Global rate optimality of integral curve estimators

C. Banerjee   ; L. A. Sakhanenko   ; D. C. Zhu

Teoriâ veroâtnostej i ee primeneniâ, Tome 68 (2023) no. 2, pp. 301-321

Résumé

Motivated by an application in neuroimaging, we consider the problem of establishing global minimax lower bound in a high order tensor model. In particular, the methodology we describe provides the global minimax bound for the integral curve estimator proposed in [O. Carmichael and L. Sakhanenko, Linear Algebra Appl., 473 (2015), pp. 377–403] under a semiparametric estimation setting. The theoretical results in this paper guarantee that the estimator used for tracing the complex fiber structure inside a live human brain obtained from high angular resolution diffusion imaging (HARDI) data is not only optimal locally but also optimal globally. The global minimax bound on the asymptotic risk of the estimators thus will provide a quantification of uncertainty for the estimation method in the whole domain of the imaging field. In addition to theoretical results, we also provide a detailed simulation study in order to find the optimal number of gradient directions for the imaging protocols, which we further illustrate with a real data analysis of a live human brain scan to showcase the uncertainty quantification of the estimation method in [O. Carmichael and L. Sakhanenko, Linear Algebra Appl., 473 (2015), pp. 377–403]. Furthermore, based on the global minimax bound, we propose a method for comparing the relative accuracy of several commonly used neuroimaging protocols in diffusion tensor imaging (DTI).

Détail

Keywords: global minimax lower bound, semiparametric estimation, magnetic resonance high angular resolution diffusion imaging.

Bibliographie

[1] C. Banerjee, L. Sakhanenko, D. C. Zhu, “Lower bounds for accuracy of estimation in magnetic resonance high angular resolution diffusion imaging data”, J. Indian Soc. Probab. Statist., 21:1 (2020), 1–41 | DOI

[2] C. Banerjee, L. Sakhanenko, D. C. Zhu, “Global rate optimality of integral curve estimators in high order tensor models: Supplemental material”, Theory Probab. Appl. (to appear)

[3] O. Carmichael, L. Sakhanenko, “Estimation of integral curves from high angular resolution diffusion imaging (HARDI) data”, Linear Algebra Appl., 473 (2015), 377–403 | DOI | MR | Zbl

[4] O. Carmichael, L. Sakhanenko, “Integral curves from noisy diffusion MRI data with closed-form uncertainty estimates”, Stat. Inference Stoch. Process., 19:3 (2016), 289–319 | DOI | MR | Zbl

[5] Soo-Eun Chang, D. C. Zhu, “Neural network connectivity differences in children who stutter”, Brain, 136:12 (2013), 3709–3726 | DOI

[6] C. Chicone, Ordinary differential equations with applications, Texts Appl. Math., 34, 2nd ed., Springer, New York, 2006, xx+636 pp. | DOI | MR | Zbl

[7] L. de Lathauwer, B. de Moor, J. Vandewalle, “On the best rank-1 and rank-$(R_1,R_2,\dots,R_N)$ approximation of higher-order tensors”, SIAM J. Matrix Anal. Appl., 21:4 (2000), 1324–1342 | DOI | MR | Zbl

[8] L. Devroye, A course in density estimation, Progr. Probab. Statist., 14, Birkhäuser Boston, Inc., Boston, MA, 1987, xx+183 pp. | MR | Zbl

[9] T. S. Ferguson, Mathematical statistics. A decision theoretic approach, Probab. Math. Statist., 1, Academic Press, New York–London, 1967, xi+396 pp. | MR | Zbl

[10] A. Guntuboyina, B. Sen, “Global risk bounds and adaptation in univariate convex regression”, Probab. Theory Related Fields, 163:1-2 (2015), 379–411 | DOI | MR | Zbl

[11] R. Hasminskii, I. Ibragimov, “On density estimation in the view of Kolmogorov's ideas in approximation theory”, Ann. Statist., 18:3 (1990), 999–1010 | DOI | MR | Zbl

[12] I. A. Ibragimov, R. Z. Has'minskii, Statistical estimation. Asymptotic theory, Appl. Math., 16, Springer-Verlag, New York–Berlin, 1981, vii+403 pp. | DOI | MR | MR | Zbl | Zbl

[13] I. A. Ibragimov, R. Z. Khas'minskii, “Estimation of distribution density”, J. Soviet Math., 21:1 (1983), 40–57 | DOI | MR | Zbl

[14] I. A. Ibragimov, R. Z. Khas'minskii, “More on the estimation of distribution densities”, J. Soviet Math., 25:3 (1984), 1155–1165 | DOI | MR | Zbl

[15] A. K. H. Kim, R. J. Samworth, “Global rates of convergence in log-concave density estimation”, Ann. Statist., 44:6 (2016), 2756–2779 | DOI | MR | Zbl

[16] V. Koltchinskii, L. Sakhanenko, Songhe Cai, “Integral curves of noisy vector fields and statistical problems in diffusion tensor imaging: nonparametric kernel estimation and hypotheses testing”, Ann. Statist., 35:4 (2007), 1576–1607 | DOI | MR | Zbl

[17] D. Le Bihan, J. F. Mangin, C. Poupon, C. A. Clark, S. Pappata, N. Molko, H. Chabriat, “Diffusion tensor imaging: concepts and applications”, J. Magn. Reson. Imaging, 13:4 (2001), 534–546 | DOI

[18] E. Mammen, S. van de Geer, “Locally adaptive regression splines”, Ann. Statist., 25:1 (1997), 387–413 | DOI | MR | Zbl

[19] Guyan Ni, Yiju Wang, “On the best rank-1 approximation to higher-order symmetric tensors”, Math. Comput. Modelling, 46:9-10 (2007), 1345–1352 | DOI | MR | Zbl

[20] E. Özarslan, T. Mareci, “Generalized diffusion tensor imaging and analytical relationships between diffusion tensor imaging and high angular resolution diffusion imaging”, Magn. Reson. Med., 50:5 (2003), 955–965 | DOI

[21] G. Raskutti, M. J. Wainwright, Bin Yu, “Minimax-optimal rates for sparse additive models over kernel classes via convex programming”, J. Mach. Learn. Res., 13 (2012), 389–427 | MR | Zbl

[22] L. Sakhanenko, “A global optimal convergence rate in a model for the diffusion tensor imaging”, Theory Probab. Appl., 55:1 (2011), 77–90 | DOI | DOI | MR | Zbl

[23] L. Sakhanenko, “How to choose the number of gradient directions for estimation problems from noisy diffusion tensor data”, Contemporary developments in statistical theory, A festschrift for Hira Lal Koul, Springer Proc. Math. Stat., 68, Springer, Cham, 2014, 305–310 | DOI | MR | Zbl

[24] L. Sakhanenko, M. DeLaura, “A comparison study of statistical tractography methodologies for diffusion tensor imaging”, Int. J. Stat. Adv. Theory Appl., 1:1 (2017), 93–110

[25] C. J. Stone, “Optimal global rates of convergence for nonparametric regression”, Ann. Statist., 10:4 (1982), 1040–1053 | DOI | MR | Zbl

[26] M. J. Wainwright, High-dimensional statistics. A non-asymptotic viewpoint, Camb. Ser. Stat. Probab. Math., 48, Cambridge Univ. Press, Cambridge, 2019, xvii+552 pp. | DOI | MR | Zbl

[27] Leslie Ying, Yi Ming Zou, D. P. Klemer, Jiun-Jie Wang, “Determination of fiber orientation in MRI diffusion tensor imaging based on higher-order tensor decomposition”, Proceedings of the 29th annual international conference of the IEEE engineering in medicine and biology society (Lyon, 2007), IEEE, 2007, 2065–2068 | DOI

[28] D. C. Zhu, S. Majumdar, “Integration of resting-state FMRI and diffusion-weighted MRI connectivity analyses of the human brain: limitations and improvement”, J. Neuroimaging, 24:2 (2014), 176–186 | DOI

[29] D. C. Zhu, S. Majumdar, I. O. Korolev, K. L. Berger, A. C. Bozoki, “Alzheimer's disease and amnestic mild cognitive impairment weaken connections within the default mode network: a multi-modal imaging study”, J. Alzheimer's Dis., 34:4 (2013), 969–984 | DOI