Back

Brain-responsive music enables non-invasive, targeted and unobtrusive neurostimulation

Zaaimi, B.; Clay, R.; Shivakumar, A.; Rickert, J.; Jackson, A.

2024-07-31 neuroscience
10.1101/2024.07.30.605888 bioRxiv
Show abstract

ObjectiveWe are developing a new closed-loop brain stimulation method by embedding, within music, auditory elements that respond to the listeners brain activity. Here we show that this brain-responsive music has systematic and targeted effects on neural oscillations implicated in a variety of neurological and mental health disorders. ApproachWe recorded magnetoencephalogram (MEG) or electroencephalogram (EEG) signals from participants as they listened to music synthesized by commercial audio software. Brain signals were bandpass filtered, phase-shifted and used to control the timbre and/or timing of notes within the music. Main resultsListening to brain-responsive music induced peaks and troughs in spectral power at frequencies that depended systematically on the phase-shift applied to the brain signal. Phase-dependent modulation was greatest at the centre frequency of the filter. As a result, by calibrating these parameters we could achieve selective enhancement or suppression of either theta (5 Hz) or alpha (10 Hz) oscillations. Moreover, by chosing different sensor locations we could target power modulation to either frontal or temporal cortex. The phase-dependent power modulation observed with brain-responsive music was significantly attenuated when participants listened to identical music as a conventional, open-loop stimulus. Finally, we demonstrate that brain activity could be modulated by more complex compositions combining a variety of brain-responsive musical elements controlled by a wireless, wearable EEG headband suitable for home use. SignificanceBrain-responsive music provides an unobtrusive and targeted method of modulating neural oscillations in the listeners brain, and may enable both creative and therapeutic applications of Brain Computer Interface technologies.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.