Detection and analysis of cortical beta bursts in developmental EEG data

Abstract

Developmental EEG research often involves analyzing signals within various frequency bands, based on the assumption that these signals represent oscillatory neural activity. However, growing evidence suggests that certain frequency bands are dominated by transient burst events in single trials rather than sustained oscillations. This is especially true for the beta band, with adult 'beta burst' timing a better predictor of motor behavior than slow changes in average beta amplitude. No developmental research thus far has looked at beta bursts, with techniques used to investigate frequency-specific activity structure rarely even applied to such data. Therefore, we aimed to: i) provide a tutorial for developmental EEG researchers on the application of methods for evaluating the rhythmic versus transient nature of frequency-specific activity; and ii) use these techniques to investigate the existence of sensorimotor beta bursts in infants. We found that beta activity in 12-month-olds did occur in bursts, however differences were also revealed in terms of duration, amplitude, and rate during grasping compared to adults. Application of the techniques illustrated here will be critical for clarifying the functional roles of frequency-specific activity across early development, including the role of beta activity in motor processing and its contribution to differing developmental motor trajectories.

Keywords: Beta bursts; EEG; Infant; Lagged coherence; Oscillations; Power spectral density.

Fig. 1

Sensorimotor beta activity is dominated by transient bursts. A) The power spectral density (blue) after removing the 1/f aperiodic component and lagged coherence averaged over 2 – 2.5 (dark red) and 3 – 3.5 (light red) cycles in the C3 cluster of the adult participants. Dark lines show the mean across participants and the shaded areas indicate the standard error. A prominent peak appears in the alpha range in the power spectrum and lagged coherence at 2 – 2.5 and 3 – 3.5 cycles, whereas a beta peak is visible in the power spectrum and only in lagged coherence at 2 – 2.5 but not 3 – 3.5 cycles. B) Mean lagged coherence over all adult participants in the C3 cluster over a range of lags. There is high lagged coherence in the alpha band over a large range of lags, but beta lagged coherence rapidly decreases after 2 cycles. This same pattern of power and lagged coherence in the alpha and beta bands was also apparent in the adult C4 cluster (C and D), as well as the 12 m infant C3 (E and F) and C4 (G and H) clusters, but each band was evident at a lower peak frequency in the infant data. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)

Fig. 2

Lagged coherence localizes beta activity to central electrode clusters. A) Topography of power in the alpha band, averaged over adult participants. Alpha power is most prominent in central electrodes. B) Alpha lagged coherence topography averaged between 2 and 2.5 cycles (left) and 3 – 3.5 cycles (right), averaged over adult participants. Alpha lagged coherence also localizes to central electrodes, and remains high over several cycles. C) As in (A), for the beta band in adult participants. As with alpha, adult beta power is highest in central electrodes. D) As in (B), for the beta band in adult participants. Lagged coherence in the beta band localizes to central electrodes, but lagged coherence drops off more rapidly with increasing cycles. E) Topographic distribution of power in the alpha band, averaged over infant participants. Infant alpha power localizes to the central and occipital electrodes. F) Topography of alpha lagged coherence averaged between 2 and 2.5 cycles (left) and 3 – 3.5 cycles (right), averaged over infant participants. As in adults, lagged coherence in infant alpha localizes centrally and does not greatly decrease with increasing cycles. G) As in (E), for the beta band in infant participants. Unlike adults, infant beta power is highest in frontal and temporal electrodes. H) As in (F), for beta lagged coherence in infant participants. As in adults, beta lagged coherence is strongest in central electrodes and decreases with increasing cycles.

Fig. 3

Beta burst probability tracks changes in beta amplitude. A) Bursts extracted from the C3 cluster of adult subjects during the baseline and grasp execution epochs of all trials. A raster plot of bursts in each trial (top panel; dashed red lines denote the trials from each subject) shows that bursts occurred more frequently during the baseline epoch. The burst rate (light green) closely tracked changes in beta amplitude (dark green). Solid lines show the mean across participants and shaded regions indicate the standard error. B) Beta amplitude and burst rate also decreased in the C4 cluster of adult subjects during grasp execution. Time zero in the grasp execution epoch represents the first contact of the hand with the object. Burst rate in the C3 (C) and C4 (D) clusters of infants also closely tracked changes in beta amplitude, but neither were modulated by grasp execution. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)

Fig. 4

Infant and adult beta bursts differ in duration and amplitude. A) Infant beta bursts in C3 and C4 have longer duration than adult beta bursts. Each dot shows the duration of an individual beta burst, color-coded by subject. B) When expressed in terms of cycles rather than absolute duration, infant beta bursts in C3 and C4 lasted fewer cycles than adult bursts. C) Infant beta bursts in C3 and C4 tended to have a higher amplitude than adult bursts. D) Even when using the lower threshold identified for burst detection in adults, C3 and C4 infant beta bursts had a higher amplitude than in adults.