Atomic vibrations control all thermally activated processes in materials including ionic, atomic and electron diffusion, heat transport, phase transformations and surface chemical reactions. The jump frequency characterizing thermally activated processes is of great practical importance and is determined by the local phonon and molecular vibrational modes of the system. Atomic and molecular heterogeneities and defects such as vacancies, interstitials, dislocations and grain boundaries often regulate kinetic pathways and are associated with vibrational modes which are substantially different from bulk modes. High spatial resolution vibrational spectroscopy is required to probe these defect modes.Recent developments in aberration corrected, monochromated, scanning transmission electron microscopy (STEM) have enabled nanoscale probing of vibrational modes via electron energy-loss spectroscopy (EELS) 1,2 . Nanoscale vibrational spectroscopy is already impacting a wide range of important scientific problems such as measurement of surface and bulk vibrational excitations in MgO nanocubes 3 , probing hyperbolic phonon polaritons in nanoflakes of hBN 4 , measuring temperature in nanometer-sized areas with 1°K precision 5,6 and determining phonon dispersion in nanoparticles 7 . The delocalized nature of certain vibrational signals allows damagefree nanoscale detection for a variety of organic and inorganic material-systems 8-11 . This progress has been impressive, however, to date there have been no experimental methods to spectroscopically probe individual vibrational modes in materials with atomic resolution. Theoretical treatments have explored the question of spatial resolution 12,13 with some treatments suggesting that atomic resolution vibrational EELS should be possible [14][15][16] . Here we demonstrate atomic resolution vibrational spectroscopy in STEM for signals predominantly excited by impact scattering. The resulting order of magnitude advance in spatial resolution will