Sulcal Pattern Matching with the Wasserstein Distance

Moo K. Chung; Soumya Das; Zijian Chen

arxiv: 2307.00385 · v2 · submitted 2023-07-01 · 🧬 q-bio.NC · eess.IV

Sulcal Pattern Matching with the Wasserstein Distance

Zijian Chen , Soumya Das , Moo K. Chung This is my paper

Pith reviewed 2026-05-24 07:57 UTC · model grok-4.3

classification 🧬 q-bio.NC eess.IV

keywords sulcal patternsWasserstein distancebrain MRIpattern matchingdeformation fieldsex differencesoptimal transport

0 comments

The pith

Wasserstein distance aligns sulcal patterns from brain MRI despite topological differences across subjects.

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

The paper develops a computational framework that represents sulcal patterns extracted from MRI as measures and uses the Wasserstein distance to compute nonlinear alignments between them. These patterns differ topologically from one brain to the next, so direct comparison requires a deformation field obtained through optimal transport. The authors derive the necessary mathematics and implement gradient descent to estimate that field, then measure how well the resulting registrations perform. The same procedure is used to test for systematic differences between male and female sulcal patterns. A reader would care because reliable alignment of variable brain folds could support studies of structural variation linked to sex or other traits.

Core claim

The central claim is that the Wasserstein distance supplies a practical way to align topologically different sulcal patterns nonlinearly; the mathematical details are worked out, gradient-descent algorithms are given for recovering the deformation field, registration quality is quantified, and the method is shown to identify differences between male and female sulcal patterns.

What carries the argument

The Wasserstein distance between sulcal patterns treated as measures or point sets, used to recover a deformation field by gradient descent.

If this is right

Nonlinear registration of sulcal patterns becomes possible even when their topologies differ across subjects.
Registration performance can be quantified directly from the Wasserstein distance itself.
The same pipeline can be applied to detect systematic differences between male and female sulcal patterns.

Where Pith is reading between the lines

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

The framework could be tested on longitudinal MRI data to track how sulcal patterns change within the same individual over time.
If the point-set representation holds, the method might extend to other cortical landmarks beyond sulci.
Differences detected between sexes could be checked against independent anatomical measures to see whether they reflect folding geometry or simply registration artifacts.

Load-bearing premise

Sulcal patterns extracted from MRI can be represented as measures whose topological differences are adequately captured by a deformation field obtained via Wasserstein optimal transport.

What would settle it

Running the Wasserstein alignment on a set of MRI-derived sulcal patterns and finding that the resulting deformation fields produce no measurable reduction in misalignment compared with rigid or affine registration would falsify the central claim.

read the original abstract

We present the unified computational framework for modeling the sulcal patterns of human brain obtained from the magnetic resonance images. The Wasserstein distance is used to align the sulcal patterns nonlinearly. These patterns are topologically different across subjects making the pattern matching a challenge. We work out the mathematical details and develop the gradient descent algorithms for estimating the deformation field. We further quantify the image registration performance. This method is applied in identifying the differences between male and female sulcal patterns.

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.

Desk Editor's Note private letter to a colleague

Wasserstein OT applied to sulcal patterns is a direct extension of existing transport methods with limited visible validation or topology handling.

read the letter

The paper sets up a framework that treats sulcal patterns as measures and uses the Wasserstein distance plus gradient descent to compute a nonlinear deformation field for alignment across subjects. It then applies the result to compare male and female sulcal patterns. That is the core of what is presented. The work does supply a concrete computational pipeline for this specific registration task and states that it quantifies registration performance, which is a step beyond pure theory. The math details and algorithm development are claimed to be worked out, so the authors have at least produced usable code-level steps even if they are not shown here. The application to sex differences gives it a downstream use case that some neuroimaging groups might actually run. The soft spots are straightforward. The abstract gives no equations, no error metrics, no baseline comparisons, and no description of how topological differences (branching, connectivity) are preserved or penalized under standard OT. The stress-test point holds: nothing in the visible text adds explicit regularization for sulcal geometry, so the claim that the deformation field adequately aligns topologically varying patterns rests on an untested assumption. Without those checks the registration numbers cannot be evaluated for bias or failure modes. This is the kind of paper that belongs in a specialized neuroimaging methods journal or conference track. Readers working on brain surface registration or optimal transport in medical imaging could extract the algorithm and test it themselves. It is not ready for broad citation yet because the evidence is still at the level of a methods sketch. A serious editor should send it to peer review so referees can see the full derivations, any synthetic or real-data experiments, and whether the topology issue is actually addressed in the implementation.

Referee Report

1 major / 2 minor

Summary. The manuscript presents a computational framework for modeling sulcal patterns extracted from MRI using the Wasserstein distance to perform nonlinear alignment of topologically varying patterns across subjects. It derives gradient-descent algorithms to estimate the deformation field, quantifies registration performance, and applies the method to detect differences between male and female sulcal patterns.

Significance. A validated method for aligning sulcal patterns despite topological mismatches would provide a principled, measure-theoretic approach to brain morphometry with direct relevance to studies of sex differences and inter-subject variability. The explicit construction of gradient-descent OT updates is a positive technical contribution if the resulting alignments preserve or correctly handle sulcal topology.

major comments (1)

[Abstract] Abstract, paragraph 2: the central claim that a Wasserstein-derived deformation field adequately aligns sulcal patterns despite topological differences across subjects is load-bearing, yet the description provides no explicit regularization, connectivity constraint, or unbalanced-transport penalty to preserve branching or genus; standard balanced OT minimizes transport cost between supports but contains no built-in mechanism for these topological features.

minor comments (2)

The abstract states that registration performance is quantified but supplies no metrics, error measures, or comparison baselines in the provided text.
Notation for the sulcal patterns as measures or point sets is not introduced in the visible description, making it difficult to assess how the support geometry is discretized.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the detailed review and the constructive comment on the abstract. We address the point below and will revise the manuscript accordingly to improve clarity.

read point-by-point responses

Referee: [Abstract] Abstract, paragraph 2: the central claim that a Wasserstein-derived deformation field adequately aligns sulcal patterns despite topological differences across subjects is load-bearing, yet the description provides no explicit regularization, connectivity constraint, or unbalanced-transport penalty to preserve branching or genus; standard balanced OT minimizes transport cost between supports but contains no built-in mechanism for these topological features.

Authors: We agree that the abstract is brief and does not detail the handling of topological variations. The full manuscript derives a gradient-descent procedure on the deformation field that maps one sulcal measure to another; because the supports are allowed to differ in cardinality and local connectivity (as sulcal patterns vary across subjects), the optimal transport plan implicitly accommodates unmatched branches without requiring a bijection. However, we acknowledge that the abstract should clarify this point and will expand it to note that the method operates on probability measures without explicit topology-preserving penalties, relying instead on the flexibility of the Wasserstein metric between non-isomorphic supports. We will also add a sentence referencing the unbalanced-transport literature if future extensions are considered. revision: yes

Circularity Check

0 steps flagged

No circularity detected; derivation chain not visible or self-referential

full rationale

The provided abstract and context describe a framework applying Wasserstein distance and gradient descent to align sulcal patterns represented as measures, with no equations, fitted parameters, self-citations, or derivation steps shown. No load-bearing step reduces to an input by construction, no uniqueness theorem is invoked, and no renaming or ansatz smuggling occurs. The central claim relies on standard OT properties applied to the problem rather than any internal redefinition or statistical forcing, making the derivation self-contained against external benchmarks where visible.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; the method implicitly assumes sulcal patterns admit a measure-theoretic representation suitable for Wasserstein distance and that a deformation field can resolve topological mismatches. No free parameters, axioms, or invented entities are stated in the provided text.

pith-pipeline@v0.9.0 · 5599 in / 950 out tokens · 27257 ms · 2026-05-24T07:57:39.612679+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages · 1 internal anchor

[1]

INTRODUCTION The concave regions in the highly convoluted cerebral cor- tex of the human brain are referred to as the sulci (Fig. 1). These complex tree-shaped sulcal curves are highly variable in length, area, depth, curvature and topology across different subjects [1]. There have been extensive studies that connect the variabilities of such biomarkers w...

work page
[2]

Sulcal Pattern Matching with the Wasserstein Distance

METHODS 2.1. Sulcal pattern data We used the processed T1-weighted MRI of 456 subjects (age- matched 274 females and 182 males) in the Human Con- nectome Project (HCP) [8]. The MRI were obtained using a Siemens 3T Connectome Skyra scanner with a 32-channel head coil [9, 10]. The MRI were registered to the MNI space with a FLIRT affine and FNIRT nonlinear ...

work page internal anchor Pith review Pith/arXiv arXiv 2023
[3]

Validation against the Hungarian Algorithm We matched the two sets ofn random generated scatter points in [0, 1]2

EXPERIMENTS 3.1. Validation against the Hungarian Algorithm We matched the two sets ofn random generated scatter points in [0, 1]2. Fig.3 displays the result of one realization withn =

work page
[4]

We then applied heat kernel kernel smoothing with bandwidth σ = 0 .01 on the scatter points and compute the Wasserstein distance D′ W ( ˜f1, ˜f2) using the gradient descent (9)

The Wasserstein distanceDW (f1, f2) can be computed ex- actly through (3) using the Hungarian Algorithm [19]. We then applied heat kernel kernel smoothing with bandwidth σ = 0 .01 on the scatter points and compute the Wasserstein distance D′ W ( ˜f1, ˜f2) using the gradient descent (9). The av- erage percentage reduction of distance for 100 independent si...

work page
[5]

We pro- vided the theoretical justification for performing heat kernel smoothing before computing the Wasserstein distance

DISCUSSION In this paper, we presented the new framework of match- ing sulcal patterns of human brain across subjects. We pro- vided the theoretical justification for performing heat kernel smoothing before computing the Wasserstein distance. Smooth- ing reduces the Wasserstein distance between the sulcal pat- terns and spatial pattern variabilities. It i...

work page
[6]

ACKNOWLEDGEMENT We would like to thank Soheil Kolouri of Vanderbilt Univer- sity and Tahmineh Azizi of University of Wisconsin-Madison for discussion on the Wasserstein distance, and Ilwoo Lyu of Ulsan National Institute of Science and Technology for assis- tant with the TRACE algorithm

work page
[7]

A primal sketch of the cortex mean cur- vature: A morphogenesis based approach to study the variability of the folding patterns,

A. Cachia, J.F. Mangin, D. Riviere, F. Kherif, N. Bod- daert, A. Andrade, D. Papadopoulos-Orfanos, J.B. Po- line, I. Bloch, M. Zilbovicius, P. Sonigo, F. Brunelle, and J. R ´egis, “A primal sketch of the cortex mean cur- vature: A morphogenesis based approach to study the variability of the folding patterns,” IEEE Trans. Med. Imag., vol. 22, no. 6, pp. 75...

work page 2003
[8]

Sulcal pits and patterns in devel- oping human brains,

K. Im and P.E. Grant, “Sulcal pits and patterns in devel- oping human brains,” NeuroImage, vol. 185, pp. 881– 890, 2019

work page 2019
[9]

Fast polynomial approximation of heat kernel convolution on manifolds and its application to brain sulcal and gyral graph pattern analysis,

S.-G. Huang, I. Lyu, A. Qiu, and M.K. Chung, “Fast polynomial approximation of heat kernel convolution on manifolds and its application to brain sulcal and gyral graph pattern analysis,” IEEE Transactions on Medical Imaging, vol. 39, pp. 2201–2212, 2020

work page 2020
[10]

Shape analysis with hyperbolic wasserstein distance,

J. Shi, W. Zhang, and Y . Wang, “Shape analysis with hyperbolic wasserstein distance,” in 2016 IEEE Con- ference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 5051–5061

work page 2016
[11]

Sliced wasserstein distance for learning gaussian mixture mod- els,

S. Kolouri, G.K. Rohde, and H. Hoffmann, “Sliced wasserstein distance for learning gaussian mixture mod- els,” in2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 3427–3436

work page 2018
[12]

Un- paired deep learning for accelerated mri using optimal transport driven cyclegan,

G. Oh, B. Sim, H. Chung, L. Sunwoo, and J.C. Ye, “Un- paired deep learning for accelerated mri using optimal transport driven cyclegan,”IEEE Transactions on Com- putational Imaging, vol. 6, pp. 1285–1296, 2020

work page 2020
[13]

A gradient descent solution to the monge-kantorovich problem,

R. Chartrand, B. Wohlberg, K. Vixie, and E. Bollt, “A gradient descent solution to the monge-kantorovich problem,” Applied Mathematical Sciences , vol. 3, pp. 1071–1080, 01 2009

work page 2009
[14]

The human connectome project: a data acquisition perspective,

D.C. Van Essen, K. Ugurbil, E. Auerbach, D. Barch, T.E.J. Behrens, R. Bucholz, A. Chang, L. Chen, M. Cor- betta, and S.W. Curtiss, “The human connectome project: a data acquisition perspective,” NeuroImage, vol. 62, pp. 2222–2231, 2012

work page 2012
[15]

The minimal preprocessing pipelines for the Human Connectome Project,

M.F. Glasser, S.N. Sotiropoulos, J.A. Wilson, T.S. Coal- son, B. Fischl, J.L. Andersson, J. Xu, S. Jbabdi, M. Web- ster, J.R. Polimeni, D.C. Van Essen, and M. Jenkinson, “The minimal preprocessing pipelines for the Human Connectome Project,” NeuroImage, vol. 80, pp. 105– 124, 2013

work page 2013
[16]

Resting-state fMRI in the Human Connectome Project,

S.M. Smith, C.F. Beckmann, J. Andersson, E.J. Auer- bach, J. Bijsterbosch, E. Douaud, G.and Duff, D.A. Feinberg, L. Griffanti, M.P. Harms, et al., “Resting-state fMRI in the Human Connectome Project,”NeuroImage, vol. 80, pp. 144–168, 2013

work page 2013
[17]

Improved optimization for the robust and accurate lin- ear registration and motion correction of brain images,

M. Jenkinson, P. Bannister, and S. Brady, M.and Smith, “Improved optimization for the robust and accurate lin- ear registration and motion correction of brain images,” NeuroImage, vol. 17, no. 2, pp. 825–841, 2002

work page 2002
[18]

Cortical folding patterns and predicting cy- toarchitecture,

B. Fischl, N. Rajendran, E. Busa, J. Augustinack, O. Hinds, B.T Yeo, H. Mohlberg, K. Amunts, and K. Zilles, “Cortical folding patterns and predicting cy- toarchitecture,” Cerebral Cortex, vol. 18, no. 8, pp. 1973–1980, 2007

work page 1973
[19]

TRACE: A topological graph repre- sentation for automatic sulcal curve extraction,

I. Lyu, S.H. Kim, N.D. Woodward, M.A. Styner, and B.A. Landman, “TRACE: A topological graph repre- sentation for automatic sulcal curve extraction,” IEEE Trans. Med. Imag., vol. 37, no. 7, pp. 1653–1663, 2018

work page 2018
[20]

Weighted fourier series representation and its application to quantifying the amount of gray matter,

M.K. Chung, K.M. Dalton, L. Shen, A.C. Evans, and R.J. Davidson, “Weighted fourier series representation and its application to quantifying the amount of gray matter,” IEEE Transactions on Medical Imaging , vol. 26, no. 4, pp. 566–581, 2007

work page 2007
[21]

Computational optimal trans- port,

G. Peyr ´e and M. Cuturi, “Computational optimal trans- port,” Foundations and Trends in Machine Learning , vol. 11, no. 5-6, pp. 355–607, 2019

work page 2019
[22]

A class of wasserstein metrics for probability distributions.,

C.R. Givens and R.M. Shortt, “A class of wasserstein metrics for probability distributions.,”Michigan Mathe- matical Journal, vol. 31, no. 2, pp. 231–240, 1984

work page 1984
[23]

On the optimal mapping of distributions,

M. Knott and C. S. Smith, “On the optimal mapping of distributions,” Journal of Optimization Theory and Applications, vol. 43, no. 1, pp. 39–49, 1984

work page 1984
[24]

Optimal mass transport: Signal processing and machine-learning applications,

S. Kolouri, S.R. Park, M. Thorpe, D. Slepcev, and G.K. Rohde, “Optimal mass transport: Signal processing and machine-learning applications,” IEEE Signal Process- ing Magazine, vol. 34, no. 4, pp. 43–59, 2017

work page 2017
[25]

Algorithms for the assignment and trans- portation problems,

J. Munkres, “Algorithms for the assignment and trans- portation problems,” Journal of the Society for Indus- trial and Applied Mathematics, vol. 5, no. 1, pp. 32–38, 1957

work page 1957
[26]

A computational fluid me- chanics solution to the monge-kantorovich mass transfer problem,

J. Benamou and Y . Brenier, “A computational fluid me- chanics solution to the monge-kantorovich mass transfer problem,” Numerische Mathematik, vol. 84, no. 3, pp. 375–393, 2000

work page 2000
[27]

Rapid acceleration of the permutation test via transpositions,

M.K. Chung, L. Xie, S.-G. Huang, Y . Wang, J. Yan, and L. Shen, “Rapid acceleration of the permutation test via transpositions,” International Workshop on Connec- tomics in Neuroimaging, vol. 11848, pp. 42–53, 2019

work page 2019
[28]

Why sex matters: brain size independent differences in gray matter distributions between men and women,

E. Luders, C. Gaser, K.L. Narr, and A.W. Toga, “Why sex matters: brain size independent differences in gray matter distributions between men and women,” Journal of Neuroscience, vol. 29, pp. 14265–14270, 2009

work page 2009
[29]

Gender difference analysis of cortical thickness in healthy young adults with surface- based methods,

K. Im, J.-M. Lee, J. Lee, Y .-W. Shin, I. Y . Kim, J. S. Kwon, and S.I. Kim, “Gender difference analysis of cortical thickness in healthy young adults with surface- based methods,” NeuroImage, vol. 31, pp. 31–38, 2006

work page 2006

[1] [1]

INTRODUCTION The concave regions in the highly convoluted cerebral cor- tex of the human brain are referred to as the sulci (Fig. 1). These complex tree-shaped sulcal curves are highly variable in length, area, depth, curvature and topology across different subjects [1]. There have been extensive studies that connect the variabilities of such biomarkers w...

work page

[2] [2]

Sulcal Pattern Matching with the Wasserstein Distance

METHODS 2.1. Sulcal pattern data We used the processed T1-weighted MRI of 456 subjects (age- matched 274 females and 182 males) in the Human Con- nectome Project (HCP) [8]. The MRI were obtained using a Siemens 3T Connectome Skyra scanner with a 32-channel head coil [9, 10]. The MRI were registered to the MNI space with a FLIRT affine and FNIRT nonlinear ...

work page internal anchor Pith review Pith/arXiv arXiv 2023

[3] [3]

Validation against the Hungarian Algorithm We matched the two sets ofn random generated scatter points in [0, 1]2

EXPERIMENTS 3.1. Validation against the Hungarian Algorithm We matched the two sets ofn random generated scatter points in [0, 1]2. Fig.3 displays the result of one realization withn =

work page

[4] [4]

We then applied heat kernel kernel smoothing with bandwidth σ = 0 .01 on the scatter points and compute the Wasserstein distance D′ W ( ˜f1, ˜f2) using the gradient descent (9)

The Wasserstein distanceDW (f1, f2) can be computed ex- actly through (3) using the Hungarian Algorithm [19]. We then applied heat kernel kernel smoothing with bandwidth σ = 0 .01 on the scatter points and compute the Wasserstein distance D′ W ( ˜f1, ˜f2) using the gradient descent (9). The av- erage percentage reduction of distance for 100 independent si...

work page

[5] [5]

We pro- vided the theoretical justification for performing heat kernel smoothing before computing the Wasserstein distance

DISCUSSION In this paper, we presented the new framework of match- ing sulcal patterns of human brain across subjects. We pro- vided the theoretical justification for performing heat kernel smoothing before computing the Wasserstein distance. Smooth- ing reduces the Wasserstein distance between the sulcal pat- terns and spatial pattern variabilities. It i...

work page

[6] [6]

ACKNOWLEDGEMENT We would like to thank Soheil Kolouri of Vanderbilt Univer- sity and Tahmineh Azizi of University of Wisconsin-Madison for discussion on the Wasserstein distance, and Ilwoo Lyu of Ulsan National Institute of Science and Technology for assis- tant with the TRACE algorithm

work page

[7] [7]

A primal sketch of the cortex mean cur- vature: A morphogenesis based approach to study the variability of the folding patterns,

A. Cachia, J.F. Mangin, D. Riviere, F. Kherif, N. Bod- daert, A. Andrade, D. Papadopoulos-Orfanos, J.B. Po- line, I. Bloch, M. Zilbovicius, P. Sonigo, F. Brunelle, and J. R ´egis, “A primal sketch of the cortex mean cur- vature: A morphogenesis based approach to study the variability of the folding patterns,” IEEE Trans. Med. Imag., vol. 22, no. 6, pp. 75...

work page 2003

[8] [8]

Sulcal pits and patterns in devel- oping human brains,

K. Im and P.E. Grant, “Sulcal pits and patterns in devel- oping human brains,” NeuroImage, vol. 185, pp. 881– 890, 2019

work page 2019

[9] [9]

Fast polynomial approximation of heat kernel convolution on manifolds and its application to brain sulcal and gyral graph pattern analysis,

S.-G. Huang, I. Lyu, A. Qiu, and M.K. Chung, “Fast polynomial approximation of heat kernel convolution on manifolds and its application to brain sulcal and gyral graph pattern analysis,” IEEE Transactions on Medical Imaging, vol. 39, pp. 2201–2212, 2020

work page 2020

[10] [10]

Shape analysis with hyperbolic wasserstein distance,

J. Shi, W. Zhang, and Y . Wang, “Shape analysis with hyperbolic wasserstein distance,” in 2016 IEEE Con- ference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 5051–5061

work page 2016

[11] [11]

Sliced wasserstein distance for learning gaussian mixture mod- els,

S. Kolouri, G.K. Rohde, and H. Hoffmann, “Sliced wasserstein distance for learning gaussian mixture mod- els,” in2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 3427–3436

work page 2018

[12] [12]

Un- paired deep learning for accelerated mri using optimal transport driven cyclegan,

G. Oh, B. Sim, H. Chung, L. Sunwoo, and J.C. Ye, “Un- paired deep learning for accelerated mri using optimal transport driven cyclegan,”IEEE Transactions on Com- putational Imaging, vol. 6, pp. 1285–1296, 2020

work page 2020

[13] [13]

A gradient descent solution to the monge-kantorovich problem,

R. Chartrand, B. Wohlberg, K. Vixie, and E. Bollt, “A gradient descent solution to the monge-kantorovich problem,” Applied Mathematical Sciences , vol. 3, pp. 1071–1080, 01 2009

work page 2009

[14] [14]

The human connectome project: a data acquisition perspective,

D.C. Van Essen, K. Ugurbil, E. Auerbach, D. Barch, T.E.J. Behrens, R. Bucholz, A. Chang, L. Chen, M. Cor- betta, and S.W. Curtiss, “The human connectome project: a data acquisition perspective,” NeuroImage, vol. 62, pp. 2222–2231, 2012

work page 2012

[15] [15]

The minimal preprocessing pipelines for the Human Connectome Project,

M.F. Glasser, S.N. Sotiropoulos, J.A. Wilson, T.S. Coal- son, B. Fischl, J.L. Andersson, J. Xu, S. Jbabdi, M. Web- ster, J.R. Polimeni, D.C. Van Essen, and M. Jenkinson, “The minimal preprocessing pipelines for the Human Connectome Project,” NeuroImage, vol. 80, pp. 105– 124, 2013

work page 2013

[16] [16]

Resting-state fMRI in the Human Connectome Project,

S.M. Smith, C.F. Beckmann, J. Andersson, E.J. Auer- bach, J. Bijsterbosch, E. Douaud, G.and Duff, D.A. Feinberg, L. Griffanti, M.P. Harms, et al., “Resting-state fMRI in the Human Connectome Project,”NeuroImage, vol. 80, pp. 144–168, 2013

work page 2013

[17] [17]

Improved optimization for the robust and accurate lin- ear registration and motion correction of brain images,

M. Jenkinson, P. Bannister, and S. Brady, M.and Smith, “Improved optimization for the robust and accurate lin- ear registration and motion correction of brain images,” NeuroImage, vol. 17, no. 2, pp. 825–841, 2002

work page 2002

[18] [18]

Cortical folding patterns and predicting cy- toarchitecture,

B. Fischl, N. Rajendran, E. Busa, J. Augustinack, O. Hinds, B.T Yeo, H. Mohlberg, K. Amunts, and K. Zilles, “Cortical folding patterns and predicting cy- toarchitecture,” Cerebral Cortex, vol. 18, no. 8, pp. 1973–1980, 2007

work page 1973

[19] [19]

TRACE: A topological graph repre- sentation for automatic sulcal curve extraction,

I. Lyu, S.H. Kim, N.D. Woodward, M.A. Styner, and B.A. Landman, “TRACE: A topological graph repre- sentation for automatic sulcal curve extraction,” IEEE Trans. Med. Imag., vol. 37, no. 7, pp. 1653–1663, 2018

work page 2018

[20] [20]

Weighted fourier series representation and its application to quantifying the amount of gray matter,

M.K. Chung, K.M. Dalton, L. Shen, A.C. Evans, and R.J. Davidson, “Weighted fourier series representation and its application to quantifying the amount of gray matter,” IEEE Transactions on Medical Imaging , vol. 26, no. 4, pp. 566–581, 2007

work page 2007

[21] [21]

Computational optimal trans- port,

G. Peyr ´e and M. Cuturi, “Computational optimal trans- port,” Foundations and Trends in Machine Learning , vol. 11, no. 5-6, pp. 355–607, 2019

work page 2019

[22] [22]

A class of wasserstein metrics for probability distributions.,

C.R. Givens and R.M. Shortt, “A class of wasserstein metrics for probability distributions.,”Michigan Mathe- matical Journal, vol. 31, no. 2, pp. 231–240, 1984

work page 1984

[23] [23]

On the optimal mapping of distributions,

M. Knott and C. S. Smith, “On the optimal mapping of distributions,” Journal of Optimization Theory and Applications, vol. 43, no. 1, pp. 39–49, 1984

work page 1984

[24] [24]

Optimal mass transport: Signal processing and machine-learning applications,

S. Kolouri, S.R. Park, M. Thorpe, D. Slepcev, and G.K. Rohde, “Optimal mass transport: Signal processing and machine-learning applications,” IEEE Signal Process- ing Magazine, vol. 34, no. 4, pp. 43–59, 2017

work page 2017

[25] [25]

Algorithms for the assignment and trans- portation problems,

J. Munkres, “Algorithms for the assignment and trans- portation problems,” Journal of the Society for Indus- trial and Applied Mathematics, vol. 5, no. 1, pp. 32–38, 1957

work page 1957

[26] [26]

A computational fluid me- chanics solution to the monge-kantorovich mass transfer problem,

J. Benamou and Y . Brenier, “A computational fluid me- chanics solution to the monge-kantorovich mass transfer problem,” Numerische Mathematik, vol. 84, no. 3, pp. 375–393, 2000

work page 2000

[27] [27]

Rapid acceleration of the permutation test via transpositions,

M.K. Chung, L. Xie, S.-G. Huang, Y . Wang, J. Yan, and L. Shen, “Rapid acceleration of the permutation test via transpositions,” International Workshop on Connec- tomics in Neuroimaging, vol. 11848, pp. 42–53, 2019

work page 2019

[28] [28]

Why sex matters: brain size independent differences in gray matter distributions between men and women,

E. Luders, C. Gaser, K.L. Narr, and A.W. Toga, “Why sex matters: brain size independent differences in gray matter distributions between men and women,” Journal of Neuroscience, vol. 29, pp. 14265–14270, 2009

work page 2009

[29] [29]

Gender difference analysis of cortical thickness in healthy young adults with surface- based methods,

K. Im, J.-M. Lee, J. Lee, Y .-W. Shin, I. Y . Kim, J. S. Kwon, and S.I. Kim, “Gender difference analysis of cortical thickness in healthy young adults with surface- based methods,” NeuroImage, vol. 31, pp. 31–38, 2006

work page 2006