This decade-long review examined BCI-based working memory interventions published between 2014 and 2024, focusing on the design elements that influence usability, engagement, and training effectiveness. While existing studies often report positive outcomes, the design features critical to user experience and long-term applicability remain underreported. Conducted in accordance with the PRISMA 2020 guidelines, this systematic review analyzed 13 eligible studies identified through a structured search across four databases to identify and compare design elements. Continuous visual feedback and personalized neurofeedback thresholds were the most common, with N-back, Sternberg, and Corsi tasks frequently used to modulate cognitive load. Adaptivity was widely implemented, yet older adults were underrepresented, and usability evaluations were rare. Gamification appeared in fewer than half of the studies, often without systematic assessment of its impact. Although many systems incorporated personalization and adaptive difficulty, the lack of user-centered design, especially for older or clinical populations, highlights a gap between experimental prototypes and real-world applications. The review concludes with preliminary design element guidelines to inform the development of effective, accessible, and engaging BCI-based working memory systems.