Abstract Multimedia systems involve the association of various types of information: verbal information presented visually or auditorily, static or dynamic pictorial information, and sound information. In a cognitive approach, integrating this information involves complex processes constrained by properties of the learner's cognitive system, and especially by the capacity of working memory. This paper, reports on an experiment focused on the integration of verbal and pictorial information when students learn a series of physics concepts. The involvement of the visuo-spatial working memory was investigated by means of a dual-task paradigm. Results show that pictorial information enhances the learning process. They also suggest that the visual and the spatial components of visuospatial working memory should be considered separately. Finally, they emphasise the need to consider the limitations in cognitive resources available to the learner.