Not yet reviewed by Pith; the record is open.

Central depression of nucleonic density, i.e., a reduction of density in the nuclear interior, has been attributed to many factors. E.g., bubble structures in superheavy (SH) nuclei are believed to be due to the electrostatic repulsion. In light nuclei, the mechanism has been discussed in terms of shell effects associated with occupations of s-orbits. The objective of this work is to reveal mechanisms behind the formation of central depression in nucleonic densities in light and heavy nuclei. We introduce several measures of the internal nucleonic density. Through the statistical analysis, we study the information content of these measures with respect to nuclear matter properties. We apply nuclear density functional theory with Skyrme functionals. Using the statistical tools of linear least square regression, we inspect correlations between various measures of central depression and model parameters, including nuclear matter properties. We study bivariate correlations with selected quantities as well as multiple correlations with groups of parameters. Detailed correlation analysis is carried out for $^{34}$Si for which a bubble structure has been reported recently, $^{48}$Ca, and N=82, 126, and 184 isotonic chains. We show that the central depression in medium-mass nuclei is very sensitive to shell effects, whereas for SH systems it is firmly driven by the electrostatic repulsion. An appreciable semi-bubble structure in proton density is predicted for $^{294}$Og, which is currently the heaviest nucleus known experimentally. Our correlation analysis reveals that the central density indicators in nuclei below $^{208}$Pb carry little information on parameters of nuclear matter; they are predominantly driven by shell structure. On the other hand, in the SH nuclei there exists a clear relationship between the central nucleonic density and symmetry energy.