Small ball probability estimates, \(\psi_2\)-behavior and the hyperplane conjecture (Q2267510)

A convex body \(K\) in \({\mathbb R}^n\) is called isotropic if it has volume \(|K|=1\), center of mass at the origin, and its inertia matrix is a multiple of the identity; equivalently, if there is a constant \(L_K>0\) such that \(\int_K\langle x, \theta\rangle^2\, dx=L_K^2\) for every \(\theta\) in the Euclidean unit sphere \({\mathbf S}^{n-1}\). It is not hard to see that for every convex body \(K\) in \({\mathbb R}^n\) there exists an affine transformation \(T\) of \({\mathbb R}^n\) such that \(T(K)\) is isotropic. Moreover, this isotropic image is unique up to orthogonal transformations. Consequently, one can define the isotropic constant \(L_K\) as an invariant of the affine class of \(K\). The isotropic constant is closely related to the hyperplane conjecture (also known as the slicing problem) which asks if there exists an absolute constant \(c>0\) such that \(\max_{\theta\in {\mathbf S}^{n-1}}|K\cap \theta^\bot|\geq c\) for every convex body \(K\) of volume 1 in \({\mathbb R}^n\) and any \(\theta\in {\mathbf S}^{n-1}\), the function \(t\mapsto |K\cap(\theta^\bot + t\,\theta)|^{1/(n-1)}\) is concave on its support, and this implies that \(\int_K\langle x, \theta\rangle^2\, dx\simeq |K\cap\theta^\bot|^{-2}\). Using this relation it is not difficult to check that an affirmative answer to the slicing problem is equivalent to the following statement: There exists an absolute constant \(C>0\) such that \(L_K\leq C\) for every convex body \(K\). The isotropic constant and the hyperplane conjecture can be studied in the more general setting of log-concave measures. Let \(f:\,{\mathbb R}^n\to {\mathbb R}_+\) be an integrable function with \(\int_{{\mathbb R}^n}f(x)\, dx=1\). \(f\) is isotropic if \(f\) has center of mass at the origin and \(\int_K\langle x, \theta\rangle^2\,f(x)\, dx=1\) for every \(\theta\in {\mathbf S}^{n-1}\). It is well known that the hyperplane conjecture for convex bodies is equivalent to the following statement: There exists an absolute constant \(C>0\) such that, for every isotropic log-concave function \(f\) in \({\mathbb R}^n\), \(f(0)^{1/n}\leq C\). The purpose of the paper is to introduce a different method which leads to upper bounds for \(L_K\). The authors prove that a positive answer to the hyperplane conjecture is equivalent to some very strong probability estimates for the Euclidean norm on isotropic convex bodies. The main idea of the paper is to start from an extremal isotropic convex body in \({\mathbb R}^n\) with maximal isotropic constant \(L_K\simeq L_n=\sup\{L_K: K\) is a convex body in \(\mathbb R^n\}\). Building on ideas from the work [\textit{J. Bourgain, B. Klartag} and \textit{V. D. Milman}, ``Symmetrization and isotropic constants of convex bodies'', Lect. Notes Math. 1850, 101--115 (2004; Zbl 1070.52002)], the authors construct a second isotropic convex body \(K_1\) which is also extremal. The construction of \(K_1\) from \(K\) can be done inside any subclass of isotropic log-concave measures which is stable under the operations of taking marginals or products. In Section 4 the authors introduce a coherent class of probability measures. It should be noted that the class of isotropic convex bodies is not coherent. This is the reason for working with the more general class of log-concave measures.