Annals of Statistics

Tilted Euler characteristic densities for Central Limit random fields, with application to “bubbles”

N. Chamandy, K. J. Worsley, J. Taylor, and F. Gosselin

Full-text: Open access


Local increases in the mean of a random field are detected (conservatively) by thresholding a field of test statistics at a level u chosen to control the tail probability or p-value of its maximum. This p-value is approximated by the expected Euler characteristic (EC) of the excursion set of the test statistic field above u, denoted $\mathbb{E}\varphi(A_{u})$. Under isotropy, one can use the expansion $\mathbb{E}\varphi(A_{u})=\sum_{k}\mathcal{V}_{k}\rho_{k}(u)$, where $\mathcal{V}_{k}$ is an intrinsic volume of the parameter space and ρk is an EC density of the field. EC densities are available for a number of processes, mainly those constructed from (multivariate) Gaussian fields via smooth functions. Using saddlepoint methods, we derive an expansion for ρk(u) for fields which are only approximately Gaussian, but for which higher-order cumulants are available. We focus on linear combinations of n independent non-Gaussian fields, whence a Central Limit theorem is in force. The threshold u is allowed to grow with the sample size n, in which case our expression has a smaller relative asymptotic error than the Gaussian EC density. Several illustrative examples including an application to “bubbles” data accompany the theory.

Article information

Ann. Statist., Volume 36, Number 5 (2008), 2471-2507.

First available in Project Euclid: 13 October 2008

Permanent link to this document

Digital Object Identifier

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

Primary: 60G60: Random fields 62E20: Asymptotic distribution theory 62M40: Random fields; image analysis
Secondary: 53A99: None of the above, but in this section 58E05: Abstract critical point theory (Morse theory, Ljusternik-Schnirelman (Lyusternik-Shnirel m an) theory, etc.) 60B12: Limit theorems for vector-valued random variables (infinite- dimensional case) 60F05: Central limit and other weak theorems

Saddlepoint approximation isotropic field Poisson process threshold excursion set


Chamandy, N.; Worsley, K. J.; Taylor, J.; Gosselin, F. Tilted Euler characteristic densities for Central Limit random fields, with application to “bubbles”. Ann. Statist. 36 (2008), no. 5, 2471--2507. doi:10.1214/07-AOS549.

Export citation


  • [1] Adler, R. J. (1981). The Geometry of Random Fields. Wiley, New York.
  • [2] Adler, R. J. and Taylor, J. E. (2007). Random Fields and Geometry. Springer, New York.
  • [3] Adolphs, R., Gosselin, F., Buchanan, T. W., Tranel, D., Schyns, P. and Damasio, A. R. (2005). A mechanism for impaired fear recognition after amygdala damage. Nature 433 68–72.
  • [4] Barndorff-Nielsen, O. E. and Cox, D. R. (1989). Asymptotic Techniques for Use in Statistics. Chapman and Hall, London.
  • [5] Bhattacharya, R. N. and Ranga Rao, R. (1976). Normal Approximation and Asymptotic Expansions. Wiley, New York.
  • [6] Cao, J. and Worsley, K. J. (1999). The detection of local shape changes via the geometry of Hotelling’s t2 fields. Ann. Statist. 27 925–942.
  • [7] Carbonell, F., Galan, L. and Worsley, K. J. (2005). The geometry of the Wilks’s Lambda random field. Submitted.
  • [8] Catelan, P., Lucchin, F. and Matarrese, S. (1988). Peak number density of non-Gaussian random-fields. Phys. Rev. Lett. 61 267–270.
  • [9] Chamandy, N. (2007). Inference for asymptotically Gaussian random fields. Ph.D. thesis, McGill University, Montréal.
  • [10] Chamandy, N. (2007). Thresholding asymptotically Gaussian random fields using the tilted mean EC, with application to lesion density maps. In Proceedings of the 56th Session of the International Statistical Institute.
  • [11] Chamandy, N., Taylor, J. E. and Worsley, K. J. (2008). Euler characteristic densities for asymptotically Gaussian random fields: A geometric approach. In preparation.
  • [12] Daniels, H. E. (1954). Saddlepoint approximations in statistics. Ann. Math. Statist. 25 631–650.
  • [13] Daniels, H. E. (1980). Exact saddlepoint approximations. Biometrika 67 59–63.
  • [14] Daniels, H. E. (1987). Tail probability approximations. Internat. Statist. Rev. 55 37–48.
  • [15] Field, C. and Ronchetti, E. (1990). Small Sample Asymptotics. IMS, Hayward, CA.
  • [16] Gosselin, F. and Schyns, P. G. (2001). Bubbles, a technique to reveal the use of information in recognition. Vision Research 41 2261–2271.
  • [17] Hadwiger, H. (1957). Vorlesüngen Über Inhalt, Oberfläche und Isoperimetrie. Springer, Berlin.
  • [18] Hoyle, F., Vogeley, M. S., Gott, J. R., Blanton, M., Tegmark, M., Weinberg, D., Bachall, N., Brinchman, J. and York, D. (2002). Two-dimensional topology of the SDSS. Astrophys. J. 580 663–671.
  • [19] Loader, C. R. and Sun, J. (1997). Robustness of tube formulae. J. Comput. Graph. Statist. 6 242–250.
  • [20] Lugannani, R. and Rice, S. O. (1980). Saddle point approximations for the distribution of the sum of independent random variables. Adv. in Appl. Probab. 12 475–490.
  • [21] Matsubara, T. (1994). Analytic expression of the genus in weakly non-Gaussian field induced by gravity. Astrophys. J. 434 L43–L46.
  • [22] Matsubara, T. and Yokoyama, J. (1996). Genus statistics of the large-scale structure with non-Gaussian density fields. Astrophys. J. 463 409–419.
  • [23] Nardi, Y., Siegmund, D. and Yakir, B. (2008). The distribution of maxima of approximately Gaussian random fields. Ann. Statist. To appear.
  • [24] Padmanabhan, N., Schlegel, D. J., Seljak, U., Makarov, A. et. al. (2006). The clustering of luminous red galaxies in the sloan digital sky survey imaging data. Mon. Not. R. Astron. Soc. Submitted. Available at–ph/0605302.
  • [25] Rabinowitz, D. and Siegmund, D. (1997). The approximate distribution of the maximum of a smoothed Poisson random field. Statist. Sinica 7 167–180.
  • [26] Robinson, J. (1982). Saddlepoint approximations for permutation tests and confidence intervals. J. Roy. Statist. Soc. Ser. B 44 91–101.
  • [27] Salmond, C. H., Ashburner, J., Vargha-Khadem, F., Connelly, A., Gadian, D. G. and Friston, K. J. (2002). Distributional assumptions in Voxel-based morphometry. Neuroimage 17 1027–1030.
  • [28] Schneider, R. (1993). Convex Bodies: The Brunn–Minkowski Theory. Cambridge Univ. Press.
  • [29] Sun, J., Loader, C. R. and McCormick, W. P. (2000). Confidence bands in generalized linear models. Ann. Statist. 28 429–460.
  • [30] Taylor, J. E. (2003). Gaussian volumes of tubes, Euler characteristic densities and correlated conjunction. In Proceedings of Singular Models and Geometric Methods in Statistics 7–33. IMS, Hayward, CA.
  • [31] Taylor, J. E. (2006). A Gaussian kinematic formula. Ann. Probab. 34 122–158.
  • [32] Taylor, J. E. and Adler, R. J. (2003). Euler characteristics for Gaussian fields on manifolds. Ann. Probab. 31 533–563.
  • [33] Taylor, J. E., Takemura, A. and Adler, R. J. (2005). Validity of the expected Euler characteristic heuristic. Ann. Probab. 33 1362–1396.
  • [34] Taylor, J. E. and Worsley, K. J. (2008). Detecting sparse signal in random fields, with an application to brain mapping. J. Amer. Statist. Assoc. 102 913–928.
  • [35] Taylor, J. E. and Worsley, K. J. (2008). Random fields of multivariate test statistics, with applications to shape analysis and fMRI. Ann. Statist. 36 1–27.
  • [36] Taylor, J. E., Worsley, K. J. and Gosselin, F. Maxima of discretely sampled random fields, with an application to “bubbles.” Biometrika 94 1–18.
  • [37] Viviani, R., Beschoner, P., Ehrhard, K., Schmitz, B. and Thöne, J. (2007). Non-normality and transformations of random fields, with an application to voxel-based morphometry. Neuroimage 35 121–130.
  • [38] Worsley, K. J. (1994). Local maxima and the expected Euler characteristic of excursion sets of χ2, F and t fields. Adv. in Appl. Probab. 26 13–42.
  • [39] Worsley, K. J. (1995). Boundary corrections for the expected Euler characteristic of excursion sets of random fields, with an application to astrophysics. Adv. in Appl. Probab. 27 943–959.
  • [40] Worsley, K. J. (1996). The geometry of random images. Chance 9 27–40.