Tagging the neuronal entrainment to beat and meter

Abstract

Feeling the beat and meter is fundamental to the experience of music. However, how these periodicities are represented in the brain remains largely unknown. Here, we test whether this function emerges from the entrainment of neurons resonating to the beat and meter. We recorded the electroencephalogram while participants listened to a musical beat and imagined a binary or a ternary meter on this beat (i.e., a march or a waltz). We found that the beat elicits a sustained periodic EEG response tuned to the beat frequency. Most importantly, we found that meter imagery elicits an additional frequency tuned to the corresponding metric interpretation of this beat. These results provide compelling evidence that neural entrainment to beat and meter can be captured directly in the electroencephalogram. More generally, our results suggest that music constitutes a unique context to explore entrainment phenomena in dynamic cognitive processing at the level of neural networks.

Figure 1.

Experimental design. A, A 6 s excerpt of the 33 s auditory stimulus (x-axis: time; y-axis: sound amplitude). Note the pseudo-periodic beat structure, visible as a slightly irregular modulation of amplitude. From this pseudo-periodic stimulus, subjects perceived the 2.4 Hz beat represented in blue. In the binary and ternary meter conditions (shown in red and green, respectively), subjects were asked to imagine a binary (1.2 Hz) and ternary (0.8 Hz) metric structure onto this beat. B, The frequency spectrum of the sound stimulus. Note the peak corresponding to the frequency of the tone (333.33 Hz), as well as the sideband frequencies resulting from the convolution of that carrier frequency with the two different amplitude modulation frequencies (2.4 and 11 Hz). C, The frequency content of the sound envelope obtained by convoluting the two different amplitude modulation frequencies (2.4 and 11 Hz). Note that the sound envelope contains a peak at the frequency corresponding to the beat (2.4 Hz), but does not contain any sideband frequencies at the frequency of the binary (1.2 Hz) and ternary (0.8 Hz) meters.

Figure 2.

Beat- and meter-related steady-state EPs recorded in a single representative subject. Bottom, The EEG amplitude spectrum (in microvolts) from 0 to 45 Hz, averaged across all scalp electrodes, after applying the noise subtraction procedure (see Materials and Methods). The EEG spectrum obtained in the control condition is shown in blue, whereas the EEG spectra obtained in the binary and ternary meter imagery conditions are shown in red and green, respectively. Middle, The EEG amplitude spectrum (in microvolts) within a frequency range comprising the frequency of the beat (2.4 Hz) and the frequency of the imagined binary and ternary meters (1.2 and 0.8 Hz, respectively). Note that in all three conditions, the auditory stimulus elicited, at f = 2.4 Hz, a clear beat-related steady-state EP. Also note the emergence of a meter-related steady-state EP at 1.2 Hz in the binary meter imagery condition, and at 0.8 and 1.6 Hz in the ternary meter imagery condition. Top, The topographical maps of EEG signal amplitude at 0.8, 1.2, 1.6, and 2.4 Hz, obtained in each of the three conditions.

Figure 3.

Group-level average of the beat- and meter-related steady-state EPs elicited by the 2.4 Hz auditory beat in the control condition (top), the binary meter imagery condition (middle), and the ternary meter imagery condition (bottom). The frequency spectra represent the amplitude of the EEG signal (in microvolts) as a function of frequency, averaged across all scalp electrodes, after applying the noise subtraction procedure (see Materials and Methods). The group-level average frequency spectra are shown using a thick colored line, while single-subject spectra are shown in gray lines. Note that in all three conditions, the auditory stimulus elicited a clear beat-related steady-state EP at f = 2.4 Hz. Also, note the emergence of a meter-related steady-state EP at 1.2 Hz in the binary meter imagery condition, and at 0.8 and 1.6 Hz in the ternary meter imagery condition.

Figure 4.

Amplitude of the beat- and meter-related steady-state EPs elicited in the control condition, the binary meter imagery condition, and the ternary meter imagery condition. Dots represent individual amplitude values of the EEG signal for each experimental condition at each target frequency (1.2, 0.8, 1.6, and 2.4 Hz), averaged across all scalp electrodes after applying the noise subtraction procedure. The whisker plots represent the group-level median and interquartile range.

Figure 5.

Transient auditory event-related potentials elicited by the 33 s sound stimulus (group-level average waveforms recorded at FCz) in the control condition (blue), the binary meter condition (red), and the ternary meter condition (green). Top, The onset of the auditory stimulus elicited a clear auditory evoked potential consisting of a negative peak (N1) followed by a positive peak (P2). In contrast, beat onsets (represented by the dashed vertical lines) did not elicit a measurable transient event-related potential.