A standardized file format and open-source analysis framework for Brillouin microscopy data

Carlo Bevilacqua; Kareem Elsayad; Pierre Bouvet; Robert Prevedel; Salvatore La Cavera III; Sebastian Hambura

REVIEW 3 major objections 5 minor 24 references

A standardized file format and open-source analysis framework for Brillouin microscopy data

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Two open file formats aim to make Brillouin microscopy data reproducible and comparable across labs.

desk verdict A serious, well-executed standards proposal for Brillouin microscopy; the broad compatibility claim is ahead of the demonstrated conversion path. read the letter →

arxiv 2509.07566 v1 pith:5E4BBN25 submitted 2025-09-09 physics.optics

Carlo Bevilacqua , Sebastian Hambura , Pierre Bouvet , Salvatore La Cavera III , Kareem Elsayad , Robert Prevedel This is my paper · ORCID

classification physics.optics

keywords BrillouinmicroscopyfileformatstandardizationZarrHDF5open-sourcesoftwarereproducibilitybiomechanicalimaging

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Brillouin microscopy maps the mechanical properties of living samples by measuring the tiny frequency shift of scattered light, but each lab currently stores and processes its data differently, making results hard to compare when the relevant differences are only a few percent. This paper proposes a standardized file format, .brim, built on Zarr v3, that stores raw spectra, the frequency axis, spatial scanning information, and analysis results together with their metadata. A complementary format, .brimX, stores whole multi-parameter experiments in HDF5, and the authors provide Python libraries, a browser-based viewer, a desktop GUI, and a converter so data from spontaneous or stimulated scattering, confocal or line-scanning systems, and time- or frequency-domain detection can share one analysis pipeline. The central claim is that adopting these formats would make Brillouin studies reproducible, shareable, and compatible with existing bioimaging tools, easing the field's transition from niche technique to mainstream bioimaging.

What carries the argument

The PSD/Frequency/Spatial-scan triple: every Brillouin dataset is represented as a power spectral density array, a frequency axis, and a flexible Scanning group mapping spectra to voxel positions, while different fits live in separate Analysis_m groups. This canonical decomposition is what lets one set of tools ingest data from diverse instrument designs and keeps processing assumptions machine-readable.

What would settle it

The claim would be falsified by an existing Brillouin modality whose raw data cannot be represented as a PSD array plus Frequency axis and Scanning group without loss or flattening—for example, a time-domain instrument with nonuniform sampling that requires per-pixel calibration metadata—such that faithful round-tripping through .brim and .brimX is impossible.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central contribution is a standardization scheme: every Brillouin measurement is reduced to a common pair of arrays, a power spectral density (PSD) and a frequency axis, plus a flexible Scanning group linking spectra to spatial positions and optional Analysis groups storing fitted quantities like shift and linewidth. The .brim format encodes this scheme in Zarr v3, chosen for cloud compatibility and parallel I/O, with defined metadata groups and extensible 'subtypes' for instrument-specific additions. A second format, .brimX, stores complete experiments in HDF5 with hierarchical attributes and a 'Brillouin_type' tag so any HDF5-compatible tool can read it. The a

Load-bearing premise

The metadata schema and array structures are complete enough to represent every Brillouin modality without losing instrument-specific calibration details; the paper demonstrates this on four datasets but not on the full range of existing instruments.

Editorial extensions

If this is right

Published Brillouin datasets would carry the optical configuration, calibration, and fitting model needed to reproduce reported shift and linewidth values.
Cross-lab comparisons become feasible even when the parameter changes of interest are only a few percent, addressing the reproducibility problem documented by the field's consensus statement.
Commercial Brillouin microscopes could export data that existing open-source viewers and pipelines read directly, lowering the barrier for new users.
Cloud storage plus a browser-based viewer enables sharing large datasets without requiring local software installation.
The same PSD/Frequency structure could extend to other hyperspectral imaging modalities and to correlating Brillouin with Raman or fluorescence measurements.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Not in the paper: if the PSD/Frequency pair becomes the shared exchange representation, each new instrument type only needs one converter into the canonical pair, not bespoke analysis software; the paper demonstrates four such conversions but leaves this one-converter-per-modality principle implicit.
Not in the paper: machine-readable metadata and standardized array layouts would make pooled multi-lab datasets usable for training machine-learning spectral analysis models, an application the authors do not discuss.
Not in the paper: the framework's generality could be tested by saving a non-Brillouin hyperspectral dataset, such as Raman or fluorescence emission maps, in .brim; the paper notes this possibility in passing but does not perform the experiment.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a real contribution. Brillouin microscopy needs a common data container, and the authors have shipped working code, a sensible schema, and four example datasets spanning several instrument types. If you work in this field, you should look at it. The weak spot is not the idea, it's the gap between the "supports all modalities" claim and what the conversion layer actually preserves.

What's new: the .brim format (Zarr v3) with required PSD/Frequency/Scanning groups, the .brimX HDF5 wrapper, and the BrimConverter/BrimView/HDF5_BLS toolchain. Building on OME-Zarr and the recent consensus statement is the right move; the metadata recommendations are concrete and align with the field's push toward FAIR data. The code is open, documented, and the examples cover confocal, line-scan, stimulated, and FTBM modalities. That's real evidence of effort and partial support for the claim.

Soft spots, in order of weight. First, the multidimensional claim. The spec says PSDs can be multidimensional (angle-resolved, etc.), but BrimConverter enforces shape (z, y, x, f) when converting from .brimX. There is no round-trip test and no example with a non-spatial extra axis. So the format may support such data in principle, but the main conversion path will flatten or drop those axes. That needs a clarifying statement and ideally a test case. Second, the evidence for completeness is four imaging datasets; there is no systematic test that the schema captures all essential metadata for every modality. The paper acknowledges it's a starting point, which is fair, but the abstract's "diverse range" phrasing overreaches. Third, figure data is "available upon request"—for a reproducibility-focused paper, that undercuts the message. And the BrimView URL in the code block is mangled. Minor, but sloppy.

My verdict: the central premise holds up. This is community infrastructure, not a physics breakthrough, and it should be judged as such. The issues are addressable with a revision that tightens the modality claims and adds a conversion test. I'd send it to peer review and ask for those changes.

Referee Report

3 major / 5 minor

Summary. The paper proposes two standardized file formats for Brillouin microscopy data: .brim, a Zarr-based format for imaging data, and .brimX, an HDF5-based format for broader spectroscopy and multi-parameter experiments. It presents an open-source software stack including the brimfile Python library, the HDF5_BLS library and GUI, the BrimConverter library for .brim↔.brimX conversion, the BrimView browser-based visualization/analysis application, and a Napari plugin. The authors claim the framework supports the diversity of Brillouin microscope implementations — spontaneous and stimulated scattering, confocal and line-scanning modalities, and time- or frequency-domain detection — and illustrate this with four example datasets. The paper positions the work as a community standardization effort to improve FAIR data practices, reproducibility, and cross-laboratory comparability.

Significance. If the framework delivers on its claims, it would be a valuable contribution to the Brillouin microscopy community. The software is genuinely open-source, documented, and accompanied by public example datasets, which is a practical strength. The proposal addresses a real and increasingly urgent need for standardized data storage and metadata in a field where reproducibility is a known bottleneck. The use of Zarr for cloud compatibility and HDF5 for flexibility is reasonable, and the provision of GUI tools lowers adoption barriers for non-expert users. However, the paper's central breadth claim — support for all major Brillouin modalities — is only partially demonstrated, and the file-conversion design appears to impose shape constraints that may contradict the stated support for multidimensional spectra. The manuscript is best viewed as a promising standards proposal that needs additional validation and scoping before the broadest claims can be accepted.

major comments (3)

[SI, 'Brim file format for imaging data' and 'BrimConverter library'] The .brim specification states that the PSD array 'can optionally store multidimensional spectra', including angle-resolved measurements. However, the BrimConverter section states that HDF5_BLS arrays of arbitrary dimensions are 'reshaped into the array-structure enforced by Brimfile, e.g., an array containing PSDs should be of shape (z, y, x, f)'. If this constraint is enforced generally, then any measurement with an extra physical axis (angle, time delay, pump-probe phase) is either collapsed into the spatial dimensions or lost in round-trip conversion. The manuscript provides no round-trip test and no example with a non-spatial extra dimension. This is load-bearing because the paper's central claim is support for the full diversity of Brillouin implementations. Please either relax the shape constraint, define an explicit representation for parameter axes, or provide a concrete example
[SI, 'Example .brim files' and Abstract] The paper claims support for 'the diverse range of Brillouin microscope implementations — including spontaneous and stimulated scattering, confocal and line-scanning modalities, and time- or frequency-domain detection'. The support is demonstrated with four imaging datasets only: one line-scanning spontaneous, one confocal spontaneous, one pulsed stimulated, and one full-field FTBM. All four are spatial-imaging datasets; none exercises a non-spatial extra parameter axis, a pure spectroscopy workflow, or a non-imaging acquisition. This makes the breadth claim stronger than the evidence. I recommend either adding at least one example that exercises the multidimensional/parameter-axis path or reformulating the claim to state explicitly that these modalities are supported for imaging data and that non-imaging and higher-dimensional cases are planned/partially supported.
[SI, 'Metadata' and 'BrimX file format'] The metadata schema (brim_file_metadata.md and attributes_v1.0.xlsx) is asserted to be sufficient for reproducible acquisition and processing, but the manuscript contains no systematic validation that the attribute set covers instrument-specific calibrations, arbitrary experimental conditions, or the variations described in the cited consensus statement (Ref. 11/8). The paper also notes that additional groups/attributes can be added without breaking compatibility, but such extensions are not readable by standard tools unless defined as subtypes. This is a correct design compromise, but the completeness of the core metadata schema remains untested. Please provide a validation table mapping the attributes used in the four example datasets to the schema, and state explicitly which metadata fields are required, recommended, or optional, so that the reproducibility claim is verifiable.

minor comments (5)

[Code availability] The BrimView URL appears corrupted in the manuscript text: 'https://biobrillouin.org/brihttps://github.com/prevedel-lab/BrimView'. Please correct this.
[SI, 'Brimfile Python library'] The class name 'AnalysysResults' contains a typo; it should be 'AnalysisResults'. This affects the API documentation and code examples.
[General] The term 'multidimensional spectra' is used in the .brim specification but never formally defined. Please clarify whether it means multiple parameter axes, multiple spectral regions, or something else, and show how such data are dimensioned and interpreted in the file.
[SI, 'BrimView'] Supplementary Table 1 uses fill characters (stars, checks, circles) without a legend. Please make the table self-explanatory or add a legend.
[Data availability] The statement 'Exemplary imaging data presented in figures are available upon request' is not consistent with the FAIR principles the paper advocates. If possible, provide persistent identifiers for the figure datasets as well.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is a standards/software proposal with no derivation-to-input reduction; self-citations are illustrative data, not load-bearing.

full rationale

This manuscript proposes file formats and software tools; it contains no mathematical derivation, no fitted parameters renamed as predictions, and no uniqueness theorem imported from the authors' prior work. The example datasets are partly drawn from the authors' own earlier publications, but those are independent, already-published measurements used as demonstrations of interoperability, not as premises that force the paper's conclusions. The closest concern, noted by the skeptic, is that BrimConverter reshapes arbitrary HDF5_BLS arrays into the (z, y, x, f) structure enforced by Brimfile, which may conflict with the claim that .brim PSDs can store multidimensional spectra with extra external-parameter axes. That is a potential limitation or internal inconsistency in the generality claim, not a circular reduction: no prediction is made equivalent to an input by construction. Defining a file format is inherently stipulative, which is normal for a standards proposal rather than circular. Therefore the score is 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 2 invented entities

No free parameters are fit. The paper's contribution is a data standard; its axioms are design choices about the suitability of Zarr and HDF5 and about the completeness of its metadata schema. The two invented entities are software formats, both backed by public code and sample data.

assumptions (4)

domain assumption Zarr v3 is an appropriate container for Brillouin spectral imaging data, offering efficient cloud storage and metadata handling.
Stated in the .brim file format section when justifying the choice over HDF5.
domain assumption HDF5 is a robust, widely supported format that cannot host executable code, which is assumed to mitigate security risks.
Stated in the BrimX section of the Supplementary Information.
domain assumption The Brillouin community currently lacks standardized formats and would benefit from the proposed standard.
Motivates the work in the Introduction, citing reference 11 (consensus statement).
domain assumption The metadata attributes defined in the proposed spreadsheets sufficiently capture instrument configuration and processing steps needed for reproducibility.
This is the key assumption behind the FAIR and reproducibility claims; it is asserted but not empirically validated across all modalities.

invented entities (2)

.brim file format independent evidence
purpose: Standard hierarchical Zarr-based container for Brillouin spectral images and analysis results.
The format specification, Python library (brimfile), and four public example files provide outside verification handles.
.brimX file format independent evidence
purpose: HDF5-based container for general Brillouin spectroscopy data, experiments, and multi-parameter studies.
Spec is documented in the HDF5_BLS Tutorial, library is open-source, and example .h5 files can be created with the provided code.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A standardized file format and open-source analysis framework for Brillouin microscopy data." pith.science (2026). https://pith.science/paper/5E4BBN25

@misc{pith2026250907566,
  author       = {Pith},
  title        = {Pith review of: A standardized file format and open-source analysis framework for Brillouin microscopy data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5E4BBN25}},
  note         = {Machine review of arXiv:2509.07566}
}

read the original abstract

Brillouin microscopy is rapidly emerging as a powerful technique for imaging the mechanical properties of biological specimens in a label-free, non-contact manner. We present a standardized file format and open-source tools to facilitate the uptake and analysis of Brillouin microscopy related data and to unify this growing field.

Figures

Figures reproduced from arXiv: 2509.07566 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

[1]

& Antonacci, G

Prevedel, R., Diz-Muñoz, A., Ruocco, G. & Antonacci, G. Brillouin microscopy: an emerging tool for mechanobiology. Nat. Methods 16, 969–977 (2019)

work page 2019
[2]

Kabakova, I. et al. Brillouin microscopy. Nat. Rev. Methods Prim. 4, 8 (2024)

work page 2024
[3]

Berne, B. J. & Pecora, R. Dynamic Light Scattering: With Applications to Chemistry, Biology, and Physics. (Dover Publications, 2000)

work page 2000
[4]

& Fioretto, D

Palombo, F. & Fioretto, D. Brillouin Light Scattering: Applications in Biomedical Sciences. Chem. Rev. 119, 7833–7847 (2019)

work page 2019
[5]

& Bilenca, A

Remer, I., Shaashoua, R., Shemesh, N., Ben-Zvi, A. & Bilenca, A. High-sensitivity and high-specificity biomechanical imaging by stimulated Brillouin scattering microscopy. Nat. Methods 17, 913–916 (2020)

work page 2020
[6]

& Scarcelli, G

Fiore, A., Bevilacqua, C. & Scarcelli, G. Direct three-dimensional measurement of refractive index via dual photon-phonon scattering. Phys. Rev. Lett. 122, 103901 (2019)

work page 2019
[7]

J., Akhenblit, P., McKiernan, K

Koski, K. J., Akhenblit, P., McKiernan, K. & Yarger, J. L. Non-invasive determination of the complete elastic moduli of spider silks. Nat. Mater. 12, 262–267 (2013)

work page 2013
[8]

Scarcelli, G. et al. Noncontact three-dimensional mapping of intracellular hydromechanical properties by Brillouin microscopy. Nat. Methods 12, 1132–1134 (2015)

work page 2015

Show all 24 references

[9]

& Kabakova, I

Poon, C., Chou, J., Cortie, M. & Kabakova, I. Brillouin imaging for studies of micromechanics in biology and biomedicine: from current state-of-the-art to future clinical translation. J. Phys. Photonics 3, 012002 (2020)

2020
[10]

& Scarcelli, G

Zhang, J. & Scarcelli, G. Mapping mechanical properties of biological materials via an add-on Brillouin module to confocal microscopes. Nat. Protoc. 16, 1251–1275 (2021)

2021
[12]

Moore, J. et al. OME-Zarr: a cloud-optimized bioimaging file format with international community support. Histochem. Cell Biol. 160, 223–251 (2023)

2023
[13]

Besson, S. et al. Bringing Open Data to Whole Slide Imaging. in 3–10 (2019). doi:10.1007/978-3-030-23937-4_1

2019 doi
[14]

Wilkinson, M. D. et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci. Data 3, 160018 (2016)

2016
[15]

Goldberg, I. G. et al. The Open Microscopy Environment (OME) Data Model and XML file: open tools for informatics and quantitative analysis in biological imaging. Genome Biol. 6, R47 (2005)

2005
[16]

multidimensional

Sarkans, U. et al. REMBI: Recommended Metadata for Biological Images—enabling reuse of microscopy data in biology. Nat. Methods 18, 1418–1422 (2021). Figure 1 Fig. 1: A standardized file format and open-source analysis framework for Brillouin microscopy data. a) Overview of Br...

2021
[17]

Brillouin

BrimX file format for general Brillouin spectroscopy data and local usage Brillouin scattering serves a variety of applications beyond spatial imaging, some of which may lead to data protection issues. In an effort to unify the storage of data falling in these scenarios, we pr...

2002
[18]

Bevilacqua, C. et al. High-resolution line-scan Brillouin microscopy for live imaging of mechanical properties during embryo development. Nat. Methods 20, 755–760 (2023)

2023
[19]

& Prevedel, R

Bevilacqua, C., Diz-mu, A. & Prevedel, R. A 3D Brillouin microscopy dataset of the in-vivo zebrafish eye. Data Br. 105427 (2020) doi:10.1016/j.dib.2020.105427

2020
[20]

Yang, F. et al. Pulsed stimulated Brillouin microscopy enables high-sensitivity mechanical imaging of live and fragile biological specimens. Nat. Methods 20, 1971–1979 (2023)

1971
[21]

& Prevedel, R

Bevilacqua, C. & Prevedel, R. Full-field Brillouin microscopy based on an imaging Fourier-transform spectrometer. Nat. Photonics 19, 494–501 (2025)

2025
[22]

Lee, S. et al. A Case Study on Parallel HDF5 Dataset Concatenation for High Energy Physics Data Analysis. (2022)

2022
[23]

Dougherty, M. T. et al. Unifying biological image formats with HDF5. Commun. ACM 52, 42–47 (2009)

2009
[24]

& Michalet, X

Ingargiola, A., Laurence, T., Boutelle, R., Weiss, S. & Michalet, X. Photon-HDF5: An Open File Format for Timestamp-Based Single-Molecule Fluorescence Experiments. Biophys. J. 110, 26–33 (2016)

2016
[25]

Bouvet, P. et al. Consensus statement on Brillouin light scattering microscopy of biological materials. Nat. Photonics 19, 681–691 (2025)

2025

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.

[1] [1]

& Antonacci, G

Prevedel, R., Diz-Muñoz, A., Ruocco, G. & Antonacci, G. Brillouin microscopy: an emerging tool for mechanobiology. Nat. Methods 16, 969–977 (2019)

work page 2019

[2] [2]

Kabakova, I. et al. Brillouin microscopy. Nat. Rev. Methods Prim. 4, 8 (2024)

work page 2024

[3] [3]

Berne, B. J. & Pecora, R. Dynamic Light Scattering: With Applications to Chemistry, Biology, and Physics. (Dover Publications, 2000)

work page 2000

[4] [4]

& Fioretto, D

Palombo, F. & Fioretto, D. Brillouin Light Scattering: Applications in Biomedical Sciences. Chem. Rev. 119, 7833–7847 (2019)

work page 2019

[5] [5]

& Bilenca, A

Remer, I., Shaashoua, R., Shemesh, N., Ben-Zvi, A. & Bilenca, A. High-sensitivity and high-specificity biomechanical imaging by stimulated Brillouin scattering microscopy. Nat. Methods 17, 913–916 (2020)

work page 2020

[6] [6]

& Scarcelli, G

Fiore, A., Bevilacqua, C. & Scarcelli, G. Direct three-dimensional measurement of refractive index via dual photon-phonon scattering. Phys. Rev. Lett. 122, 103901 (2019)

work page 2019

[7] [7]

J., Akhenblit, P., McKiernan, K

Koski, K. J., Akhenblit, P., McKiernan, K. & Yarger, J. L. Non-invasive determination of the complete elastic moduli of spider silks. Nat. Mater. 12, 262–267 (2013)

work page 2013

[8] [8]

Scarcelli, G. et al. Noncontact three-dimensional mapping of intracellular hydromechanical properties by Brillouin microscopy. Nat. Methods 12, 1132–1134 (2015)

work page 2015