Posture influences motor imagery: An fMRI study

Abstract

Motor imagery is widely used to study cognitive aspects of the neural control of action. However, what is exactly simulated during motor imagery is still a matter of debate. On the one hand, it is conceivable that motor imagery is an embodied cognitive process, involving a simulation of movements of one’s own body. The alternative possibility is that, although motor imagery relies on knowledge of the motor processes, it does not entail an actual motor simulation that is influenced by the physical configuration of one’s own body. Here we discriminate between these two hypotheses, in the context of an established motor imagery task: laterality judgments of rotated hand drawings.

We found that reaction times of hand laterality judgments followed the biomechanical constraints of left or right hand movements. Crucially, the position of subjects’ own left and right arm influenced laterality judgments of left and right hands. In neural terms, hand laterality judgments activated a parieto-frontal network. The activity within this network increased with increasing biomechanical complexity of the imagined hand movements, even when the amount of stimulus rotation was identical. Moreover, activity in the intraparietal sulcus was modulated by subjects’ own hand position: a larger incongruence in orientation between the subjects’ hand and the stimulus hand led to a selective increase in intraparietal activity.

Our results indicate that motor imagery generates motor plans that depend on the current configuration of the limbs. This motor plan is calculated by a parieto-frontal network. Within this network, the posterior parietal cortex appears to incorporate proprioceptive information related to the current position of the body into the motor plan.

Introduction

It is widely held that imagining performing an action and actually performing the action rely on partly overlapping mechanisms. However, the exact nature of the relationship between imagined and executed movements remains controversial. In particular, it is not well understood what is simulated during motor imagery. Is it a simulation of a movement of one’s own body (Jeannerod, 1994) or is it a more abstract implementation of general kinematic rules of biological motion (Viviani and Schneider, 1991, Fitts and Peterson, 1964)? If motor imagery entails a simulation of one’s own body movements, then this simulation should depend not only on the action requirements but also on the configuration of one’s own body in space. This would be an instance of embodied cognition (Gallese, 2003). Conversely, if motor imagery entails a simulation of a third person’s body movement, or merely the selection of an action’s goal, then this simulation would require knowledge of the motor processes involved, but no actual motor simulation. This would be an instance of a cognitive process that is independent from the contingent physical characteristics of the agent.

The issue whether motor imagery is embodied remains controversial. Some studies observed an influence of hand posture on motor imagery (Nico et al., 2004, Shenton et al., 2004, Parsons, 1987, Sirigu and Duhamel, 2001, Vargas et al., 2004). Other studies, using hemiplegic patients, showed that motor simulations can be successfully carried out in absence of the ability to produce these actions (Johnson et al., 2002b, Johnson, 2000). The issue goes well beyond the boundaries of motor imagery: movement simulation is a process at the basis of several models of motor planning (Wolpert and Ghahramani, 2000), action observation (Jeannerod, 2001), and social cognition (Gallese et al., 2004, Jacob and Jeannerod, 2005), and it is therefore relevant to precisely define the characteristics of this process. In this study we test these contrasting predictions, both at the behavioral and at the neural level, in the context of a well-established imagery task: the hand laterality judgment task.

Psychophysical studies suggest that, when subjects observe a hand in a certain orientation and have to decide whether it is a left or a right hand, they solve this task by imagining their own hand moving into the stimulus orientation for comparison (Parsons, 1987, Parsons, 1994, Sekiyama, 1982). These studies found that the time required for a hand laterality judgment is similar to the time taken to execute a corresponding movement. Crucially, the trajectory imagined during laterality judgments of left and right hands is strongly influenced by the biomechanical constraints of actual left-hand and right-hand movements. A series of neuroimaging experiments also suggests a large cognitive and neural overlap between the hand laterality judgment task and execution of actions (de Lange et al., 2005, Rumiati et al., 2001, Sekiyama, 1982, Parsons, 1994, Parsons et al., 1998, Tomasino et al., 2003). On the basis of these findings, the hand laterality task is taken as an instance of motor imagery rather than a purely visual mental rotation task, and it has been extensively used to study cognitive aspects of the action system in a range of motor-related neuropathologies (Schwoebel et al., 2001, Fiorio et al., 2006, Tomasino et al., 2003, Johnson, 2000, de Lange et al., 2004).

In this study, we test for the embodied nature of movement simulation, as operationalized by the hand laterality task, both at the behavioral and at the neural level. First, we used fMRI to identify which regions increase their neural activity with increasing biomechanical complexity of the imagined movement, even when the amount of stimulus rotation is identical. Our approach exploits the fact that rotating right hands in clockwise orientations requires biomechanically more complex movements than counterclockwise rotations, whereas the opposite holds true for left hands (Parsons, 1994). This novel approach (in the context of neuroimaging) allows to identify the network that is specifically involved in imagined hand movements, while avoiding the interpretational and methodological problems that arise when using a secondary task to control for visuospatial processes only loosely related to movement simulation (de Lange et al., 2005, Kosslyn et al., 1998). Second, we assessed to what extent the hand laterality judgment task entails motor simulations of one’s own body, by probing the influence of posture on performance of this task. For this, we manipulated the posture of subjects left and right hand, independently. If motor imagery evokes motor simulations of one’s own body, then the current position of the left and right hand should have an effector-specific effect on behavioral performance. Moreover, this should influence the activity of the neural architecture subserving the integration of postural information in the motor plan.