Neural signal-informed adaptive system for drumming collaboration with AI (BrAIn Jam)

• BrAIn Jam is a neural signal-informed adaptive system for drumming collaboration with an AI-driven virtual musician.¹
• Neural signals drive real-time closed-loop interaction in the system; the use case is BCI-adjacent.¹
• The approach is transferable to other adaptive interfaces and is reported in Frontiers as a proof-of-concept.^{1 1}

Weekly enrichment (2026-07-20)

• BrAIn Jam uses functional near-infrared spectroscopy (fNIRS) to monitor a human drummer’s brain state during real-time improvisation with an AI-driven virtual musician.^{2 3}
• The work was published in Frontiers in Computer Science (2025, volume 7, article 1570249) by Hopkins et al. at the University of Colorado Boulder.²
• The study used a two-phase design: (1) train individualized machine-learning models from data collected in a controlled experiment, then (2) use those models to inform the embodied AI musician during live improvised collaboration.²
• A real-time algorithm preprocesses and classifies fNIRS-derived “rhythmic predictability,” which the AI uses to dynamically adjust its rhythmic patterns.^{2 3}
• fNIRS-driven adaptation carries an inherent ~3–5 second lag because it relies on the slower hemodynamic response rather than direct electrical activity.³
• The authors compared several machine-learning models and ran post hoc brain-activation analysis to corroborate the network involved in music improvisation.²
• The team released their real-time fNIRS preprocessing algorithm to support future fNIRS-based brain–computer interface research.³
• BCI implication: the system demonstrates neural/affective state as a communication channel between humans and embodied AI, but it is a small formative proof-of-concept rather than a clinical BCI, and the participant count is not reported in the primary abstract.^{2 3}

Footnotes

1. https://news.google.com/rss/articles/CBMimgFBVV95cUxQT2l1WUdydFpacWhoY0xKc3lIVk5UQ0l0ZUhxUm5LMDVWR291T2Y4aWdUOHJCSmVrUWFuSGc1Q2FZbUJDUFByNmhsSXJIbDlFemZoZVNwTExsUjBTRnl0YUhzdnI3MFNtNXhZaHdkLVNnVWFPTWI0VHZRaWEzVDI4dlVsWTFnWk1nUFJBa0tLNHhUdEI1OEtIRTVB?oc=5 ↩ ↩² ↩³ ↩⁴

2. https://www.frontiersin.org/journals/computer-science/articles/10.3389/fcomp.2025.1570249/full ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

3. https://doi.org/10.3389/fcomp.2025.1570249 ↩ ↩² ↩³ ↩⁴ ↩⁵