The use of decision directed (DD) channel estimation in a multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) downlink receiver is studied in this paper. The 3GPP long term evolution (LTE) based pilot structure is used as a benchmark. The space-alternating generalized expectation-maximization (SAGE) algorithm is used to improve the performance from that of the pilot symbol based least-squares (LS) channel estimator. The DD channel estimation improves the performance with high user velocities, where the pilot symbol density is not sufficient. It can also be used to reduce the pilot overhead without any performance degradation. Minimum mean square error (MMSE) filtering can also be used in estimating the channel in between pilot symbols. The pilot based LS, MMSE and the SAGE channel estimators are implemented and the performance-complexity trade-offs are studied.