As part of on-going advances in virtual reality (VR) and robotics technologies for rehabilitation applications, augmented sensory feedback (e.g., knowledge of result [KR] and knowledge of performance [KP]) systems have been developed to motivate patients through visual, auditory or haptic modalities. The objective of the present research was to provide a systematic review on types of augmented feedback and modalities that could produce optimum outcomes in terms of learning rate and skill retention. Based on the review, we proposed links between the constructs of a human information processing (HIP) model and motor learning processes. Our findings indicate human perceptual, cognitive and motor performance in training tasks mainly depend on the flow of information processing; whereas, skill learning (expertise) is primarily related to the formation of long-term memory (LTM) and how frequently communications occur among the perceptual process, working memory and LTM. On this basis, KP feedback was expected to support greater learning than KR since it facilitates more frequent memory communications. Similarly, auditory and haptic cues were expected to support greater learning than visual cues due to their retention strength in short-term memory stores. Future directions of empirical research were identified, including assessing the proposed model of human learning under various feedback conditions as well as the effectiveness of recommended modes of feedback in actual training systems.