Spectro-astrometry of the pre-transitional star LkCa 15 does not reveal an accreting planet but extended Hα emission

(Abridged) The detection of forming planets in disks around young stars remains elusive, and state-of-the-art observational techniques provide somewhat ambiguous results. It has been reported that the pre-transitional T Tauri star LkCa 15 could host three planets; candidate planet b is in the process of formation, as inferred from its H$\alpha$ emission. However, a more recent work casts doubts on the planetary nature of the previous detections. We have observed LkCa 15 with ISIS/WHT. The spectrograph's slit was oriented towards the last reported position of LkCa 15 b (parallel direction) and 90degr from that (perpendicular). The photocenter and full width half maximum (FWHM) of the Gaussians fitting the spatial distribution at H$\alpha$ and the adjacent continuum were measured. A well-known binary (GU CMa) was used as a calibrator to test the spectro-astrometric performance of ISIS/WHT, recovering consistent photocenter and FWHM signals. However, the photocenter shift predicted for LkCa 15 b is not detected, but the FWHM in H$\alpha$ is broader than in the continuum for both slit positions. Our simulations show that the photocenter and FWHM observations cannot be explained simultaneously by an accreting planet. In turn, both spectro-astrometric observations are naturally reproduced from a roughly symmetric Halpha emitting region centered on the star and extent comparable to the orbit originally attributed to the planet at several au. The extended H$\alpha$ emission around LkCa 15 could be related to a variable disk wind, but additional multi-epoch data and detailed modeling are necessary to understand its physical nature. Spectro-astrometry in H$\alpha$ is able to test the presence of accreting planets and can be used as a complementary technique to survey planet formation in circumstellar disks.