Background: The integration of multimedia e-content into higher education has transformed how adult learners engage with complex information. But preventing instructional failure requires a deep understanding of human cognitive architecture. Designing multimedia based on intuition rather than empirical evidence risks overloading a learner's working memory. Consequently, this systematic literature review analyzes empirical evidence published before 2019, investigating the effects of interactive multimedia instructional design on cognitive processing, working memory capacity, and the retention and transfer of information.

Objectives: This review addresses three core objectives. First, it examines how diverse multimedia formats influence the intrinsic, extraneous, and germane dimensions of cognitive load. Second, it critically evaluates the research methodologies and measurement instruments employed to quantify cognitive load. Lastly, it synthesizes evidence-based instructional design principles—such as instructional cueing, the modality effect, and the split-attention effect—to offer practical strategies that demonstrably maximize adult learner retention in digital environments.

Methods: To establish a baseline of multimedia research prior to the global pandemic shift, a systematic retrieval was conducted using the database of Scopus, Web of Science, EBSCOhost. This was executed strictly adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2009 guidelines. The temporal inclusion criterion strictly excluded any study published on or after January 1, 2020. Data extraction rigorously mapped multimedia formats, statistical effect sizes, such as Cohen’s , and learner metrics, encompassing both subjective survey instruments and objective physiological measures like EEG.

Results: An analysis of the 20 key studies meeting the rigorous inclusion criteria—spanning large-scale meta-analyses to highly controlled randomized experiments—reveals critical insights into cognitive load and methodology. Methodologically, the existing literature underscores widespread challenges, particularly the common conflation of intrinsic and germane load in subjective questionnaires (Klepsch et al., 2017) ^[10], alongside the rare but essential adoption of objective physiological measurement tools, such as EEG-based Partial Directed Coherence (Mazher et al., 2017) ^[14]. Regarding instructional outcomes, the findings overwhelmingly demonstrate that minimizing total cognitive load through instructional cueing significantly enhances both transfer and retention (Xie et al., 2017) ^[20]. Additionally, integrating representational video with auditory narration strategically boosted comprehension for non-native English speakers ( ) while avoiding extraneous overload (Mayer et al., 2014).