The aim of this study was to experimentally test the effects of different electromyographic-based prosthetic control modes on user task performance, cognitive workload, and perceived usability to inform further human-centered design and application of these prosthetic control interfaces. We recruited 30 able-bodied participants for a between-subjects comparison of three control modes: direct control (DC), pattern recognition (PR), and continuous control (CC). Multiple human-centered evaluations were used, including task performance, cognitive workload, and usability assessments. To ensure that the results were not task-dependent, this study used two different test tasks, including the clothespin relocation task and Southampton hand assessment procedure-door handle task. Results revealed performance with each control mode to vary among tasks. When the task had high-angle adjustment accuracy requirements, the PR control outperformed DC. For cognitive workload, the CC mode was superior to DC in reducing user load across tasks. Both CC and PR control appear to be effective alternatives to DC in terms of task performance and cognitive load. Furthermore, we observed that, when comparing control modes, multitask testing and multifaceted evaluations are critical to avoid task-induced or method-induced evaluation bias. Hence, future studies with larger samples and different designs will be needed to expand the understanding of prosthetic device features and workload relationships.