The present study joins a series of studies that used the dual-task paradigm to measure cognitive load while learning with multimedia instruction. The goal of the current work was to develop a secondary task, to measure cognitive load in a direct and continuous way using intra-individual, behavioral measures. The new task is achieved by utilizing internalized cues. More specifically, a previously practiced rhythm is executed continuously by foot tapping (secondary task) while learning (primary task). Precision of the executed rhythm was used as indicator for cognitive load-the higher the precision, the lower cognitive load. The suitability of this method was examined by two multimedia experiments (n₁=30; n₂=50). Cognitive load was manipulated by seductive details (Experiment 1: with vs. without) and modality (Experiment 2: on-screen text vs. narration). Learners who learned under low cognitive load conditions (Experiment 1: without seductive details; Experiment 2: narration) showed significantly higher rhythm precision. These results provide evidence that rhythm precision allows for a precise and continuous measurement of cognitive load during learning.