Y. Ikegaya
Among the human senses, hearing is distinguished by its high frequency resolution and sensitivity to spatiotemporal patterns. Accordingly, sonification--the rendering of neural activity as sound for direct perception--has a long history in neuroscience. Although sonification can facilitate exploration of the temporal structure of population activity, dense spike trains are often difficult to track by ear. Here, I present HERO (Hierarchy-guided Event-faithful Reproducible Orchestration), a pipeline that assigns stable musical voices to recorded units while preserving spike onsets in a MIDI event representation. Each spike within a selected interval produces a single Note On message whose timing is determined by a global temporal dilation followed by bounded clock rounding. A separate analysis pathway uses multiscale correlations and block-bootstrap co-association to organize units into a hierarchy. This hierarchy guides instrument assignment, whereas each unit is assigned a fixed key, tuning offset, velocity, and stereo position. A derived SoundFont stores tuning and pan parameters independently for units that share a MIDI channel. Rate-dependent velocity and channel-volume settings are intended to reduce the dominance of frequently firing units. An online audio example illustrating the audibility of individual activity during large-scale firing is available at: https://www.youtube.com/watch?v=zOYGXcp727A. HERO provides a reproducible auditory representation of population activity that may help direct attention toward features warranting subsequent quantitative analysis.