December 19, 2019

Your Brain on the Dance Floor

Neural Oscillations Support Musical Rhythm Perception

While it seems rather banal, humans’ ability to detect the beat in music is, in fact, a rare feat of computation. Although recent evidence suggests that some animal species may be capable of synchronizing their motor activity to a repeating auditory stimulus (see Alex the parrot and Ronan the sea lion [1]), humans have the unique ability to infer periodicity in auditory stimuli which may not explicitly contain that periodicity.
What does that mean, exactly? Give a listen to James Brown’s 1968 hit “I Got the Feelin’”[2]. The song is rhythmically complex, with accents falling in unexpected places, and yet, moving to the beat is gratifyingly easy. However, in order for you to continue feeling the pulse while Brown and his band weave in and around the beat, significant cognitive and perceptual computations are required. Neural oscillations provide the means for those computations, and thus, subsequent hip shakin’. But, before delving deeper, let’s define some terminology at the intersection of music and brain science.
Musical Time and Rhythm
Time in music is hierarchical, with divisions broader and finer than the fundamental pulse at which you might tap your foot. In Western music, that fundamental pulse is called tempo. Tempo describes the rate of evenly spaced musical beats and is quantified in beats per minute (BPM). However, as scientists, we may prefer to describe tempo in Herz (oscillations per second) as a measure of frequency. If you venture into a Berlin club, the techno music you hear will likely have a tempo near 120 BPM. Described as a beat frequency, this tempo would be 2 Hz (120 BPM/60 sec = 2 beats per second) [3].
The hierarchical level above the beat is called meter, which describes the grouping of beats into a unit called a bar or measure. In popular Western music, a bar typically has 4 beats, creating a quadruple meter. A waltz is an example of a musical style in which 3 beats define a bar, creating a triple meter [4].
Tempo and meter provide a temporal framework against which we can perceive rhythm, which we’ll define as an expressive pattern of musical note onsets, durations and accents. The music of James Brown is full of excellent examples of rhythmic syncopation, where the musical content does not always fall on the beat [5]. While the familiar ‘4 on the floor’ of techno music clearly indicates each beat, in the case of “I Got the Feelin’”, some notes in the rhythm fall in-between beats, while other beats are left silent. Syncopation builds a sensation of rhythmic tension and, when wielded properly by Brown, et al., it creates a powerful urge to dance.
Entrainment to the Beat
If we were to look at “I Got the Feelin’” as a time-series of sonic amplitude events, we would see that musical sound does not occur on every beat, yet we feel the beat continuing unbroken as the band syncopates. We perceive a periodic pulse despite the fact that the music does not explicitly sound that pulse [9]. In order for a listener to feel the beat as a regular framework continuing steadily despite rhythmic variations in the song, recent music neuroscience research indicates that neural oscillations in the auditory pathway entrain to the beat. Entrainment describes the ability of ensembles of sensory neurons to adjust the phase or period of their oscillations in order to synchronize to a regularly repeating external stimulus [6]. In the case of our James Brown song, neurons in your auditory pathway adapt their cycles so that their oscillations align with the foot-tapping pulse.
Several lines of cognitive neuroscientific theory converge on the idea that by synchronizing the rhythmic activity of neurons in perceptual networks to the rhythmic activity of repeating external stimuli, organisms are able to improve their perceptual accuracy and predict the future activity of that stimulus [7, 8]. If oscillatory activity in a listener’s auditory networks is synchronized to a repeating musical beat, that listener is able to predict when the next beat will happen. Dancers, musicians and even passive listeners are not reacting to every beat with surprise. Rather, they can predict that the beat is coming, given the internal metronome provided by entrained neural oscillations.
Ongoing entrained oscillations allow the human brain to track a musical rhythmic stream and also provide a temporal framework against which syncopation and other complex rhythms can be judged. When James Brown’s band plays an unexpected phrase, the listener understands this rhythmic surprise as a variation against the internalized, ongoing temporal framework. Interestingly, this phenomenon is proposed to exist in the visual domain as well. When we look at a bistable percept like Rubin’s famous vase/face, where one might see a white vase against a black background or two black faces in profile against a white background, neural oscillations in the visual pathway code for figure vs. ground. Similarly, neural oscillations underlie the perception of a musical background (the pulse of the tempo) against which we perceive a musical figure (a syncopated rhythm or an expressive solo) [5, 6].
Neural Oscillations Track the Beat and Meter
In examining EEG data recorded from the brains of subjects exposed to rhythmic music, oscillations can be identified whose frequency corresponds to the tempo of the music they hear. If we examine the EEG time-series of a person listening passively (EEG is no good when you’re dancing) to that techno track at 120 BPM, we would see ongoing oscillations, stable in phase and period (called steady-state evoked potentials or SS-EPs), occurring at 2 Hz. Additionally, because musical meter is often marked by accentuation on the first note of the bar, we would observe a sub-harmonic of the beat frequency oscillation corresponding to the meter frequency; in the case of our quadruple meter techno song, 0.5 Hz [6].
Given the steady beat of techno, it’s easy to imagine cells in your auditory pathway firing every time the bass drum hits. However, in the case of Brown’s “I Got the Feelin’”, where not every beat is sounded and some notes happen in unexpected places, it becomes clear that entrained oscillations are the key to tracking the beat despite complexities and variations in rhythm. The repetition in Brown’s music reinforces a listener’s perception of the beat, but even in the case of improvised, non-repetitious jazz music, listeners’ entrained neural oscillations provide a guide through which the rhythmic acrobatics of the band can be understood.
By allowing a listener to infer the beat, entrained neural oscillations give James Brown the opportunity to create rich rhythmic complexity safe in the knowledge that his fans will not trip over themselves. And, for those who feel inherently un-rhythmic, there is hope. Musical training has been shown to increase the strength of neural entrainment to musical rhythm and to improve musical beat tracking [3]. Few people are born with the rhythmic aptitude of the Godfather of Soul, but practice can improve anyone’s beat keeping. In addition, the hierarchical nature of musical time and the non-linear dynamics of the neural oscillations involved in perceiving that time mean that it’s possible for different listeners to feel the beat at different harmonics of the fundamental pulse. So, leave any shyness behind you and take your brain for a workout on the dance floor. Let your oscillating neurons guide you and you’ll never lose the beat.

Steve's Master's Thesis
Neural entrainment is the process by which ensembles of neurons in sensory networks synchronize their oscillation in phase and period to that of rhythmic, repeating external stimuli. In music perception, evidence suggests that neural oscillations entrain to the frequencies of musical beat and meter. Polyrhythms are complex rhythmic structures in which two non-factorial rhythms co-occur over a common tempo. For example, if you walk down the street, your feet create a duple rhythm (based in two): left-right, left-right. If, while walking, you repeatedly count to three (a triple rhythm) in time with your footsteps, you will be the embodiment of a three-over-two (3:2) polyrhythm.
To investigate the effect of musical experience on entrainment strength to musical polyrhythm and to determine if stronger entrainment correlates with improved musical performance, Prof. Dr. Gabriel Curio and Dr. Gunnar Waterstraat of the Charité Neurophysics Group, Carola Bothe of the Freie Universität Department of Computer Science, and myself conducted an EEG study with a musical behavior performance task. While EEG was recorded, subjects listened to a 3:2 polyrhythm and were asked to complete either the duple or triple rhythm by striking an electronic drum pad when cued for that rhythm. We collected information on subjects’ musical background and compared this with their accuracy on the musical task and with EEG measures of neural entrainment to the polyrhythm stimulus.
We found strong relationships between musical experience and performance on the musical task. Increased musical experience correlated negatively with performance error and positively with performance consistency. Additionally, musical experience correlated positively with entrained oscillatory power at rhythm-related frequencies. A strong positive correlation was found between musical experience and entrainment at the first common harmonic of the duple and triple rhythm frequencies. This suggests that with musical training, the brain can increasingly track the separate components of a polyrhythm in an integrated manner, via oscillations entrained at frequencies that encompass both rhythms.
A significant relationship between entrainment strength and accuracy on the musical task was also present in the data. By employing spatial filtering algorithms adapted to the purpose by Dr. Waterstraat, a strong negative correlation was found between the power of oscillations entrained at the duple rhythm frequency and error on duple rhythm trials of the musical behavioral task.
The results of the study strongly suggest that by entraining oscillations in the auditory pathway to the components of musical rhythm, the brain gains accuracy in musical rhythm performance, and that musical training can increase the strength of this entrainment. Additionally, the work showed the value of interdisciplinary scientific collaboration, given the team’s skill sets which ranged from neurology and electrophysiology to advanced algorithmic coding to experience as a professional musician. Dr. Curio’s interest in leading a team with a diversity of perspectives and training created a very productive and collaborative environment for research, which I’m grateful to have been a part of.

Steve Garofano
Berlin School of Mind & Brain, M.Sc.
Neurophysics Group, Charité-Universitätsmedizin Berlin 

References:
1: Rouse et al., Front Neurosci 2016
2: Brown, J. (1968). I got the feelin’. On I got the feelin’ [LP]. Cincinnati, OH: King Records.
3: Levitin et al., Annu Rev Psychol 2017
4: Grahn, Top Cogn Sci 2012
5: Vuust, P. et al Neural underpinnings of music: the polyrhythmic brain. In Neurobiology of interval timing, 2014
6: Nozaradan, Philos Trans Roy Soc 2014
7: Jones and Boltz, Psychol Rev 1989
8: Lakatos et al., Science 2008
9: Large and Snyder, Ann NY Acad Sci 2009

No comments:

Post a Comment