If ultraluminous X-ray sources (ULXs) are powered by accretion onto stellar remnant black holes, then many must be accreting at super-Eddington rates. It is predicted that such high accretion rates should give rise to massive, radiatively-driven winds. However, observational evidence of a wind, in the form of absorption or emission features, has remained elusive. As such, the reported detection of X-ray spectral residuals in XMM-Newton spectra of NGC 5408 X-1, which could be related to absorption in a wind is potentially very exciting. However, it has previously been assumed by several authors that these features simply originate from background diffuse plasma emission related to star-formation in the ULX's host galaxy. In this work we utilise the spatial resolving power of Chandra to test whether we can rule out this latter interpretation. We demonstrate that the majority of the luminosity in these spectral features is emitted from a highly localised region close to the ULX, and appears point-like even with Chandra. It is therefore highly likely that the spectral features are associated with the ULX itself, and little of the flux in this spectral component originates from spatially extended emission in the host galaxy. This may be consistent with the suggestion of absorption in an optically thin phase of a super-Eddington wind. Alternatively, we could be seeing emission from collisionally ionised material close to the black hole, but critically this would be difficult to reconcile with models where the source inclination largely determines the observed X-ray spectral and timing properties. ; Comment: 9 pages, 5 figures, 3 tables, accepted for publication in ApJ