Sensorimotor Oscillations in Human Infants during an Innate Rhythmic Movement

Abstract

The relationship between cerebral rhythms and early sensorimotor development is not clear. In recent decades, evidence revealed a rhythmic modulation involving sensorimotor processing. A widely corroborated functional role of oscillatory activity is to coordinate the information flow across sensorimotor networks. Their activity is coordinated by event-related synchronisation and desynchronisation in different sensorimotor rhythms, which indicate parallel processes may be occurring in the neuronal network during movement. To date, the dynamics of these brain oscillations and early sensorimotor development are unexplored. Our study investigates the relationship between the cerebral rhythms using EEG and a typical rhythmic movement of infants, the non-nutritive sucking (NNS) behaviour. NNS is an endogenous behaviour that originates from the suck central pattern generator in the brainstem. We find, in 17 infants, that sucking frequency correlates with beta synchronisation within the sensorimotor area in two phases: one strongly anticipating (~3 s) and the other encompassing the start of the motion. These findings suggest that a beta synchronisation of the sensorimotor cortex may influence the sensorimotor dynamics of NNS activity. Our results reveal the importance of rapid brain oscillations in infants and the role of beta synchronisation and their possible role in the communication between cortical and deep generators.

Keywords: beta synchronisation; brain oscillations; infants; non-nutritive sucking; sensorimotor.

Figure 1

Methods. (A) View at the infra-red camera during the recording. (B) Illustration of International 10–20 System of EEG electrodes. The red and grey cycles represent the selected electrodes (C3, C4, T3 and T4). The area selected by the red dashed line represent the target region of interest (ROI) analysed. The area selected by the grey dashed line represent the control region of interest (ROI) analysed. (C) Illustration of typical EEG power spectral density (PSD) plot. Blue and yellow areas highlight the EEG rhythms, respectively mu (8–12 Hz) and beta (14–32 Hz). (D) Representative EEG traces of baseline and sucking conditions. ((D), left) Baseline condition is characterized by an EEG recording of a resting-state period sensory deprivation environment during which infants had eyes open. ((D), right) Sucking condition is characterized by an EEG recording during which infants produce a visible NNS behaviour (with pacifier or thumb). Blue and yellow areas highlight examples of mu and beta the EEG rhythms, respectively.

Figure 2

Global power spectral analysis. Figures show the significant correlations between relative spectral power (expressed in %) in mu and beta of the EEG sub-bands during baseline or during sucking conditions and sucking frequency. In red is represented the significant results (p < 0.05 e R² > 0.30) related to baseline condition and in orange the results related to sucking condition. Each point represents a single subject. The shaded areas represent the 95% CI. (A) Correlation between sucking frequency and relative power of mu1 activity (8–10) Hz during resting-state. (B) Correlation between sucking frequency and relative power of mu1 activity (8–10) Hz during sucking condition. (C) Correlation between sucking frequency and relative power of beta activity (14–20) Hz during sucking condition.

Figure 3

Temporal pattern of the beta ERSP and sucking frequency. This pattern emerges from single trial ERSP analysis adopting the normalization described in [46] and expressing ERSP in arbitrary units (AU). 0.5 s time-windows were considered from 3 s to 2.5 s after with respect to the sucking onset. The shaded grey areas represent the significant time-windows emerging from the model. (A) Plot of mean ERSP (larger transparent green circles) and single subjects ERSP (smaller green dots) in each analysed time window. The dashed line points out y = 0 highlighting that all points are positive. (B) Plot of the slopes estimated by the linear mixed model (blue points) and related 95% CI (vertical line) in each analysed time windows. The dashed line points out y = 0 highlighting positive and negative correlations. * indicates p < 0.05, ** indicates p < 0.01.