pith. sign in

arxiv: 2510.25654 · v1 · submitted 2025-10-29 · ⚛️ physics.comp-ph · physics.data-an

Bayesian MINFLUX localization microscopy

Steffen Schultze , Helmut Grubm\"uller This is my paper

Pith reviewed 2026-05-18 03:21 UTC · model grok-4.3

classification ⚛️ physics.comp-ph physics.data-an
keywords MINFLUX microscopyBayesian localizationsuper-resolution microscopyphoton efficiencyfluorophore positioningnanometer precisiontargeted illumination
0
0 comments X

The pith

A Bayesian method for MINFLUX microscopy achieves 1 nm resolution with about four times fewer photons than heuristic approaches.

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

The paper develops a Bayesian framework to choose optimal scanning patterns and procedures for localizing fluorophores in MINFLUX microscopy. Current techniques depend on heuristic choices that may not extract the maximum information from each detected photon. Simulations of localization runs show that the new method reduces the photons required for 1 nm resolution by a factor of roughly four. This matters because lower photon counts can reduce damage to light-sensitive samples while still delivering high precision.

Core claim

MINFLUX microscopy localizes fluorophores with nanometer precision using targeted scanning with an illumination profile that has a minimum. Current scanning patterns and procedures are based on heuristics and may therefore be suboptimal. We present a rigorous Bayesian method that offers maximal resolutions from either minimal detected photons or minimal exposures. We estimate using simulated localization runs that this approach should reduce the number of photons required for 1 nm resolution by a factor of about four.

What carries the argument

Bayesian estimation of fluorophore position that optimizes the targeted illumination scanning pattern to maximize information per detected photon

If this is right

  • A fixed photon budget yields higher localization precision.
  • Lower light exposure reduces phototoxicity and photobleaching in live samples.
  • The same framework can minimize the number of exposures rather than total photons.
  • Optimized patterns improve performance when background noise or blinking varies.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The approach could be adapted to improve efficiency in other targeted-illumination or scanning microscopy techniques.
  • Hardware implementation would require real-time calculation of the next scan position from the running posterior.
  • The method sets a benchmark for comparing photon efficiency across different localization strategies.

Load-bearing premise

The simulated localization runs accurately model real experimental conditions including photon statistics, background noise, fluorophore blinking, and optical aberrations present in actual MINFLUX setups.

What would settle it

An experiment in a real MINFLUX microscope that measures the actual photon number needed to reach 1 nm localization precision with the Bayesian-optimized patterns versus standard patterns.

Figures

Figures reproduced from arXiv: 2510.25654 by Helmut Grubm\"uller, Steffen Schultze.

Figure 1
Figure 1. Figure 1: FIG. 1. Selected Bayesian MINFLUX localization. Shown are the posterior densities [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Current maximum a posteriori (MAP) estimates [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: ). The conventional approach achieved slightly better res￾olutions in our simulations than reported from experi￾ments [4], likely due to higher experimental noise-levels or other additional sources of uncertainty. The relative a ‖r‖ (nm) 0 100 101 102 103 EIG(r) 0.000 0.025 0.050 lo g10 𝜎 (n m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 b 𝜎 (nm) 100 101 102 D(𝜎) (nm) 0 100 101 102 μ 0.1 0.2 0.5 1 2 5 10 2… view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Median localization accuracy [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗
read the original abstract

MINFLUX microscopy allows for localization of fluorophores with nanometer precision using targeted scanning with an illumination profile with a minimum. However, current scanning patterns and the overall procedure are based on heuristics, and may therefore be suboptimal. Here we present a rigorous Bayesian that offers maximal resolutions from either minimal detected photons or minimal exposures. We estimate using simulated localization runs that this approach should reduce the number of photons required for 1 nm resolution by a factor of about four.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper introduces a Bayesian framework for optimizing MINFLUX localization microscopy, replacing heuristic scanning patterns with a principled approach to achieve maximal resolution from either minimal detected photons or minimal exposures. It estimates, via forward simulations of localization performance, that the method reduces the photon count needed for 1 nm resolution by a factor of about four relative to existing procedures.

Significance. If the simulation results prove robust, the work could meaningfully advance super-resolution microscopy by lowering photon budgets, thereby reducing phototoxicity and acquisition times in live-cell imaging. The shift from heuristics to a Bayesian formulation provides a reproducible, optimizable foundation that may generalize across MINFLUX variants and related targeted-illumination techniques.

major comments (1)
  1. [Simulation results / Methods] The central performance claim (factor-of-four photon reduction for 1 nm resolution) rests exclusively on simulated localization runs. The manuscript provides no quantitative specification of the photon-arrival statistics, background-noise model, fluorophore blinking kinetics, or residual aberration terms used in those runs (see the simulation/results section and associated methods). Any systematic mismatch with real MINFLUX data would directly scale the reported improvement factor.
minor comments (1)
  1. [Abstract] Abstract: the phrase 'a rigorous Bayesian that offers' is grammatically incomplete; it should read 'a rigorous Bayesian framework/method that offers'.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive report and the opportunity to clarify our work. We address the single major comment below and agree that additional detail will improve the manuscript.

read point-by-point responses

  1. Referee: [Simulation results / Methods] The central performance claim (factor-of-four photon reduction for 1 nm resolution) rests exclusively on simulated localization runs. The manuscript provides no quantitative specification of the photon-arrival statistics, background-noise model, fluorophore blinking kinetics, or residual aberration terms used in those runs (see the simulation/results section and associated methods). Any systematic mismatch with real MINFLUX data would directly scale the reported improvement factor.

    Authors: We agree that the current manuscript lacks sufficient quantitative detail on the simulation parameters, which limits reproducibility and makes it harder for readers to judge robustness. In the revised version we will expand the Methods section to specify the photon-arrival process (inhomogeneous Poisson with intensity set by the MINFLUX doughnut profile), the background model (Poisson-distributed with mean calibrated to typical experimental counts), the fluorophore blinking kinetics (two-state Markov chain with on-time and off-time distributions drawn from literature values for common dyes), and residual aberrations (small random Zernike terms added to the PSF). Because both the Bayesian and heuristic scanning strategies are evaluated under identical forward models, the reported relative improvement factor of approximately four is expected to be insensitive to uniform rescaling of noise parameters; only differential mismatches between the two strategies would alter the ratio. We therefore view the addition of these explicit specifications as a straightforward and necessary revision rather than a fundamental change to the central claim. revision: yes

Circularity Check

0 steps flagged

No circularity: performance estimate obtained from independent forward simulations

full rationale

The paper derives a Bayesian localization procedure and then evaluates its photon-efficiency gain exclusively through forward Monte-Carlo localization runs that simulate photon detection, background, and scanning under the new versus heuristic patterns. This constitutes an out-of-sample predictive test rather than a parameter fit or self-referential definition. No equation or claim reduces the reported factor-of-four improvement to a fitted quantity or to a self-citation chain; the simulation model is stated as an external forward model whose fidelity is an empirical question separate from the derivation itself. The derivation chain therefore remains self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities are described. The Bayesian method presumably relies on priors over fluorophore position and intensity plus a likelihood model for photon counts, but these are not detailed.

pith-pipeline@v0.9.0 · 5590 in / 1073 out tokens · 39643 ms · 2026-05-18T03:21:59.743459+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

13 extracted references · 13 canonical work pages

  1. [1]

    Balzarotti, Y

    F. Balzarotti, Y. Eilers, K. C. Gwosch, A. H. Gynn˚ a, V. Westphal, F. D. Stefani, J. Elf, and S. W. Hell, Nanometer resolution imaging and tracking of fluores- cent molecules with minimal photon fluxes, Science355, 606 (2017)

    work page 2017

  2. [2]

    Eilers, H

    Y. Eilers, H. Ta, K. C. Gwosch, F. Balzarotti, and S. W. Hell, MINFLUX monitors rapid molecular jumps with su- perior spatiotemporal resolution, Proceedings of the Na- tional Academy of Sciences115, 6117 (2018)

    work page 2018

  3. [3]

    K. C. Gwosch, J. K. Pape, F. Balzarotti, P. Hoess, J. El- lenberg, J. Ries, and S. W. Hell, MINFLUX nanoscopy delivers 3D multicolor nanometer resolution in cells, Na- ture Methods17, 217 (2020)

    work page 2020

  4. [4]

    Schmidt, T

    R. Schmidt, T. Weihs, C. A. Wurm, I. Jansen, J. Rehman, S. J. Sahl, and S. W. Hell, MINFLUX nanometer-scale 3D imaging and microsecond-range tracking on a common fluorescence microscope, Nature Communications12, 1478 (2021)

    work page 2021

  5. [5]

    S. J. Sahl, J. Matthias, K. Inamdar, M. Weber, T. A. Khan, C. Br¨ user, S. Jakobs, S. Becker, C. Griesinger, J. Broichhagen, and S. W. Hell, Direct optical measure- ment of intramolecular distances with angstrom preci- sion, Science386, 180 (2024)

    work page 2024

  6. [6]

    Scheiderer, Z

    L. Scheiderer, Z. Marin, and J. Ries, MINFLUX achieves molecular resolution with minimal photons, Nature Pho- tonics19, 238 (2025)

    work page 2025

  7. [7]

    Scheiderer, H

    L. Scheiderer, H. von der Emde, M. Hesselink, M. We- ber, and S. W. Hell, MINSTED tracking of single biomolecules, Nature Methods21, 569 (2024)

    work page 2024

  8. [8]

    Weber, M

    M. Weber, M. Leutenegger, S. Stoldt, S. Jakobs, T. S. Mihaila, A. N. Butkevich, and S. W. Hell, MINSTED fluorescence localization and nanoscopy, Nature Photon- ics15, 361 (2021)

    work page 2021

  9. [9]

    J. O. Wirth, L. Scheiderer, T. Engelhardt, J. Engel- hardt, J. Matthias, and S. W. Hell, MINFLUX dissects the unimpeded walking of kinesin-1, Science379, 1004 (2023)

    work page 2023

  10. [10]

    Deguchi and J

    T. Deguchi and J. Ries, Simple and robust 3D MINFLUX excitation with a variable phase plate, Light: Science & Applications13, 134 (2024)

    work page 2024

  11. [11]

    Bezanson, A

    J. Bezanson, A. Edelman, S. Karpinski, and V. B. Shah, Julia: A Fresh Approach to Numerical Computing, SIAM Review59, 65 (2017)

    work page 2017

  12. [12]

    P. K. Mogensen and A. N. Riseth, Optim: A mathe- matical optimization package for Julia, Journal of Open Source Software3, 615 (2018). Implementation details A proof-of-principle implementation of the Bayesian approach presented in the Julia programming language

    work page 2018

  13. [13]

    The posteriorP k was approximated on a uniform Cartesian grid with positionsx ij = (x i, yj) and asso- ciated probabilitiesp ij

    here is available athttps://gitlab.gwdg.de/ sschult/minflux. The posteriorP k was approximated on a uniform Cartesian grid with positionsx ij = (x i, yj) and asso- ciated probabilitiesp ij. The Bayesian updates were per- formed according to eq. (2) and after each step the pos- terior was normalized such that P ij pij = 1. The initial grid consisted of 60×...

    work page