Electronic states are paramount to most physical and chemical properties of materials. For instance, electronic orbitals are responsible for chemical bonding between atoms of a crystal. Their experimental observation at defects and interfaces would help understanding material properties better and developing nanostructures with novel functionalities. Nevertheless, the visualization of orbitals in real space at the atomic scale is extremely elusive, and is mostly determined by surface microscopy tools, thus only with surface sensitivity. However, using electron energy-loss spectroscopy (EELS) in the scanning transmission electron microscope (STEM) allows to probe electronic transitions from core levels to momentum- and site-projected empty states, i.e., orbitals, in the bulk of the crystal, as demonstrated in rutile [1]. In this work [2], the capability to map individual electronic states in the TEM is explored from a combined experimental-theoretical approach in epitaxially-grown graphene layers, whose inherent 2-dimensionality presents surfaces as structural discontinuities. Core-level EELS spectra were recorded in a Nion HERMES scanning transmission electron microscope, equipped with a high-energy-resolution monochromator, a Cs aberration corrector up to the fifth order, and operated at 60 kV. The experimental signals are interpreted on the basis of inelastic channelling calculations (ICCs) of the energy filtered maps [3]. The extent of the π* (1s → 2pz) and σ* (1s → 2px, y) state distributions is determined in epitaxial graphene layers observed in side-view, in a 25-nm-thick TEM lamella. Despite higher intensity of the π* fine structures relative to the σ* in between the graphene layers, the absolute π* intensity is higher on the C planes, and as a result π* states appear essentially localized on the graphene layers. However, the π*/σ* ratio map shows intensity maxima in between graphene layers, as highlighted by the π*/σ* and HAADF intensity profiles. All maps are well reproduced by ICCs, as exemplified by the remarkable overlap of experimental and calculated π*/σ* intensity profiles. ICCs also enable an evaluation of the effect of thickness on the orbital contrast. The π* map of an extremely (unrealistically) thin specimen (0.43 nm) displays lobes outside the C planes, and additional intensity is also expected on the C columns. For larger and more realistic thicknesses up to the experimental value of ~25 nm, the π* intensity becomes progressively more important on the C planes than outside, in agreement with experiments. This work highlights the current potential of core-level EELS towards the direct visualization of electronic orbitals in a wide range of materials, which is of interest to better understand chemical bonding among many other properties at interfaces and defects in solids [4].