Weight dependent modulation of motor resonance induced by weight estimation during observation of partially occluded lifting actions

Highlights

• Participants perform a weight estimation task while MEPs are recorded.

• Videos show a hand lifting a box of three different weights hidden behind a screen.

• MEPs are modulated with weight in both observed and unobserved muscles.

• Motor resonance is triggered by direct observation.

• Motor resonance extends to cortical representation of muscles that are unobserved.

Abstract

Seeing others performing an action induces the observers' motor cortex to “resonate” with the observed action. Transcranial magnetic stimulation (TMS) studies suggest that such motor resonance reflects the encoding of various motor features of the observed action, including the apparent motor effort. However, it is unclear whether such encoding requires direct observation or whether force requirements can be inferred when the moving body part is partially occluded. To address this issue, we presented participants with videos of a right hand lifting a box of three different weights and asked them to estimate its weight. During each trial we delivered one transcranial magnetic stimulation (TMS) pulse over the left primary motor cortex of the observer and recorded the motor evoked potentials (MEPs) from three muscles of the right hand (first dorsal interosseous, FDI, abductor digiti minimi, ADM, and brachioradialis, BR). Importantly, because the hand shown in the videos was hidden behind a screen, only the contractions in the actor's BR muscle under the bare skin were observable during the entire videos, while the contractions in the actor's FDI and ADM muscles were hidden during the grasp and actual lift. The amplitudes of the MEPs recorded from the BR (observable) and FDI (hidden) muscle increased with the weight of the box. These findings indicate that the modulation of motor excitability induced by action observation extends to the cortical representation of muscles with contractions that could not be observed. Thus, motor resonance appears to reflect force requirements of observed lifting actions even when the moving body part is occluded from view.

Introduction

When observing a box being lifted by somebody, most of the time, people can easily estimate its weight. The mere observation of another person acting has been shown to activate several brain areas that are also engaged during action execution (Caspers et al., 2010). Additionally it has been shown that observing actions modulates the excitability of the primary motor cortex (M1) of the observer, i.e. that it induces “motor resonance in M1” (Fadiga et al., 2005).

Fadiga et al. (1995) were the first to use single pulse transcranial magnetic stimulation (TMS) to assess the excitability of M1 during action perception. They found that seeing others' actions increased the amplitude of TMS-induced motor-evoked potentials (MEPs) and that this increase in corticospinal excitability was specific to the muscles used to perform the observed actions. Since the work of Fadiga et al. (1995), the facilitation of M1 during action observation has been replicated numerous times (Aziz-Zadeh et al., 2002, Candidi et al., 2010, Fadiga et al., 2005, Roosink and Zijdewind, 2010, Sartori et al., 2012, Schütz-Bosbach et al., 2009, Strafella and Paus, 2000). This motor facilitation appears to (1) be present for transitive actions (Fadiga et al., 1995, Sartori et al., 2012) and intransitive movements (Borroni et al., 2005, Burgess et al., 2013, Fadiga et al., 1995, Romani et al., 2005); (2) be temporally coupled with the phases of the observed actions (Alaerts et al., 2012, Borroni et al., 2005, Gangitano et al., 2002, Gangitano et al., 2001, Urgesi et al., 2010); (3) depend on muscular involvement rather than direction features of observed movements (Alaerts et al., 2009, Urgesi et al., 2006a); (4) be causally linked to signals from the same premotor and parietal regions that are involved in action performance (Avenanti et al., 2007, Avenanti et al., 2013, Catmur et al., 2011, Koch et al., 2010). Remarkably, such fronto-parietal regions correspond to the regions where mirror neurons were first discovered in the monkey brain (di Pellegrino et al., 1992, Fogassi et al., 2005, Gallese et al., 1996) and where blood-oxygen-level-dependent (BOLD) signal is increased during action observation (Caspers et al., 2010).

These findings have supported the assumption that motor resonance is dependent on activity in the mirror neurons (Fadiga et al., 1995) and similar somatotopical organization of motor resonance has also been found in premotor and parietal brain regions using functional magnetic resonance imaging (fMRI). Buccino et al. (2001) showed that the activations in the premotor and parietal regions corresponded to the effector used in the observed actions. Their results show that when participants observed actions performed with the mouth, hand or foot, different parts of frontal and parietal cortex were activated. Brain stimulation studies indicate that activity in parietal and premotor regions is necessary for action perception (Avenanti et al., 2013). In particular, Pobric and Hamilton (2006) first used repetitive transcranial magnetic stimulation (rTMS) to investigate whether the motor system plays a crucial role in weight estimation. They found that when participants watched a hand lifting a box and were instructed to estimate its weight, repetitive rTMS, delivered to the inferior frontal gyrus (IFG) disrupted their performance. In contrast, rTMS to the occipital cortex did not affect performance, and rTMS over the frontal or the occipital cortex did not affect performance when people had to judge the weight of a bouncing ball. This observation implies that motor system activation is crucial for weight estimation when a human hand is lifting the object and not for weight estimation of objects as such.

Alaerts et al. (2010) have shown that when watching somebody lift an object M1 excitability is proportionally modulated by the weight of the object being lifted. They recorded MEPs from the first dorsal interosseous (FDI) muscle while participants passively observed an actor lifting two different objects of different weights using a precision grip. Results showed that the amplitude of the MEPs was modulated in accordance with the weight of the objects being lifted. In a complementary experiment, Alaerts et al. (2010) compared the amplitudes of the MEPs measured from the opponens pollicis (OP, thumb), flexor carpi radialis (FCR, wrist) and extensor carpi radialis (ECR, wrist) muscles while participants passively observed videos of a hand lifting an empty, half-full or full bottle. The amplitude of the MEPs measured from the OP and ECR muscles was higher when the observed videos showed a hand lifting a transparent half-full or full (heavier) bottle as compared with an empty one. Additional studies demonstrated that the weight-dependent modulation of motor resonance in M1 persisted when the agent lifted objects that were visually identical but had different weights (Alaerts et al., 2010, Senot et al., 2012, Tidoni et al., 2013). Weight dependent motor resonance in the observers' M1 was observed even when only kinematic information (e.g. trajectories, speed, acceleration) associated to lifting light and heavy objects was available. In some experiments, weight related information conveyed by muscle contraction and local skin tone changes associated to grasping and moving objects with different weights were minimized either using digital movie editing (Tidoni et al., 2013) or by asking the moving agent to wear a glove (Alaerts et al., 2010, Alaerts et al., 2012). However, in these studies, the moving hand and arm were entirely visible for the observers.

In the monkey brain, a significant proportion of mirror neurons in the premotor cortex has been shown to fire also when the object is occluded from view (Umiltá et al., 2001). Umiltá et al. (2001) have shown that “grasping” mirror neurons in the ventral premotor cortex (vPM) of the monkey fire both during the observation of a hand reaching for and grasping an object in full sight, and of a hand reaching behind an occluding screen to grasp a hidden object. Recently, Villiger et al. (2011) have used TMS to investigate motor resonance in M1 using a similar occluding paradigm. The authors presented participants with videos of actual or mimed grasping movements of visible or hidden objects and measured the MEPs from the grasping muscles of the observers. Results showed that MEPs are modulated by object presence. Thus, even when the object was hidden behind a screen, but participants were aware of its existence, MEPs were larger compared to when the object was visible. This effect was pronounced during the grasping phase, but not when seeing the hand at rest. We therefore hypothesize that, motor resonance in M1 will be present when the final phase of the action is occluded from view and that the amount of motor resonance in M1 measured from muscles not directly observed by the participant will be proportional to the inferred force requirements of the action. In the present study we therefore asked participants to estimate the relative weight of a box being lifted. In the videos, only the hand and arm could be observed during the reaching phase, while the grasp and lift of the object were hidden behind a screen and only the forearm proximal to the wrist was visible. This approach allowed us to determine whether modulation of the MEPs according to the weight of the box being lifted occurs only in the muscles in which contractions can be observed by the participants (arm muscles) or also in the muscles involved in the action but that are hidden behind a screen during the actual lift. In the latter (hand) muscles contractions are not directly observable. The distinction between observable and not-observable (occluded from view) muscles allows to explore whether motor resonance in M1 is triggered only by the observation (detectable only for the observable forearm muscle) or whether it can be detected for the hand muscles for which involvement in the action can only be inferred based on the motion and muscle contraction in the visible part of the arm.