Effects of different slow paced breathing regimes on cerebrospinal fluid (CSF) oscillations
Nasseri, P., Banks, M., Li, A. Y., Lewis, L. D., Yoo, H. J., Mather, M.
neuroscience
Quick verdict
Claims infraslow paced sighing (0.02-0.03 Hz) strengthens CSF-BOLD coupling in waking humans; well-powered, but sigh-vs-frequency confound and motion limits.
Summary
This preprint reports a within-subject fMRI study testing whether different paced respiratory regimes modulate CSF oscillations in 35 healthy adults aged 18-30. Participants completed four conditions — paced breathing at 0.09-0.11 Hz and 0.05-0.07 Hz, and paced sighing at 0.05-0.07 Hz and 0.02-0.03 Hz — using fast fMRI (TR 367 ms) with the slab edge placed at the fourth ventricle to capture CSF inflow. All four manipulations increased respiratory volume, low-frequency respiratory power, fourth-ventricle CSF spectral power below 0.11 Hz, and BOLD global signal amplitude relative to rest, with CSF power indistinguishable among conditions. The key finding is a dissociation: infraslow paced sighing at 0.02-0.03 Hz yielded the strongest CSF-negative-BOLD-derivative coupling (r = 0.548 at 3 s lag), exceeding all other conditions (Bonferroni p < 0.05), while paced breathing at 0.05-0.07 Hz — despite the greatest respiratory power — fell below rest (r = 0.247). Cardiac total spectral power was lowest under infraslow sighing. The authors argue the imposed frequency, not response amplitude, drives coupling, possibly via sleep-like infraslow vasomotion. Main limitations, acknowledged by the authors: no CO2 or blood-pressure monitoring, no EEG/pupillometry arousal measures, fourth-ventricle-only measurement, and an unresolvable confound between sighing form and infraslow frequency; the infraslow condition also showed roughly threefold more motion-corrected volumes than rest (8.69% vs 2.51%). Data are promised on OpenNeuro upon publication.
Review complete
Assessment at a glance
6
risk flags identified
● high 0
● medium 3
● low 3
5 strengths · 5 weaknesses · 5 author questions
AI text signal = Binoculars zero-shot detector score over 20 sampled passages (Fast-DetectGPT co-reported) — absence does not prove human authorship, and a flag does not imply misconduct.
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- All four paced respiratory conditions (breathing at 0.09-0.11 Hz and 0.05-0.07 Hz; sighing at 0.05-0.07 Hz and 0.02-0.03 Hz) increase respiratory volume, low-frequency respiratory power, fourth-ventricle CSF spectral power below 0.11 Hz, and BOLD global signal amplitude relative to rest, with CSF power not differing among conditions.
- Paced sighing at 0.02-0.03 Hz produces significantly stronger coupling between CSF and the zero-thresholded negative BOLD derivative (r = 0.548 at a 3-second lag) than every other condition including rest (r = 0.406), while paced breathing at 0.05-0.07 Hz falls significantly below rest (r = 0.247).
- The coupling enhancement is dissociated from both respiratory and cardiac oscillation amplitude, since infraslow sighing had the greatest coupling but the lowest cardiac total spectral power among manipulations.
- Imposing an infraslow respiratory oscillation during wakefulness may recruit the same infraslow neurovascular-CSF dynamics implicated in NREM glymphatic clearance.
Strengths
Adequately powered within-subject design
A priori G*Power calculation (N=31 for 95% power, f=0.25) with 39 recruited and 35 analyzed; frequencies were individualized from each participant's resting cardiac spectrum via Kubios, reducing between-subject variance in entrainment.
Dissociative experimental logic
Crossing breathing vs sighing at the same 0.05-0.07 Hz band and comparing against 0.02-0.03 Hz sighing shows coupling does not track respiratory power (PB 0.05-0.07 Hz had the highest LF respiratory power yet the weakest, below-rest coupling), supporting a frequency-specific rather than effort-based interpretation.
Measurement approach appropriate to the question
Conservative statistics for most measures
Candid limitations section
Weaknesses
Motion burden is highest in the critical condition
Frequency and respiratory form are perfectly confounded
Inconsistent statistical treatment of the headline measure
Ad hoc signal transformation untested
Reporting errors in tables and figures
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
Report preview
Scope & Claims
This bioRxiv preprint (posted September 17, 2026; USC IRB UP-20-01022) asks whether voluntarily imposed respiratory oscillations can recruit, during wakefulness, the infraslow (0.02-0.03 Hz) neurovascular-CSF dynamics associated with NREM glymphatic clearance. Thirty-five healthy adults (18 female; aged 18-30, from 39 recruited) each completed four ~8-minute paced-respiration scans — paced breathing at 0.09-0.11 Hz and 0.05-0.07 Hz, paced sighing at 0.05-0.07 Hz and 0.02-0.03 Hz — plus a ~7-minute rest scan in randomized order. Central claims: (1) all four manipulations raise respiratory volume, low-frequency respiratory power, fourth-ventricle CSF spectral power (<0.11 Hz) and BOLD global signal amplitude versus rest, with CSF power not differing among conditions (F(4,136)=20.063 for CSF power); (2) only infraslow paced sighing strengthens CSF coupling to the zero-thresholded negative BOLD derivative (r = 0.548 at 3-s lag vs rest r = 0.406), exceeding all other conditions (Bonferroni p < 0.05); (3) coupling is dissociated from respiratory and cardiac power — paced breathing at 0.05-0.07 Hz produced the greatest respiratory power yet the weakest, below-rest coupling (r = 0.247) — so the imposed frequency, not response amplitude, drives coupling.
The remaining sections cover Methodology Assessment, Statistical Evidence, Results & Reproducibility, Limitations & Risks, and Verdict & Recommendations.
- All 5 strengths & 5 weaknesses, analyzed in detail
- 6 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 5 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
Can slow breathing increase CSF movement while you are awake?
According to this preprint, yes: all four paced respiratory conditions tested (breathing at 0.09-0.11 Hz and 0.05-0.07 Hz; sighing at 0.05-0.07 Hz and 0.02-0.03 Hz) increased fourth-ventricle CSF oscillatory power below 0.11 Hz compared with rest in 35 healthy adults, although the four conditions did not differ from each other in CSF power.
What is special about sighing at 0.02-0.03 Hz?
Only infraslow paced sighing (one sigh roughly every 33-50 seconds) strengthened the coupling between CSF inflow and the negative rate of change of the BOLD signal (r = 0.548 at a 3-second lag versus 0.406 at rest). This frequency band matches sleep-related vasomotion linked to waste clearance. Notably, normal paced breathing at 0.05-0.07 Hz — despite producing the strongest respiratory oscillations — showed coupling below rest.
Does this study prove breathing exercises clear waste like amyloid from the brain?
No. Measurements were of bulk CSF displacement in the fourth ventricle only. The study offers indirect evidence about drivers of CSF movement and did not measure clearance, perivascular flow, or interstitial fluid; the authors themselves state that the link to enhanced clearance remains untested.
Is the study data or code available?
The authors state that data will be made available on OpenNeuro.org upon journal publication. No analysis code availability is stated in this preprint, and processing spanned MATLAB, FSL, AFNI, ArtRepair and Kubios, so code transparency will matter for reproduction.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- Figures could not be analyzed from the extracted text; magnitudes and directions rely on reported statistics and captions.
- Supplemental Table 1 contains apparent typographical/reporting inconsistencies (SE 5.967 vs 55.967; p printed as 0.11), so exact pairwise values for some contrasts are uncertain.
- Only Supplemental Tables 1-2 are present in the extracted text; any additional supplementary material is not reviewable here.
- The metadata abstract was truncated mid-sentence, but the full abstract appears intact in the extracted body text.
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.