This thesis is an investigation of the feasibility of training protocols to alter dualtask performance in complex environments. The performance of a cognitive task during whole body movements is well established to cause performance decrements in either one of, or both tasks. Furthermore, these performance decrements are amplified as tasks become more challenging and in certain special populations, such as older adults. Research groups have developed training protocols to examine the possibility of reducing dual-task interference effects. However, it is currently unknown if training will transfer to improvements in more complex environments. The current thesis is composed of four experiments. The first experiment was designed to better understand the effects of structural interference on the concurrent performance of an obstacle avoidance task. Results from this experiment demonstrated larger dual-task interference effects on a concurrent obstacle crossing task in the presence of structural interference, indicating the selection of a cognitive task to pair with obstacle crossing has a major influence on performance outcomes. Experiments II & III examined different durations and compositions of training protocols in young adults designed to alter performance on a complex test of dual task performance (a concurrent obstacle crossing and auditory Stroop task). In Experiments II & III, dual-task training protocols composed of performing simultaneous walking and cognitive tasks were observed to increase minimum trail foot clearance in young adults, indicating a potentially more cautious obstacle crossing strategy following training. Conversely, in Experiment III, the training program based on computer-game based training did not alter obstacle crossing strategies to the same magnitude. Finally, the fourth experiment applied the most effective training protocol from the previous studies to a population of community dwelling older adults. Results from the experiment indicated older adult participants exposed to walking based dual-task training did not alter obstacle clearance values, but instead walked more slowly during the dual-task test following training, indicating an alteration to allow more time to implement an obstacle crossing strategy. Collectively, the results of this dissertation indicate that dual-task performance is largely modulated by the tasks included in the experimental protocol. Furthermore, the effects induced by a walking based dual-task training protocol to performance on a concurrent obstacle crossing and auditory Stroop test depend on the population tested.