The perceptual and cognitive underpinnings of emotional responses to music listening are subject to ongoing debate. Here, we investigated three different frameworks—namely, spreading activation theory, emotional processing mechanisms (i.e., the components of the BRECVEMA model), and flow theory (i.e., fluency and absorption)—in an online music listening experiment. A total of N = 99 participants were recruited to take part in a breathing meditation while listening to music of four target categories (familiar/joyful, familiar/sad, unfamiliar/joyful, unfamiliar/sad). Each participant rated their experiences by means of questionnaires. We assumed that listeners’ familiarity with the music would be associated with greater emotionality and flow, and that emotion categories evoked differential patterns of responses. Additionally, we explored the associations of variables representing the three frameworks in a network analysis. We found that familiar music evoked stronger activation and emotional mechanisms, but not stronger flow. There were no differences between sad and joyful music, implying that different models may be needed to capture the elicited and perceived expressive aspects of musical materials. The findings suggest that the three models tap into partially overlapping facets of emotional experience in music listening, in which evaluations of liking, aesthetics, musical expectancies, and autotelic experiences play central roles.