Can music and drawings be thought of as the results of physical gestures, and thus be compared? In this paper we summarize the conjecture of "gestural similarity" developed in the framework of the mathematical theory of musical gestures. Then, we outline the history of an experiment involving mathematics, music, drawing, and psychology, aiming to evaluate the cognitive relevance of the conjecture. A simple visual form and a short homophonic musical sequence can be considered "similar" if they can be thought of as produced by the same movements. Participants in an experiment were asked to assess the degree of similarity between given music examples and simple visuals (three visuals for each sound stimulus). Results were analyzed and confirmed the theoretical expectations. In addition, we describe some creative applications of this conjecture, including pedagogical and creative developments. In particular, we describe the music derived from a natural form, the essential structure of an ammonite, and the illusion of a "mathematical ocean" with sounds and images. We discuss challenges of these techniques and the characteristics of spectrograms in relation with gestural similarity.