Exploring the Mirror Neuron Mechanism through the electroencephalogram (EEG) in adult and newborn macaque monkeys

Abstract

The Mu rhythm is an electroencephalogram (EEG) oscillation detected over central scalp locations that was first described, by Gastaut and Bert in 1954, as a rolandic rhythm falling in the alpha frequency band (8-13 Hz) and reflecting the activation of sensory motor cortex. It reaches maximal amplitude when individuals are at rest, while desynchronizes (i.e. decreases in amplitude) during preparation, execution or imagination of movements and during observation of movements done by others (Pineda et al., 2005). More recent MEG investigations revealed that the rolandic rhythm recorded over central electrodes may result from two different brain components: one that peaks at around 10 Hz (alpha band) and originates from the primary somatosensory cortex and one that peaks at around 20 Hz (beta band) and is clustered anterior to the central sulcus, suggesting a primary contribution from the precentral motor cortex (Hari and Salmelin, 1997). Confirming this finding, several EEG works in humans have shown desynchronization in beta frequencies (13-25 Hz) over central electrodes, while subjects executed or observed actions (Babiloni, et al., 2002; Muthukumaraswamy & Johnson, 2004; Avanzini, et al., 2012). In the last two decades, particular attention has been paid to mu rhythm and its desynchronization/suppression during action execution and observation as it is thought to reflect neural responses related to the mirror neuron system (MNS) in humans. However, EEG represents only an indirect measurement of mirror neurons activity, while most of our knowledge about functions and properties of putative mirror neurons comes from electrophysiological investigations conducted with macaque monkeys, by means single-unit recordings, in the ventral premotor cortex (area F5) and in the posterior parietal cortex (area PFG). In Experiment 1, as a first step in bridging the knowledge gap between EEG during action observation that is recorded from the human scalp and the extensively studied MNS in macaques, we sought to determine whether an analogue of human EEG is recordable on the scalp of two adult rhesus macaques, and whether it was possible to modulate macaque EEG response through action observation. We focused on three sensorimotor frequency bands: alpha band (7-13 Hz), lower beta (13-19 Hz) and upper beta (19-25 Hz). Results revealed that the 19-25 Hz band was the most sensitive to EEG activity modulation during action observation. Moreover desynchronization in this band was greater over anterior and central electrodes compared to posterior electrodes. This finding results in line with other works in humans showing desynchronization in the beta band, while subjects are observing actions done by others. The monkey 19-25 Hz band may be correspondent to the human beta band and oscillations in this band may reflect the activation of motor areas during action observation and therefore could be considered as an indirect marker of mirror neuron activity. In Experiment 2, we investigated scalp EEG activity in newborn monkeys to explore if specific frequency bands may be sensitive to goal-directed actions observation and, possibly, to describe any developmental change, in the EEG oscillations, occurring over the first weeks of life. Previous EEG investigations in newborn monkeys have shown desyncronization in 5-7 Hz (an EEG band that falls in similar frequencies as mu rhythm in humans, during infancy) during imitation and observation of facial gesture; however, to date, no study has investigated infant macaques cortical activity during the observation of goal directed hand action. We focused on two specific frequency bands, following previous approaches used in adult and newborn monkeys: 5-7 Hz (mu band) and 15-17 Hz (beta band). Results show a significant desynchronization in the 15-17 Hz band (but not in the 5-7 Hz band) in anterior electrodes for the observation of grasping actions. This desynchronization emerges starting from the second week of life while no significant desynchronization was found during the first week of life (neither in alpha nor in beta frequencies). This finding is in line with our work on adult monkey, showing EEG suppression in the beta band, and suggests the emergence of a neural mechanism coding for observed goal-directed actions, probably along with improvement in reaching/grasping motor skills. In Experiment 3, we simultaneously recorded EEG scalp activity and multiunit activity from the ventral premotor cortex of two adult macaque monkeys. The specific aim of the study was to correlate multiunit spiking activity recorded from area F5 with EEG changes occurring over the scalp (in alpha and beta bands) during the execution and observation of goal-directed actions, in order to understand the possible link between mirror neurons activity and EEG changes recorded at scalp level. We found that mirror neurons activity in area F5 strongly correlates with the LFP in the high-γ band during both execution and observation of grasping action and that LFP activity in the high-γ band inversely correlates with the EEG desynchronization in the beta band, especially over central electrodes. This correlation is more extended during action execution than observation, suggesting the recruitment of other populations of neurons (i.e. purely motor) that contribute to the desynchronization of the EEG signal. These results confirm our previous findings and suggest that, in monkeys, the EEG beta band suppression recorded over the scalp may be a reliable marker of F5 mirror neurons activity. Moreover, activity in the beta band during action execution and observation may originate from the motor cortex. In fact, we found that LFP activity during action execution in the deeper cortical layers better correlates with the EEG beta desynchronization over the central electrodes, indicating the importance of the cortical motor output in the generation of this signal.