FatNav Motion Correction Sharpened Perivascular Space Imaging on 7T MRI
Subject motion remains one of the most stubborn sources of degradation in MRI, and the problem becomes sharper at 7 tesla, where higher resolution can come at the cost of greater sensitivity to movement. In this study, FatNav-based prospective motion correction was tested for perivascular space imaging on a T2-weighted 3D turbo spin echo sequence, with a direct focus on whether motion compensation could preserve both visual interpretability and quantitative image fidelity.
The work addressed a gap that matters in daily neuroimaging practice. Perivascular spaces are small fluid-filled structures, and their appearance is vulnerable to blur, ghosting, and contrast loss when the head shifts during acquisition. If those structures cannot be seen cleanly, both clinical interpretation and downstream measurement of PVS burden become less reliable.
Motion Burden Was Far Greater Under Instructed Movement
Ten healthy participants were scanned with and without prospective motion correction while performing gradual instructed motion. To move beyond a controlled volunteer setting, 62 patients were also scanned after being told to keep their heads still. Among those patients, 28 underwent paired PMC-on and PMC-off scans, while 34 had PMC-on scans only.
The measured motion separated the two settings by a wide margin. Instructed motion produced a mean motion score of 10.3 ± 5.3 mm, compared with 1.47 ± 0.97 mm in patients with involuntary motion. In practical terms, the volunteer condition created a much harsher test of the correction strategy, while the patient cohort reflected the more familiar clinical reality of smaller, less predictable head drift during routine imaging.
Motion on this scale is not a cosmetic issue. Even modest displacement can alter the apparent width, count, and conspicuity of perivascular spaces, especially when the structures are only a few voxels wide. Once the image blurs, the signal differences between PVS and adjacent tissue collapse quickly.
Image Quality Improved When Correction Stayed On
For instructed motion, prospective motion correction improved both the radiologists’ visual scores and the quantitative image metrics. Two experienced radiologists assessed motion and overall image quality using a 4-point Likert scale, and the corrected scans consistently performed better than the uncorrected scans under deliberate movement.
The quantitative gains extended across several measures. Structural similarity index measure, peak signal-to-noise ratio, PVS count, PVS volume, and contrast-to-noise ratio between PVS and neighboring tissue were all evaluated, and PMC improved these readouts in the motion-heavy volunteer scans. The correction also weakened the negative correlations between motion and those metrics, indicating that as displacement increased, the penalty on image quality became less severe when PMC was active.
The biological rationale is straightforward. A T2-weighted sequence depends on stable spatial sampling to preserve the thin fluid-tissue interfaces that define perivascular spaces. By reducing motion-induced spatial misregistration in real time, FatNav-based PMC helps keep those interfaces aligned long enough for the sequence to encode them cleanly.
Patient Scans Showed a More Selective but Still Useful Effect
The clinical cohort was more heterogeneous, and the motion burden was lower than in the instructed-motion experiment. Even so, PMC remained beneficial. In patients with involuntary motion, the technique significantly increased contrast-to-noise ratio between perivascular spaces and adjacent tissue. The abstract did not provide the full set of patient-specific statistical values, but the direction of effect supported a measurable gain in conspicuity even when motion was subtle rather than exaggerated.
That is an important distinction for bedside imaging. In real-world scans, radiologists are not usually dealing with dramatic head turns; they are dealing with small, repeated shifts caused by discomfort, fatigue, breathing, or inability to remain completely still. Those smaller movements can still soften the image enough to reduce confidence in fine-structure assessment, and the study suggests that FatNav-based correction can recover at least part of that lost contrast.
The paired patient scans strengthen the argument further. By comparing PMC-on and PMC-off images within the same clinical context, the study isolated the value of motion correction without relying only on simulated movement. That design makes the signal more relevant to routine neuroimaging workflows, where the central question is not whether motion can be eliminated entirely, but whether useful image quality can be preserved despite it.
Why Perivascular Space Imaging Is Sensitive to Motion
Perivascular spaces are increasingly scrutinized as imaging markers in cerebral small vessel disease and other neurologic conditions, so the quality of their visualization matters beyond anatomy alone. At 7T, the high field strength should theoretically improve spatial detail, but the same field strength also magnifies artifacts when the head shifts. A technique that stabilizes acquisition in real time can therefore have an outsized effect on the final image, especially for structures at the edge of resolution.
A T2-weighted 3D turbo spin echo sequence is a logical platform for this approach because it is widely used for high-resolution structural imaging. Pairing it with FatNav-based prospective correction adds a dynamic control layer, allowing the scanner to respond to motion as it occurs rather than attempting to repair degradation after the fact. For small structures such as perivascular spaces, preemptive correction is often more effective than retrospective cleanup because the information may be lost before reconstruction can recover it.
What the Data Support—and What They Do Not
The study supports FatNav-based PMC as a practical tool for improving PVS imaging on 7T T2 MRI, especially when motion is present. It also shows that the benefit is not limited to a single metric; visual interpretation, structural similarity, signal-to-noise behavior, and PVS conspicuity all moved in the same direction when correction was active.
Still, the evidence has limits that remain tied to the study design. The cohort was relatively small, the motion challenge in volunteers was instructed rather than spontaneous, and the abstract does not provide longer-term clinical outcome data or disease-specific diagnostic endpoints. The study also does not establish whether the observed gains translate into better detection of pathologic PVS patterns across neurologic disorders.
Even with those limits, the message from the imaging data is direct: when motion is controlled prospectively, perivascular space visualization becomes cleaner, more measurable, and less vulnerable to the losses that usually accompany movement at ultra-high field strength. For 7T neuroimaging, that is a meaningful technical advance, because the structures most likely to be affected are often the very ones clinicians are trying to study in the first place.