This study aims at investigating students' strategies--as revealed by behavioural, psychophysiological and introspective measures--which are applied during the free exploration of multimedia instructional presentations, which requires students to self-regulate their learning processes. Two multimedia presentations were constructed and presented to a sample of 20 undergraduates in two conditions: written text + pictures vs. audio text + pictures. While students were engaged in the study of the presentations, their eye movements were registered and psychophysiological indices were monitored. Students' learning outcomes were assessed and a questionnaire was employed to record students' awareness of the mental processes involved in the task. Results showed that students were able to discriminate between the written- and audio-text conditions and self-regulate their behaviour accordingly. A model, assuming psychophysiological indices as predictors of different eye-movement patterns, highlighted significant differences between the written- and the audio-text conditions. A regression model, considering learning outcomes as a dependent variable, showed that the number of correct responses could be predicted according to the level of cognitive effort needed during the exploration of the multimedia presentations.