The heartbeat music interaction has been conceivable, but it is not been scientifically pursued yet. In this study, we developed a heartrate music feedback architecture where the tempo of the music track was continuously changed in accordance with users' heart rate in a real-time manner and conducted a preliminary experiment to explore its psychophysiological effect in a laboratory setting. Compared with the control condition, in which the tempo of the music tracks was constant, there were significant differences among conditions in the respiration intervals, heart rate variability, and the beta power of brain waves. Furthermore, there was no distinct difference in the subjective scores and impressions for the musical tracks. The results imply that the difference in the physiological responses between the conditions may be derived not from perceptible or recognizable differences in music, but from purely physiological functioning at unconscious level.