Perspective-dependent activation of frontoparietal circuits during the observation of a static body effector

Highlights

• The observation of a static forearm evokes perspective-dependent responses.

• Neural processing of third-person stimuli segregates from other viewpoints.

• Motor-premotor regions are activated by all perspectives.

• First-person stimuli evoke earlier motor activity than third-person perspective.

Abstract

The perspective from which body-related stimuli are observed plays a fundamental role in modulating cerebral activity during the processing of others' bodies and actions. Previous research has shown perspective-dependent cerebral responses during the observation of both ongoing actions and static images of an acting body with implied motion information, with an advantage for the egocentric viewpoint. The present high-density EEG study assessed event-related potentials triggered by the presentation of a forearm at rest before reach-to-grasp actions, shown from four different viewpoints. Through a spatiotemporal analysis of the scalp electric field and the localization of cortical generators, our study revealed overall different processing for the third-person perspective relative to other viewpoints, mainly due to a later activation of motor-premotor regions. Since observing a static body effector often precedes action observation, our results integrate previous evidence of perspective-dependent encoding, with cascade implications on the design of neurorehabilitative or motor learning interventions based on action observation.

Introduction

Human bodies and body parts represent a special class of visual stimuli, which play a pivotal role in motor-related processes (Minnebusch and Daum, 2009, Downing and Peelen, 2011). On the one hand, visual cues regarding one's own body are relevant for the guidance of self-actions by tracking online the positions of one's own limbs for visuomotor control. On the other hand, the visual encoding of others' bodies is critical during social perception and interactions for recognizing the actions of other individuals and organizing appropriate imitative or social responses. The main aim of the present EEG study is to evaluate whether and how the perspective from which a static but ready-to-move effector is observed may influence ERP responses and cerebral visuomotor processing of such body-related stimulus.

In this latter regard, body-shape processing is a key step in observing others' actions in concomitance with the processing of ongoing motion cues. During action observation (AO), the visual perception of others' moving effectors activates frontoparietal circuits similar to those engaged during action execution through the so-called “mirror neuron system” (MNS) (Rizzolatti and Sinigaglia, 2010, Rizzolatti and Fogassi, 2014). It has been hypothesized that during AO, static representations of the moving body are encoded in high-level visual regions, such as the extrastriate body area (EBA) in lateral occipitotemporal, and integrated into the superior temporal sulcus with motion cues from middle/medial-superior temporal cortex (Downing et al., 2006, Peelen and Downing, 2007). This body-related perceptual input would provide a complete visual representation of actions to the frontoparietal nodes of the MNS, which in turn convert visual information into a motor format triggering an internal motor simulation of observed actions (Rizzolatti and Sinigaglia, 2010, Rizzolatti and Fogassi, 2014).

Noteworthy, visual information regarding others' bodies recruits the onlooker's motor system even without the actual perception of motion. Indeed, static images of effectors with implied motion are sufficient to trigger visuomotor priming effects in the observers, facilitating overt responses by images of congruent relative to incongruent movements (e.g., Craighero et al., 2002, Vogt et al., 2003). Accordingly, source analysis of scalp event-related potentials (ERPs) has shown stronger activation in regions of the action representation system, including the premotor and motor cortex, for static images of actions with a higher implied motion level than less dynamic actions (Proverbio et al., 2009).

Despite this evidence of motor system recruitment during observation of static images of moving

bodies, whether similar visuomotor priming could also be triggered by the observation of body segments at rest, without any implied motion, remains debated, with mixed results from previous research.

On the one hand, a previous TMS study has shown that mirror corticospinal facilitation was absent in response to images of upper limbs at rest (Urgesi et al. 2006). Accordingly, activations in the motor or premotor cortex are not generally reported in fMRI studies presenting images of still bodies or body segments (for reviews: Peelen and Downing, 2007, Minnebusch and Daum, 2009, De Gelder et al., 2010). On the other hand, behavioral evidence has revealed that focusing attention on a body effector devoid of any implied-motion and task-irrelevant appeared sufficient for eliciting visuomotor priming/interfering effects on observers' responses (Bach et al., 2007, Welsh et al., 2014). It is worth noting that the context in which a body part is perceived plays a relevant role in evoking motor system activity. Indeed, a static body effector per se could prime the observer's motor system in case of a predictable upcoming action, proactively triggering motor representations (Caetano et al., 2007).

Besides the controversies related to visuomotor transformations of an observed static effector, another open issue is the influence of low-level visual features, and in particular of the viewpoint, on cerebral processing of static and moving body-related stimuli.

The perspective from which body-related stimuli are observed plays a fundamental role in modulating cerebral activity during the processing of others' actions. Visual perspectives from which bodies are perceived can be defined according to two types of observer-centered reference frames. The first one is based on a Cartesian coordinate system, defining the angular disparity of the observed body segment relative to a 0° self-centered reference frame. Accordingly, the perspectives on the two main orthogonal axes can be defined as longitudinal (0°, 180°) and lateral (90°, 270°), with all possible intermediate degrees of angular disparity between them. The second type of reference frame is a body-centered coordinates system based on the anatomical compatibility of the observed moving effectors with self-body parts. Accordingly, viewpoints can be defined as egocentric when observed body segments are anatomically consistent with one's own seen body parts and allocentric when visual information is congruent only with anatomical details of another individual's body (Saxe et al., 2006). However, whether and how these two coordinates systems interact to facilitate the visuomotor encoding of an observed body onto self-body and motor representations remain open issues.

Mixed results emerged in previous studies evaluating, with different approaches, the effect of FP vs. TP on cerebral processing of body-related stimuli.

Despite some mixed results (e.g., Frenkel-Toledo et al., 2013, Fitzgibbon et al., 2014), an advantage in visuomotor mapping for egocentric stimuli has been proved for moving effectors during AO. Indeed, EEG (Fu and Franz, 2014, Drew et al., 2015, Angelini et al., 2018), fMRI (e.g., Jackson et al., 2006), and TMS studies (e.g., Maeda et al., 2002) have shown stronger motor resonance induced by AO for egocentric relative to the allocentric viewpoint.

Concerning body images with implied motion and depicting hand gestures, behavioral studies provided mixed results, showing both egocentric (e.g., Bruzzo et al., 2008) and allocentric advantage (e.g., Vogt et al., 2003) in terms of visuomotor priming effects.

Finally, concerning still bodies (Chan et al., 2004) and body segments images (Saxe et al., 2006, Zopf and Williams, 2013), fMRI studies have shown perspective-dependent activations in several brain regions, but the preference for FP or TP is region-dependent, and sometimes inconsistent across studies. For instance, stronger activations were found for FP in the right dorsolateral prefrontal cortex (DLPFC) (Saxe et al., 2006), in anterior intraparietal and occipitotemporal regions (Zopf and Williams, 2013), while TP triggered stronger activation in the right EBA (Chan et al., 2004, Saxe et al., 2006). In general, during the presentation of static body stimuli, motor and premotor regions did not appear to be modulated by the viewpoint. Nevertheless, in a TMS study, corticospinal modulation induced by pain observation occurred only for static body segments shown from FP (Bucchioni et al., 2016).

High-density EEG (hdEEG) brings an added value to fMRI data, as it allows one to obtain electrical neuroimaging findings, yet ensuring a high temporal resolution, in the order of ms. Starting from this premise and the background mentioned above, in the present hdEEG study, we assessed the ERPs triggered by presenting a static forearm and hand from different perspectives and performed a spatiotemporal analysis of the scalp electric fields (Murray et al., 2008). Such analysis returns the time windows of topographic differences among perspectives and allows one to statistically compare the underlying patterns of cortical generators (Michel et al., 2004, Murray et al., 2008). Our experimental stimuli were video clips of reach-to-grasp actions shown from four different viewpoints (i.e., first-person, third-person, lateral egocentric, and lateral allocentric perspectives) (FP, TP, LE, LA, respectively), defined by different combinations of angular disparity (longitudinal and lateral viewpoints) and anatomical compatibility (ego- and allocentric viewpoints). EEG analyses were focused on the initial part of the video clips that presented the effector at rest before the beginning of the action (Fig. 1).

The present study has two principal aims. The first one is to assess viewpoint-related modulations in event-related EEG activities triggered by the visual presentation of resting body segments (forearms and hands), possibly evaluating the interplay between the two factors defining perspectives (i.e., angular disparity and anatomical compatibility). The second one is to test whether the viewpoint from which a static effector is observed, in the context of a fully predictable upcoming action, modulates the activity of motor regions and which perspective determines the strongest motor-related representations.