@ARTICLE{10.3389/fpsyg.2013.00134, AUTHOR={Van Vugt, Floris and Jabusch, Hans-Christian and Altenmüller, Eckart}, TITLE={Individuality That is Unheard of: Systematic Temporal Deviations in Scale Playing Leave an Inaudible Pianistic Fingerprint}, JOURNAL={Frontiers in Psychology}, VOLUME={4}, YEAR={2013}, URL={https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00134}, DOI={10.3389/fpsyg.2013.00134}, ISSN={1664-1078}, ABSTRACT={Whatever we do, we do it in our own way, and we recognize master artists by small samples of their work. This study investigates individuality of temporal deviations in musical scales in pianists in the absence of deliberate expressive intention. Note-by-note timing deviations away from regularity form a remarkably consistent “pianistic fingerprint.” First, eight professional pianists played C-major scales in two sessions, separated by 15 min. Euclidian distances between deviation traces originating from different pianists were reliably larger than traces originating from the same pianist. As a result, a simple classifier that matched deviation traces by minimizing their distance was able to recognize each pianist with 100% accuracy. Furthermore, within each pianist, fingerprints produced by the same movements were more similar than fingerprints resulting in the same scale sound. This allowed us to conclude that the fingerprints are mostly neuromuscular rather than intentional or expressive in nature. However, human listeners were not able to distinguish the temporal fingerprints by ear. Next, 18 pianists played C-major scales on a normal or muted piano. Recognition rates ranged from 83 to 100%, further supporting the view that auditory feedback is not implicated in the creation of the temporal signature. Finally, 20 pianists were recognized 20 months later at above chance level, showing signature effects to be long lasting. Our results indicate that even non-expressive playing of scales reveals consistent, partially effector-unspecific, but inaudible inter-individual differences. We suggest that machine learning studies into individuality in performance will need to take into account unintentional but consistent variability below the perceptual threshold.} }