospinal fluid, a clear liquid that bathes the brain, out of the way. When the activity subsides and the blood vessels narrow again, the fluid rushes back in. This constant ebb and flow is more than just a mechanical necessity; it is the brain's primary method for flushing out waste. Scientists believe this system, which clears away toxic byproducts of daily thought, is essential for long-term brain health and may play a critical role in preventing diseases like Alzheimer's.

For decades, watching this fluid move in real time has been difficult. Traditional brain scans are designed to look at the brain tissue itself, often missing the subtle movements of the fluid that surrounds it. However, a new approach has emerged that turns a technical quirk of the scanner into a powerful tool. By positioning the bottom edge of the scan just below the brain, in the narrow channel of the spinal canal, researchers can catch the fluid as it enters and leaves the viewing area. This method allows them to see the fluid's movement without needing to inject dyes or use invasive procedures.

A team of researchers at Cornell University recently put this technique to the test in a large group of people ranging from their late teens to their mid-seventies. Their goal was twofold: to see if they could automatically find the spinal canal in the scan data without manually tracing it by hand, and to determine if brief, surprising events could trigger a measurable wave of fluid movement. They recruited 70 healthy adults and asked them to perform a simple visual task while lying inside a magnetic resonance imaging machine. The participants watched a screen where purple circles appeared at random intervals. Most of the circles were small and unremarkable, but occasionally, a slightly larger circle would appear, or a standard circle would flash with extra brightness. The participants were instructed to press a button only when they saw the larger circle. These "oddball" events were very brief, lasting only about 70 milliseconds, yet they required the brain to pay close attention and make a quick decision.

The researchers developed a computer program to automatically locate the spinal canal in the images, a task that usually requires a human expert to carefully outline the area on every single scan. Their software succeeded in finding the correct spot in more than 94 percent of the participants. Once the area was identified, the team used a special type of scan that captures images at three different speeds to separate the signal of the moving fluid from the signal of the brain tissue itself. This separation was crucial because the fluid's movement creates a different kind of signal than the brain's normal activity. They also recorded the participants' heartbeats and breathing to remove those natural rhythms from the data, ensuring that what they saw was truly the fluid moving in response to the brain's work.

The results confirmed that the brain's fluid system is highly responsive to even the smallest mental tasks. When the participants saw the surprising, larger circle and pressed their button, the fluid in the spinal canal showed a distinct pattern of movement. First, the fluid was pushed out of the spinal canal and up into the brain, a reaction that happened within a few seconds of the visual surprise. Then, as the brain settled down, the fluid flowed back in. This cycle of outflow and inflow was directly linked to the brain's activity: the stronger the brain's reaction to the surprise, the more fluid moved. The researchers found that this connection held true across all age groups, though the response was slightly less pronounced in the oldest participants.

Importantly, the study showed that this fluid movement was not just a side effect of the participants pressing the button with their fingers. The researchers compared the fluid movement during the button-pressing events with moments when the brain was equally active but no button was pressed. The fluid moved in the same way in both cases, proving that the flow was driven by the brain's internal processing of the surprise, not by the physical act of moving a hand. The study also demonstrated that this method works well even without the most advanced scanning equipment, suggesting that scientists could study this vital fluid system in many existing datasets.

The findings offer a clearer window into how the brain manages its own cleanliness. By showing that brief, salient moments of attention can trigger a wave of fluid that sweeps through the brain, the research highlights the dynamic nature of this waste-clearance system. It suggests that the brain is constantly flushing itself, even during short bursts of focus. While the study did not prove that this system fails in older age, it did show that the link between brain activity and fluid movement remains intact throughout adulthood. This work provides a new, non-invasive way to watch the brain's internal plumbing in action, opening the door to better understanding how this essential process changes as we grow older and how it might be disrupted in disease.

Technical Summary: Brief Visual Oddball Events Evoke CSF Flow in the Spinal Canal

Problem and Motivation
Measuring cerebrospinal fluid (CSF) flow is critical for understanding the glymphatic system's role in waste clearance, a process implicated in cognitive aging and neurodegenerative diseases like Alzheimer's. While CSF flow can be inferred via fMRI by exploiting edge artifacts of spin saturation at the boundary of the acquisition volume, previous methods relied on manual tracing of the fourth ventricle. This study addresses the need for a more robust, automated approach to identify CSF flow in the spinal canal—a region often captured in standard fMRI acquisitions covering the full cerebellum but historically underutilized for flow analysis. Furthermore, the authors sought to determine whether brief, salient cognitive events (visual oddballs) could evoke measurable CSF flow and how this relationship might vary across the adult lifespan.

Methodology
The study analyzed fMRI data from 70 healthy adults (ages 19–76) performing a visual oddball task. The task involved infrequent, brief (~70 ms) salient stimuli requiring a button press, alongside standard and bright stimuli.

Data Acquisition: Multi-echo echo planar imaging (EPI) was acquired at 3T. The bottom acquisition slice was positioned within the spinal canal at the foramen magnum, rather than the fourth ventricle.
Automated ROI Identification: A novel algorithm automatically identified the spinal canal region of interest (ROI) in the four lowest slices without manual tracing. The method seeded the ROI at the brightest voxel in a central-posterior window of the lowest slice and propagated it to adjacent slices based on intensity thresholds and spatial proximity.
Signal Decomposition: To isolate CSF flow from hemodynamic (BOLD) signals, the authors decomposed the multi-echo signal into two components:
- $S_0$ (Non-BOLD): Represents signal changes independent of echo time ( $TE$ ), attributed to the inflow of fresh, unsaturated spins (CSF flow) and motion.
- $R2^*$ (BOLD-like): Represents signal changes proportional to $TE$ , attributed to deoxygenated hemoglobin changes.
Denoising: Physiological noise (cardiac and respiratory) was regressed out using RETROICOR and response functions, alongside motion parameters and scanner drift.
Analysis: The study examined the correlation between spinal canal $S_0$ and the negative temporal derivative of global BOLD activity. It also used deconvolution to measure task-evoked CSF responses to oddball stimuli, comparing these against control conditions (standard stimuli, non-oddball BOLD peaks) and across age groups.

Key Results

Automated Detection: The automated ROI identification was successful in 94.3% of participants (83/88), with no significant differences in ROI size across age groups.
Validation of CSF Flow Signatures:
- The $S_0$ signal variability was 1.88 times higher in the bottom slice compared to the fourth slice, consistent with the expected gradient of spin saturation at the acquisition boundary. In contrast, the $R2^*$ signal showed only a 1.33-fold increase.
- Spinal canal $S_0$ was significantly positively correlated with the negative temporal derivative of global BOLD (indicating CSF inflow when cerebral blood volume decreases). This coupling was significant across all age groups and was stronger than correlations found in tissue $S_0$ or spinal canal $R2^*$ .
Task-Evoked CSF Flow:
- Brief visual oddball stimuli evoked a measurable CSF response: a significant outflow (decreased $S_0$ ) at 2.5 seconds post-stimulus and a significant inflow (increased $S_0$ ) at 7.5 seconds.
- This response was specific to oddball stimuli; standard and bright stimuli did not evoke significant CSF flow.
- The magnitude of CSF outflow correlated linearly with the size of the global BOLD response.
- Crucially, CSF flow patterns were similar for BOLD peaks occurring with or without a button press, suggesting the flow is driven by the global hemodynamic response rather than motor execution.
Age Effects: While the coupling between global BOLD and CSF flow remained stable across the lifespan, the absolute magnitude of the task-evoked inflow response was numerically weaker in older adults and did not reach statistical significance in that specific subgroup, though a linear decline with age was not statistically confirmed.
Multi-echo Necessity: While multi-echo decomposition is essential for distinguishing $S_0$ (flow) from $R2^*$ (BOLD) and providing rigorous controls, the primary finding of task-evoked inflow was detectable using a single echo, though with reduced sensitivity to outflow.

Significance and Claims
The paper claims to demonstrate that:

Automated Spinal Canal Analysis: It is possible to automatically identify the spinal canal in raw EPI data to measure CSF flow, removing the barrier of manual tracing and making this technique accessible for studies where the spinal canal is incidentally captured.
Stimulus-Evoked Flow: Brief, infrequent salient events are sufficient to provoke measurable CSF flow, linking cognitive task engagement directly to glymphatic dynamics.
Methodological Robustness: The combination of multi-echo decomposition and physiological denoising provides high-confidence identification of CSF flow, separating it from BOLD and motion artifacts.
Physiological Mechanism: The observed flow follows the global BOLD signal (blood volume changes), supporting the hypothesis that neural activity-driven hemodynamic changes mechanically displace CSF.

The authors maintain a modest tone regarding age-related differences, noting that while older adults showed weaker responses, the study lacked the power to conclusively differentiate age effects or isolate motor-specific contributions. The work positions itself as a methodological advancement that facilitates the study of CSF flow in standard fMRI datasets, particularly those covering the full brain and spinal canal.

Technical Summary: Brief Visual Oddball Events Evoke CSF Flow in the Spinal Canal

More like this