Elsevier

Cortex

Volume 50, January 2014, Pages 115-124
Cortex

Research report
Observing object lifting errors modulates cortico-spinal excitability and improves object lifting performance

https://doi.org/10.1016/j.cortex.2013.07.004Get rights and content

Abstract

Observing the actions of others has been shown to modulate cortico-spinal excitability and affect behaviour. However, the sensorimotor consequences of observing errors are not well understood. Here, participants watched actors lift identically weighted large and small cubes which typically elicit expectation-based fingertip force errors. One group of participants observed the standard overestimation and underestimation-style errors that characterise early lifts with these cubes (Error video – EV). Another group watched the same actors performing the well-adapted error-free lifts that characterise later, well-practiced lifts with these cubes (No error video – NEV). We then examined actual object lifting performance in the subjects who watched the EV and NEV. Despite having similar cognitive expectations and perceptions of heaviness, the group that watched novice lifters making errors themselves made fewer overestimation-style errors than those who watched the expert lifts. To determine how the observation of errors alters cortico-spinal excitability, we measured motor evoked potentials in separate group of participants while they passively observed these EV and NEV. Here, we noted a novel size-based modulation of cortico-spinal excitability when observing the expert lifts, which was eradicated when watching errors. Together, these findings suggest that individuals' sensorimotor systems are sensitive to the subtle visual differences between observing novice and expert performance.

Introduction

Before picking up an object, individuals will implicitly estimate its weight based on its visual properties, and these expectations of heaviness drive the way that they lift objects. This means that when lifting something for the first time, a lifter's fingertip forces reflect their initial predictions about an object's weight, rather than the actual mass of the object (Gordon, Forssberg, Johansson, & Westling, 1991). The feed-forward nature of human lifting behaviour often results in grip and load force errors, which can be especially dramatic when objects have an unusual weight for their appearance (e.g., Buckingham et al., 2009, Johansson and Westling, 1988). These errors do not generally persist and individuals are rapidly able to overcome their expectations of heaviness, tuning their fingertip forces to the actual, rather than expected, weight of the object(s) being lifted (Flanagan and Beltzner, 2000, Grandy and Westwood, 2006, Mon-Williams and Murray, 2000). In other words, when lifting objects repeatedly, individuals rapidly and implicitly learn to lift them with the appropriate level of grip and load forces for their actual weight.

Despite the widely held assumption that fingertip force adaptation is mediated solely by fast-adapting Type-2 afferents in the fingertips (Johansson & Flanagan, 2009), it has recently been demonstrated that vision plays a crucial role in this form of motor learning. When they are deprived of vision individuals show deficits in their ability to correct their fingertip force errors, continually lifting objects with forces that reflect how heavy the objects look, rather than how heavy the objects actually are (Buckingham and Goodale, 2010a, Buckingham et al., 2011). These findings indicate that individuals receive valuable information describing the direction and magnitude of a lifting error from visual kinematic cues.

Consistent with this proposal, a variety of studies have demonstrated that humans are surprisingly adept at acquiring useful information, such as object weight, from the observed visual kinematics of others' lifts (Bingham, 1987, Hamilton et al., 2007). Not only are individuals able to use these kinematic cues, but there is emerging evidence that the link between acting and perceiving is an automatic one. Hamilton and colleagues (Hamilton, Wolpert, & Frith, 2004) demonstrated that our perception of an actor's lift is modulated by the weight of an object the observer is holding (interestingly, in the opposite direction from what might be expected – holding a light box made the observed lift appear comparatively effortful, and vice versa). Furthermore, individuals implicitly use kinematic cues observed in other lifters when lifting objects which have an unpredictable weight (Meulenbroek, Bosga, Hulstijn, & Miedl, 2007). Perhaps the strongest argument for an automatic link between visual kinematics and action production in the context of object lifting comes from a recent series of action observation studies showing that the sensorimotor system appears to encode the force requirements of an observed lift. Using transcranial magnetic stimulation (TMS) to evoke motor potentials (MEPs) in a passive observation task, Alaerts, Swinnen, and Wenderoth (2009) demonstrated that merely watching a video of someone else lifting a heavy object elicits a larger MEP than is elicited while watching a similar video of a lighter object. Subsequent studies have revealed that this force-related modulation of cortico-spinal excitability was caused by differences between the kinematics of the effortful (heavy objects) and easy (light objects) lifts, rather than semantic or material-based visual cues to object (Alaerts et al., 2010a, Alaerts et al., 2010b; Senot et al., 2011). The effects of observing the actions of others are not limited to the modulation of cortico-spinal excitability. A recent study has shown that the forces involved in lifting can be modulated by observing others, elegantly demonstrating that, compared to viewing an object being lightly touched, watching an actor firmly pinching a target object will increase the gripping force subsequently used to lift that object (Uçar & Wenderoth, 2012).

These studies tend to be interpreted within the broader context of the putative human mirror neuron system (Gallese et al., 2011, Mukamel et al., 2010). The overlapping neuronal populations and cortical regions in human and non-human primates has been taken by some as a mechanism for observational learning, by means of implicit neural simulation of the observed action (Calvo-Merino et al., 2005, Jeannerod, 2001). However, the concept of mirror neurons, as typically discussed, offers no insight into how the sensorimotor system reacts to the observation of the commonplace errors that must drive motor learning. This question needs to be addressed at both the level of behaviour and cortico-spinal excitability. In terms of behaviour, it would presumably be maladaptive for the sensorimotor system to copy the motor output of an observed error. Although very few empirical studies have examined the consequences of error observation in any context whatsoever, some recent hints have emerged that individuals can improve their subsequent performance by observing errors. In Mattar and Gribble (2005), participants reached in a velocity-dependant force field toward a visual target – a task which normally requires a substantial amount of learning. They noted that after observing videos of others performing an aiming task, participants performing the same task learnt to overcome the dynamics of the force-field more rapidly. Furthermore, observing a different force-field from the one they eventually had to deal with substantially slowed their rate of adaptation, hinting at an automatic observational learning effect (see also Brown, Wilson, & Gribble, 2009). Crucially, Brown and colleagues parametrically varied the degree of error in these videos, noting that participants were able to benefit more from observing larger errors than smaller ones (Brown, Wilson, Obhi, & Gribble, 2010). This finding was, of course, not an unexpected result given that the correct performance in the task relied exclusively on vision, and the errors provide the only visual indications of the situational dynamics. This work does, however, provide some preliminary hints that there may be a specific and important role for error observation in subsequent behavioural outputs, leading us to predict that observing lifting errors will improve subsequent lifting performance more than observing well-practiced lifts. The role of errors in driving cortico-spinal excitability is less clear, with no work examining MEPs during the observation of motor errors. As there are indications that observing errors may help improve subsequent performance, it is possible that the errors are encoded by the sensorimotor system to drive the subsequent corrective behavioural response. If the low-level motor resonance within the sensorimotor system slavishly mimics what is observed, such a mirroring response would manifest as a large MEP for an overestimation of force and a small MEP for an underestimation of force. However, as errors appear to drive improved behaviour (i.e., in directional opposition to the initial error), the MEPs might in fact oppose the pattern of resonance normally evoked by lifting forces – a large MEP to counteract an erroneous underestimation of force and a small MEP to counteract an erroneous overestimation of force.

The goal of the current work was to examine the consequences of observing errors, within the simple motor learning framework of fingertip force adaptation during object lifting. To this end, we examined the sensorimotor consequences of watching the visual consequences of overestimations and underestimations of lifting forces (the error video – EV) as compared to well-adapted object lifting performance (the no error video – NEV) at the behavioural and cortico-spinal level. If errors are in fact crucial cues for observational learning, it is likely that observing them will (1) improve fingertip force adaptation and (2) modulate cortico-spinal excitability in a way that is specific to the overestimation or underestimation nature of the error.

Section snippets

Video stimuli

A 66 cm screen monitor at a resolution of 1024 × 768 was used to display a short video to participants, depicting five different actors [3 male, 2 female, mean age = 24.6 years ± (SD) .9] repeatedly lifting a small cube (5 cm × 5 cm × 5 cm) and large cube (10 cm × 10 cm × 10 cm) in alternation. Unbeknownst to the participants (or the actors in the videos), the cubes had been adjusted to have identical weights (700 g). These stimuli typically elicit the size-weight illusion, along with a

Behavioural experiment – lifting after observing others' lifts

After watching the video, but prior to actually lifting the cubes, participants verbally reported that they expected the large cube to weigh more than the small cube (Table 1). This expectation did not differ between the EV and NEV groups (Table 1), indicating that participants gained no conscious awareness of the cubes' identical weights from the videos alone. After the lifting portion of the experiment was completed, participants experienced a robust size-weight illusion, reporting that the

Acknowledgements

The authors would like to thank J. Ladich for stimulus construction as well as H. Yang for technical support. We would also like to thank J. Paciocco for assistance with Brainsight as well as L. Strother for helpful comments on an earlier draft of the manuscript. Finally, we would like to thank the actors who featured in the video stimuli. G. Buckingham was supported with a Banting Postdoctoral Fellowship, awarded by the Natural Sciences and Engineering Council of Canada (NSERC). The funders

References (36)

  • K. Alaerts et al.

    Force requirements of observed object lifting are encoded by the observer's motor system: a TMS study

    European Journal of Neuroscience

    (2010)
  • H.F. Bagce et al.

    Corticospinal excitability is enhanced after visuomotor adaptation and depends on learning rather than performance or error

    Journal of Neurophysiology

    (2013)
  • G.P. Bingham

    Kinematic form and scaling: further investigations on the visual perception of lifted weight

    Journal of Experimental Psychology. Human Perception and Performance

    (1987)
  • L.E. Brown et al.

    Repetitive transcranial magnetic stimulation to the primary motor cortex interferes with motor learning by observing

    Journal of Cognitive Neuroscience

    (2009)
  • L.E. Brown et al.

    Effect of trial order and error magnitude on motor learning by observing

    Journal of Neurophysiology

    (2010)
  • G. Buckingham et al.

    Living in a material world: how visual cues to material properties affect the way that we lift objects and perceive their weight

    Journal of Neurophysiology

    (2009)
  • G. Buckingham et al.

    Lifting without seeing: the role of vision in perceiving and acting upon the size weight illusion

    PLoS ONE

    (2010)
  • G. Buckingham et al.

    The influence of competing perceptual and motor priors in the context of the size–weight illusion

    Experimental Brain Research

    (2010)
  • Cited by (33)

    • The effects of explicit and implicit information on modulation of corticospinal excitability during hand-object interactions

      2022, Neuropsychologia
      Citation Excerpt :

      We argued that, if we wanted to investigate how predictive force scaling would be altered by the hidden series, a direct comparison between same weighted objects in different conditions should be made. This is the typical approach we used before (Rens and Davare, 2019; Rens et al., 2020; Van Polanen et al., 2020) and which has been used by other groups in our field as well (Alaerts et al., 2010; Baugh et al., 2012; Buckingham et al., 2014; Cashaback et al., 2017; Gordon et al., 1993; Johansson and Westling, 1993; Loh et al., 2010; Reichelt et al., 2013; Senot et al., 2011). As such, it is unclear whether the results of Mawase and Karniel (2010) reflect (i) an object lifting strategy driven by implicitly learning the hidden series or (ii) solely sensorimotor effects that are entirely driven by last lifted object.

    • Action observation facilitates anticipatory control of grasp for object mass but not weight distribution

      2022, Neuroscience Letters
      Citation Excerpt :

      As such, the larger the error (roll) observed, the more information it provides about weight distribution that can be used to facilitate performance. This is in line with other studies showing the effect of error on observation [4,26]. An important point to note is that even though roll was negatively correlated between the performer and observer, no other measure achieved significance.

    • What modulates the Mirror Neuron System during action observation?: Multiple factors involving the action, the actor, the observer, the relationship between actor and observer, and the context

      2021, Progress in Neurobiology
      Citation Excerpt :

      But they are at odds with Gueugneau et al.’s (2015) and Janssen et al.’s (2015) finding that when such deviations are detected, motor resonance is not abolished but rather continues to reflect the perceived kinematics. These contradictions, among others, suggest that expectations sometimes do, and sometimes don't, modulate the MNS during action observation (for other relevant studies see Obhi and Hogeveen, 2010; Buckingham et al., 2014; Ondobaka et al., 2015; Jacquet et al., 2016; see also the comments about expectations in the subsections on Factors 7 and 13). Finally, some additional insights come from two TMS studies that investigated MNS activity while subjects watched agents lift either heavy or light objects in either a genuine or a deceptive manner (Tidoni et al., 2013; Finisguerra et al., 2018).

    • The effect of modeled absolute timing variability and relative timing variability on observational learning

      2017, Acta Psychologica
      Citation Excerpt :

      Incidentally, the positive impact of observing errors has called into question the straight one-to-one subthreshold activation of motor neurons during action-observation as a complete explanation for the observational learning phenomenon (e.g., Buccino et al., 2001; Cross et al., 2009; Higuchi et al., 2012). Indeed, a recent study from Buckingham and colleagues (Buckingham, Wong, Tang, Gribble, & Goodale, 2014) has shown that the corticospinal excitability elicited during the observation of variable motor errors was modulated by the subsequent learning or parameterization of required forces rather than the observed movement kinematics. That is, the observation of motor errors, as indicated by greater grip force rates for large- compared to small-sized objects that were the same weight, resulted in comparatively similar corticospinal responses during cortical stimulation.

    View all citing articles on Scopus
    View full text