Topological structure and dynamics of three dimensional active nematics

Aparna Baskaran; Arvind Baskaran; Daniel A. Beller; Debarghya Banerjee; Federico Toschi; Guillaume Duclos; Itamar Kolvin; Matthew S. E. Peterson; Michael F. Hagan; Minu Varghese

REVIEW 2 major objections 4 minor 18 references

Topological structure and dynamics of three dimensional active nematics

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Three-dimensional active nematics are governed by topologically neutral disclination loops.

desk verdict A genuine experimental breakthrough in 3D active nematics, with a loop-neutrality census that is plausible but not airtight at current resolution. read the letter →

arxiv 1909.01381 v1 pith:N2ZH4474 submitted 2019-09-03 cond-mat.soft

Guillaume Duclos , Raymond Adkins , Debarghya Banerjee , Matthew S. E. Peterson , Minu Varghese , Itamar Kolvin , Arvind Baskaran , Robert A. Pelcovits

show 8 more authors

Thomas R. Powers Aparna Baskaran Federico Toschi Michael F. Hagan Sebastian J. Streichan Vincenzo Vitelli Daniel A. Beller Zvonimir Dogic

This is my paper · ORCID

classification cond-mat.soft

keywords activenematicstopologicaldefectsdisclinationloopsthree-dimensionalmatterlight-sheetmicroscopymicrotubulebundlesnematicdirectorfieldturbulence

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

Three-dimensional active nematics, the paper argues, are governed by extended disclination loops rather than the point defects that dominate two-dimensional active nematic turbulence. To show this, the authors built a bulk active nematic by dispersing extensile, kinesin-driven microtubule bundles in a passive colloidal liquid crystal, and imaged the evolving three-dimensional director field with multi-view light-sheet microscopy. They report that all 268 experimentally observed and 94 simulated disclination loops were topologically neutral, spanning the continuous family from wedge-twist to pure-twist geometries, and that loop nucleation, growth, shrinkage, self-annihilation, and reconnection with a network of disclination lines generate the turbulent-like dynamics. If this picture holds, the statistics and topology of neutral loops become the natural descriptors for three-dimensional active turbulence, extending the success of defect-based descriptions from two to three dimensions.

What carries the argument

The argument runs on the geometry and topology of disclination loops. A disclination is a line where the director winds by π about a rotation vector Ω; the angle β between Ω and the local tangent t distinguishes +1/2 wedge (β=0), −1/2 wedge (β=π), and twist (β=π/2) profiles, and Ω may rotate along the loop. The paper classifies a loop's topology by tracking the orthonormal frame {n_out, n_in, Ω} around the loop, lifting the accumulated rotation to SU(2) quaternions to obtain a Z4 index χ; χ=0 means topologically neutral and unlinked, while χ=2 would be a charged monopole loop. Experimental access comes from multi-view light-sheet microscopy plus a 3D structure-tensor reconstruction of the director, with defect loops identified as genus-1 high-distortion voxel sets and Ω measured by fitting the director winding on circuits around the loop core. The uniform-Ω, uniform-n_out family of loops is trivially χ=0, and this is the family observed almost exclusively.

What would settle it

A concrete test would be to process a large 3D director-field dataset with an automated defect finder and compute the topological index χ for every loop. The paper's claim predicts zero χ=2 loops in the bulk; finding charged loops, or showing that most detected loops disappear when the 6-micrometer smoothing window is changed, would refute it.

Watch

Extended reading notes

Core claim

The central discovery is that the elementary excitations of three-dimensional active nematics are closed, charge-neutral disclination loops. Analyzing the director field reconstructed from fluorescence images, the authors found that all 268 experimental loops and all 94 loops from two independent simulation methods carried zero topological charge: no loop had the odd or even hedgehog charge that would force it to appear with a partner. The loops are not a single geometry: their local winding character, measured by the angle β between the rotation vector Ω and the loop tangent, varies continuously around each loop, and the population spans the whole family from wedge-twist loops (γ=π/2) to pure-twist loops (γ=0). Nucleation of an isolated neutral loop is the three-dimensional analogue of unbinding a 2D ±1/2 disclination pair; wedge-twist loops grow out of bend distortions, while chaotic flows can accumulate twist distortion and relax it by creating nearly pure-twist loops. The observed dynamics—nucleation, expansion, contraction, self-annihilation, merging with, and emission from the disclination-line network—define the turbulent-like steady state.

Load-bearing premise

The load-bearing premise is that the reconstructed director field—from 6-micrometer Gaussian smoothing and 2-micrometer optical sections—resolves each disclination's winding well enough that the loop inventory and charge assignments are faithful, given that about 20% of automatically detected defects were false positives at out-of-focus edges and were removed by manual inspection.

Editorial extensions

If this is right

Because each loop is neutral, loops can nucleate and annihilate individually in the bulk, without the paired creation required for charged defects.
Wedge-twist loop nucleation provides a concrete 3D analogue of 2D ±1/2 pair unbinding, with twist segments binding the two wedge profiles into a loop.
Pure-twist loops can be born from twist distortion built up by chaotic flows, a route with no counterpart in 2D active nematics.
The measured |cos(γ)| distribution is a direct experimental observable that future theories and simulations of 3D active nematics should reproduce.
Loop reconnection with the disclination-line network, rather than only pair annihilation, is a primary mechanism for sustaining the turbulent-like steady state.

Reading between the lines

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

If neutral loops are the only bulk excitations, then defect-mediated mixing, transport, and rheology in 3D active nematics should be expressible in terms of loop creation and annihilation rates and loop reconnection events—quantities the paper measures but does not yet connect to transport coefficients.
The same imaging and topological-index pipeline could be applied to equilibrium nematic quenches, where the loop inventory predicted by Kibble–Zurek scaling could be tested directly; the paper notes the technique's suitability for quench relaxation but does not pursue it.
Varying activity, confinement, or the sign of active stress (contractile instead of extensile) might shift the loop-type distribution or even produce charged loops; measuring |cos(γ)| and χ under those conditions would test whether neutral-loop dominance is generic.
Because the detection pipeline had roughly 20% false positives near out-of-focus edges, an automated, resolution-matched detection scheme with quantified sensitivity would strengthen the loop census and could reveal rare loop geometries the manual analysis missed.

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

First the bottom line: this is the first bulk 3D active nematic with single-bundle imaging and time-resolved 3D director fields. The observation that disclination dynamics is dominated by extended, charge-neutral loops is a substantive claim, supported by two independent simulation methods and by a topological classification that is careful and mostly standard. If the loop census holds, it turns 3D active turbulence into an experimentally accessible subject and opens a new platform for soft matter.

What is new and done well: the experimental composite (fd virus liquid crystal + extensile microtubule bundles), the multi-view light-sheet imaging and structure-tensor director reconstruction, and the systematic classification of loop geometry via Ω and nout. The authors check threshold dependence for defect detection, manually verify defects against the fluorescence channel, and report a 20% false positive rate. They also show the same loop classes in two different simulations, which strengthens the qualitative picture. The distinction between wedge-twist and pure-twist loops and the nucleation mechanisms are physically reasonable.

The soft spot is the quantitative strength of 'all 268 experimental and 94 simulated loops are neutral.' The detection pipeline is built on a 6 μm Gaussian window, 2 μm optical sections, and a distortion-energy threshold. The SI gives a false positive rate but no false negative estimate. If small charged loops have curvature radii or winding variations on the scale of the smoothing window, the structure-tensor average could wash out their topology and make them appear neutral. The fact that the 94 simulated loops go through the same core-finding and quaternion routine means they do not independently certify the census; the simulation is not a ground-truth test of the imaging pipeline. This is not a fatal flaw: the isolated nucleation events (which by topological neutrality require a partner for any charged loop) and the large measured loops make it likely that neutral loops are indeed the dominant species. But the dominance claim is a quantitative claim, and the current evidence is not resolution-controlled. A synthetic-data test — injecting known charged loops into the pipeline and measuring the false negative rate as a function of loop size and smoothing — would settle it. Error bars on the loop statistics and public release of data/code would also be appropriate.

Who is this for: soft matter physicists, active matter experimentalists, and anyone working on topological defects. It deserves a serious referee; the experimental advance alone warrants it. I would recommend acceptance after the authors address the resolution/error-bar conditions, not before.

Referee Report

2 major / 4 minor

Summary. This paper reports a new experimental realization of three-dimensional active nematics, consisting of extensile microtubule bundles dispersed in a passive colloidal liquid crystal, and images the director field dynamics with multi-view light-sheet microscopy. The authors detect disclination loops, measure their local winding character (wedge/twist), and classify their topology using a quaternion-based invariant. They report that all 268 experimental and 94 simulated disclination loops are topologically neutral, and identify nucleation, growth, annihilation, merging, and splitting events. The experimental observations are supported by two independent simulations: a hybrid lattice Boltzmann method and a finite-difference Stokes solver.

Significance. If the central claim holds, this is a landmark result: it establishes that 3D active nematics are dominated by extended charge-neutral disclination loops, analogous to defect pair unbinding in 2D. The paper is strengthened by the use of multiple independent methods, a careful quaternion-based topological classification, explicit checks of threshold independence, and the demonstration of a new experimental system with single-bundle resolution. The new imaging capability and the theoretical framework will be of broad interest to the active matter and soft matter communities. However, the completeness of the loop census is the load-bearing point, and the lack of a false-negative control leaves the strong 'all loops are neutral' statement incompletely supported.

major comments (2)

[SI Sec. 4 (Data analysis) and Fig. 4D] The central claim that all 268 experimental and 94 simulated loops are topologically neutral (χ=0) is not supported without a false-negative control for the detection and classification pipeline. The director field is smoothed with a 6 μm Gaussian structure-tensor window and 2 μm optical sections, and the SI states that a larger window 'led to smoothing of the nematic director, and increased the number of undetected defects'; however, the chosen window is not tested against known charged loops. The rotation vector Ω and reference director nout are extracted on a circuit of radius 3–7 voxels (1.56–3.64 μm), a scale comparable to the smoothing window, so a charged loop whose Ω/nout winding varies on that scale could be averaged into the uniform configuration and be measured as χ=0. The SI reports that about 20% of detected defects are false positives, but no false-negative rate or loop-radius distribution is given. Because the simulated loops are processed with the same pipeline, they do not independently validate the census. I request a resolution-controlled test—for example, synthetic or simulation-generated χ=2 loops of varying radii, processed through the same code, with a reported detection rate and measured χ—before the 'dominant excitations are charge-neutral loops' claim can be considered established.
[SI Sec. 4 and Fig. S4] The topological index χ is computed from the frame F(θ) obtained at 8–12 points per loop. For loops with 'approximately uniform Ω and nout,' the classification reduces to the trivial χ=0 case, and Fig. S4 shows only the standard deviation of |cos γ|, not a comparison of the measured frame's winding with the resolution limit. The authors should report the distribution of loop radii and the ratio of the loop circumference to the structure-tensor window, and show that the measured uniformity of Ω and nout is not a smoothing artifact. This is directly relevant to the 'all loops are neutral' claim, because a pipeline that cannot resolve spatially varying Ω/nout would classify every loop as neutral by construction.

minor comments (4)

[SI Title] The supplementary title contains a typo: 'Top ological' should be 'Topological'.
[SI Eq. (8)] The expression for the local rotation vector appears garbled: the first equality has a vector ∇×n minus a vector, while the second is n × [(n·∇)n], and these are not obviously equal; please clarify the intended identity or correct the typo.
[Fig. 4D] The histograms in Fig. 4D are shown without error bars or a statement of binning; reporting the statistical uncertainty would strengthen the comparison between wedge-twist and pure-twist loop prevalence.
[SI Sec. 4 (Loop core algorithm)] The loop core tracing algorithm requires the loop core to contain 'at least ten points' before closure; please report the typical loop core length in voxels and the sensitivity of the resulting topological classification to this parameter.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the loop-neutrality claim rests on direct quaternion-index measurements and independent simulations, not on definitions or fitted parameters.

full rationale

The paper's central claim—that all 268 experimentally and 94 theoretically analyzed disclination loops are topologically neutral—is supported by direct measurement of the director-field winding structure, not by definitional substitution. The SI explicitly describes how the topological index χ is computed from measured Ω and nout along each loop: 'We measure the topological index χ of a disclination loop as follows: At roughly even intervals around a loop separated by angle Δθ, we calculate Ω and nout to find the frame F(θ)... The product q0,2π = ∏ qθ,θ+Δθ identifies the loop with one of the cases named in Sec. 3, as q0,2π must be one of the following: 1 (unlinked and neutral), −1 (unlinked and charged), or a quaternion satisfying q² = −1 (linked).' The SI statement that loops with approximately uniform Ω and nout are 'trivially χ=0 loops' is a mathematical observation about the ideal family, not the evidence for the empirical census; the manuscript reports that the measured loops carry χ=0. The experimental observation that isolated loops nucleate without partners provides an independent topological consistency check, and the numerical simulations use standard Beris-Edwards and Stokes-flow models with stated parameters rather than parameters fitted to the loop statistics. The detection-resolution limitations and manual false-positive filtering described in the SI are accuracy and completeness concerns, not circular reasoning. Therefore no load-bearing step reduces to its own inputs.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests primarily on the experimental imaging pipeline and the standard topological classification of nematic disclinations. The imaging pipeline assumes that the structure-tensor orientation of fluorescent microtubule bundles is a faithful proxy for the nematic director, and that threshold-based detection with manual verification captures the defect population. The topological classification uses standard homotopy theory, specifically the quaternion Z4 index, and the geometric family of uniform-Omega loops. Numerical simulations use standard Beris-Edwards and Stokes-flow models with chosen parameters; they serve as supporting evidence rather than as the basis of the experimental claim.

free parameters (3)

Defect detection energy threshold = 0.5 to 0.6 in distortion energy units
Chosen by hand in the SI; the paper states the results are independent of the threshold, but the threshold determines which voxels enter the defect set and therefore the loop statistics.
Structure tensor Gaussian window size = 6 micrometers
Set to balance false positives against smoothing; a smaller window gave false positives and a larger window hid defects, so the loop statistics depend on this choice.
Simulation activity and flow-aligning parameters = zeta = 0.01 (lattice Boltzmann), zeta* = 0.2 (Stokes), xi = 0.9
Standard parameter choices for extensile active nematics in the supporting simulations; not fitted to the experimental data, but the simulation evidence depends on these values.

assumptions (5)

domain assumption Beris-Edwards equations with active stress describe 3D active nematic hydrodynamics
Used for both numerical models in SI Section 2; assumes continuum nematohydrodynamics at low Reynolds number.
domain assumption One-elastic-constant approximation for the distortion energy and Frank free energy
Used in the experimental distortion-energy map and in the simulation free energy (Eq. 6); ignores differences between splay, twist, and bend constants.
standard math Standard homotopy classification of disclination loops using the quaternion Z4 index
Applied in SI Section 3; the neutrality of loops is defined through this established topological framework, referencing refs 15 to 17.
domain assumption Structure-tensor orientation of fluorescent bundles is a faithful proxy for the nematic director
The entire loop detection and topology analysis is built on this imaging assumption; the acknowledged 20% false-positive rate makes it non-trivial.
domain assumption Landau-de Gennes free energy expansion captures isotropic-nematic ordering
Used in the simulations (Eq. 6); standard in active nematic modeling but still a model assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Topological structure and dynamics of three dimensional active nematics." pith.science (2026). https://pith.science/paper/N2ZH4474

@misc{pith2026190901381,
  author       = {Pith},
  title        = {Pith review of: Topological structure and dynamics of three dimensional active nematics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N2ZH4474}},
  note         = {Machine review of arXiv:1909.01381}
}

read the original abstract

Point-like motile topological defects control the universal dynamics of diverse two-dimensional active nematics ranging from shaken granular rods to cellular monolayers. A comparable understanding in higher dimensions has yet to emerge. We report the creation of three-dimensional active nematics by dispersing extensile microtubule bundles in a passive colloidal liquid crystal. Light-sheet microscopy reveals the millimeter-scale structure of active nematics with a single bundle resolution and the temporal evolution of the associated nematic director field. The dominant excitations of three-dimensional active nematics are extended charge-neutral disclination loops that undergo complex dynamics and recombination events. These studies introduce a new class of non-equilibrium systems whose turbulent-like dynamics arises from the interplay between internally generated active stresses, the chaotic flows and the topological structure of the constituent defects.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

Castoldi, A

M. Castoldi, A. V. Popov, Protein Expression and Puriﬁcation 32, 83–88, issn: 1046-5928 (2003)

work page 2003
[2]

Hyman et al

A. Hyman et al. ,i n Molecular Motors and the Cytoskeleton (Academic Press, 1991), vol. 196, pp. 478– 485

work page 1991
[3]

S. J. DeCamp, G. S. Redner, A. Baskaran, M. F. Hagan, Z. Dogic, Nature Materials 14, 1110 (2015)

work page 2015
[4]

D. S. Martin, R. Fathi, T. J. Mitchison, J. Gelles, Proceedings of the National Academy of Sciences 107, 5453–5458, issn: 0027-8424, eprint: https://www.pnas.org/content/107/12/5453.full.pdf , (https://www.pnas.org/content/107/12/5453) (2010)

work page 2010
[5]

Subramanian et al

R. Subramanian et al. , Cell 142, 433–443, issn: 0092-8674, ( http : / / www . sciencedirect . com / science/article/pii/S0092867410007816) (2010)

work page 2010
[6]

Maniatis, J

T. Maniatis, J. Sambrook, E. Fritsch, Cold Spring Harbor Laboratory Press , 623–623 (1989)

work page 1989
[7]

S. J. Streichan, M. F. Lefebvre, N. Noll, E. F. Wieschaus, B. I. Shraiman, eLife 7,e d .b yF .J ü l i c h e r , e27454, issn: 2050-084X, ( https://doi.org/10.7554/eLife.27454) (Feb. 2018). 8.J . S c h i n d e l i net al. , Nature Methods 9, 676 (2012). 9.S . P r e i b i s c het al. , Nature Methods 11, 645–648 (2014)

work page doi:10.7554/elife.27454 2012
[10]

Rezakhaniha et al

R. Rezakhaniha et al. , Biomechanics and Modeling in Mechanobiology 11, 461 (2012)

work page 2012

Show all 18 references

[11]

Marenduzzo, E

D. Marenduzzo, E. Orlandini, M. Cates, J. Yeomans, Physical Review E 76, 031921 (2007)

2007
[12]

J. Zhao, Q. Wang, Journal of Scientiﬁc Computing 68, 1241–1266 (2016)

2016
[13]

S. P. Vanka, Journal of Computational Physics 65, 138–158 (1986)

1986
[14]

Kleman, J

M. Kleman, J. Friedel, Reviews of Modern Physics 80, 61 (2008)

2008
[15]

G. P. Alexander, B. G.-g. Chen, E. A. Matsumoto, R. D. Kamien, Reviews of Modern Physics 84, 497 (2012)

2012
[16]

Jänich, Acta Applicandae Mathematica 8, 65–74 (1987)

K. Jänich, Acta Applicandae Mathematica 8, 65–74 (1987)

1987
[17]

S. opar, S. éumer, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Science 469, 20130204 (2013)

2013
[18]

Tran et al

L. Tran et al. , Proceedings of the National Academy of Sciences 113, 7106–7111 (2016). 11

2016
[19]

Lozano-Durán, G

A. Lozano-Durán, G. Borrell, ACM Transactions on Mathematical Software (TOMS) 42, 34 (2016)

2016
[20]

Toriwaki, T

J. Toriwaki, T. Yonekura, Forma 17, 183–209 (2002). 12

2002

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Castoldi, A

M. Castoldi, A. V. Popov, Protein Expression and Puriﬁcation 32, 83–88, issn: 1046-5928 (2003)

work page 2003

[2] [2]

Hyman et al

A. Hyman et al. ,i n Molecular Motors and the Cytoskeleton (Academic Press, 1991), vol. 196, pp. 478– 485

work page 1991

[3] [3]

S. J. DeCamp, G. S. Redner, A. Baskaran, M. F. Hagan, Z. Dogic, Nature Materials 14, 1110 (2015)

work page 2015

[4] [4]

D. S. Martin, R. Fathi, T. J. Mitchison, J. Gelles, Proceedings of the National Academy of Sciences 107, 5453–5458, issn: 0027-8424, eprint: https://www.pnas.org/content/107/12/5453.full.pdf , (https://www.pnas.org/content/107/12/5453) (2010)

work page 2010

[5] [5]

Subramanian et al

R. Subramanian et al. , Cell 142, 433–443, issn: 0092-8674, ( http : / / www . sciencedirect . com / science/article/pii/S0092867410007816) (2010)

work page 2010

[6] [6]

Maniatis, J

T. Maniatis, J. Sambrook, E. Fritsch, Cold Spring Harbor Laboratory Press , 623–623 (1989)

work page 1989

[7] [7]

S. J. Streichan, M. F. Lefebvre, N. Noll, E. F. Wieschaus, B. I. Shraiman, eLife 7,e d .b yF .J ü l i c h e r , e27454, issn: 2050-084X, ( https://doi.org/10.7554/eLife.27454) (Feb. 2018). 8.J . S c h i n d e l i net al. , Nature Methods 9, 676 (2012). 9.S . P r e i b i s c het al. , Nature Methods 11, 645–648 (2014)

work page doi:10.7554/elife.27454 2012

[8] [10]

Rezakhaniha et al

R. Rezakhaniha et al. , Biomechanics and Modeling in Mechanobiology 11, 461 (2012)

work page 2012