We demonstrate neural decoding of imagined and heard musical pitch class from fMRI measurements. Our results extend previous findings by decoding specific patterns of brain-activity across multiple regions that discriminate pitch classes, both heard and imagined. Additionally, a model trained solely on heard pitch class successfully classified imagined pitch class, indicating strong representational overlap between heard and imagined pitch structures. This work deepens our understanding of auditory cognition and volitional control of executive function.

Journal article currently in prep, book chapter awaiting publication.

The efficacy of familiarity as a feature in a music emotion recognition (MER) system was evaluated by a Random Forest feature importance analysis on a novel dataset of 5000 clips with annotated familiarity and valence. Familiarity was correlated to perceived valence (r=0.250) and resulted in a statistically significant increase of 0.011 in the F-score of a baseline MER classifier upon its inclusion. This work was presented at the Music and Machine Learning session at the European Conference of Machine Learning, Wurzburg Germany, 16 Sept 2019. Full paper available here.

Burial, is a critically acclaimed UK electronic musician known for his unusual rhythms. He created his Mercury Prize-nominated album Untrue in Sound Forge 5, an audio editing software with extremely limited functionality, notably the lack of a quantization feature. Through a micro-timing analysis of Burial’s 2007 track “Archangel”, I show that the placement of certain elements off of the conventional metric aid in establishing his unique rhythms. Specifically, the jittered elements cause glitch stutters as well as a perceived rushing of implied space. This paper was given at Rhythm in Music Since 1900, CU Boulder, 18th November 2019.