Researchers develop an electroencephalography-based system that provides real-time music suggestions based on the user’s current state

Musical chills are pleasurable shivers or goosebump sensations that people feel when they resonate with the music they’re listening to. They reduce stress and have beneficial side effects, but they are difficult to induce reliably. Now, researchers from Japan have developed a practical system that uses in-ear electroencephalography sensors to measure the brain’s response to music in real time and provide music suggestions that enhance chills.

Most people are familiar with ‘musical chills’—a sudden, involuntary shiver or goosebump sensation that occurs when a song resonates perfectly with one’s emotions. These chills are not just a surface-level feeling, but a profound neurological event. When we experience intense musical pleasure, parts of the brain’s reward system activate in a manner similar to how they would respond to life-affirming stimuli, such as beloved foods or positive social connections. However, despite the universal nature of the experience, musical chills are difficult to trigger reliably. This limits our ability to harness their psychological and physiological benefits, even with today’s on-demand access to vast libraries of music.

The issue lies partly in how ‘personalized’ media streaming algorithms are designed. These systems are essentially blind to a user’s current internal state and rely instead on historical data, genre preferences, or acoustic features like tempo to suggest content. Therefore, they cannot detect whether a listener is actually connecting with a specific track at the moment to improve suggestions. But what if our earphones could read the brain’s response to music as it happens?

In a study recently published in Volume 29, Issue 1 of iScience, a research team led by Dr. Shinya Fujii (Associate Professor, Corresponding author) and Dr. Sotaro Kondoh (JSPS Research Fellow, first author) of the Faculty of Environment and Information Studies at Keio University, Japan, has developed a novel solution to bring real-time brain monitoring into everyday music listening. Their paper, published on January 16, 2026, introduces the ‘Chill Brain-Music Interface (C-BMI),’ a system that employs compact in-ear electroencephalography (EEG) sensors to decode pleasure from the listener’s brain and use that data to build an optimized playlist in real time.

The methodology involved a multi-phase approach comprising recording, modeling, and evaluation. Initially, each participant listened to self-selected songs (high pleasure) and songs selected by another participant (low pleasure) while their brain activity was monitored via the in-ear sensors. From this data, the researchers built two personalized models for each participant. The first model predicted pleasure based on the acoustic features of the music, while the second decoded the listener’s pleasurable state from the EEG signals. Finally, these models were combined in closed-loop systems that generated playlists designed to either augment or diminish pleasure, with real-time neural decoding incorporated in some playlists but not others.

The results conclusively demonstrated that the EEG-updated playlist for augmenting pleasure was more effective than the other generated playlists. In other words, participants reported the highest number of chills and superior pleasure ratings when the system was allowed to adapt to their brain activity. The researchers found that C-BMI successfully captured neural markers of pleasure with high accuracy and, when using them to suggest music, outperformed the ‘acoustics-only’ methods that mimic current commercial streaming services.

A critical advantage of the proposed system is its size, especially when compared to brain-computer interfaces with more conventional EEG recording setups used in laboratories. “Using an in-ear EEG device with earphones offers a practical, comfortable, and accessible alternative, opening new possibilities not only for neuroscience research investigating music and the brain—which is fundamental to understanding human nature—but also for future daily applications in entertainment and human–computer interaction,” says Dr. Fujii.

Notably, the research team explored how this technology could impact broader psychological well-being. Participants using the EEG-informed playlists reported higher scores in categories such as stress reduction and a sense of life purpose, showcasing the benefits of aligning music selection with a person’s fluctuating internal state. “If this non-invasive system could be integrated with in-ear earphones equipped with EEG sensors and wellness and music streaming apps, it could offer emotional support during commutes, before sleep, or in daily life,” remarks Dr. Kondoh.

As mental distress becomes increasingly prevalent, especially among young people, C-BMI may prove effective as a low-barrier intervention, ideal for those who might hesitate to seek traditional psychiatric care. The researchers will continue to refine the C-BMI and similar interfaces, hoping to develop systems that enhance the positive effects of the simple and universal act of listening to music.


Reference
DOI: https://doi.org/10.1016/j.isci.2025.114508