Motor imagery of complex everyday movements. An fMRI study

Abstract

The present study aimed to investigate the functional neuroanatomical correlates of motor imagery (MI) of complex everyday movements (also called everyday tasks or functional tasks). 15 participants imagined two different types of everyday movements, movements confined to the upper extremities (UE; e.g., eating a meal) and movements involving the whole body (WB; e.g., swimming), during fMRI scanning. Results showed that both movement types activated the lateral and medial premotor cortices bilaterally, the left parietal cortex, and the right basal ganglia. Direct comparison of WB and UE movements further revealed a homuncular organization in the primary sensorimotor cortices (SMC), with UE movements represented in inferior parts of the SMC and WB movements in superior and medial parts. These results demonstrate that MI of everyday movements drives a cortical network comparable to the one described for more simple movements such as finger opposition. The findings further are in accordance with the suggestion that motor imagery-based mental practice is effective because it activates a comparable cortical network as overt training. Since most people are familiar with everyday movements and therefore a practice of the movement prior to scanning is not necessarily required, the current paradigm seems particularly appealing for clinical research and application focusing on patients with low or no residual motor abilities.

Introduction

Motor imagery (MI) is defined as internal rehearsal of a movement without any overt physical movement (Crammond, 1997, Jeannerod, 1994). As such, MI is the fundamental basis of motor imagery-based mental practice (MP), which is defined as the repeated imagination of movements by using MI. A key finding of research in this area is that MP can result in improvements of motor performance, despite the absence of any overt movement (Driskell et al., 1994, Feltz and Landers, 1983). Accordingly, MP is frequently employed by athletes and sportsmen to accompany standard training procedures. Recognizing the potential of MP, this method has recently gained interest by clinical researchers and practitioners as a potential rehabilitation technique to improve motor performance in patients with movement disorders (Crosbie et al., 2004, Dickstein et al., 2004, Dijkerman et al., 2004, Jackson et al., 2001, Jackson et al., 2004, Johnson-Frey, 2004, Kimberley et al., 2006, Malouin et al., 2004, Sharma et al., 2006, Stevens and Stoykov, 2003).

To optimize rehabilitation and training strategies based on mental practice, an understanding of the functional neuroanatomical correlates of motor imagery would be highly beneficial (cf., Lacourse et al., 2004). However, a key characteristic of MP in training situations is that complex sequences of everyday movements are imagined, while previous evidence is restricted mainly to very basic and simple movements, such as finger/foot flexion extension or finger opposition. Taking the research on simple movements as a starting point, it has been shown that the imagery of a movement activates largely the same cortical motor areas as compared to the preparation (Jeannerod, 1994, Kosslyn et al., 2001) or even overt performance (Lotze et al., 1999, Porro et al., 1996) of that movement. Thus, from a functional neuroanatomical point of view, MI can be conceptualized as an “active” performance of the movements imagined, in the way that – although no overt movement is performed – the activity induced in the associated brain areas resembles the activity during active performance (Johnson et al., 2002). Thus, this theory suggests equivalent activations in MP and active performance, and consequently predicts functional changes in motor system organization for MP comparable to the changes described for overt training (Hlustik et al., 2004, Karni et al., 1995, Lacourse et al., 2004). Such functional changes may provide an explanation for the performance increments gained by MP. However, although this mechanism is quite appealing to account for the effectiveness of MP in improving motor performance in patients and athletes, it has not been verified yet. Such a verification would have to rely on movements which are actually used in the application of MP, i.e., everyday movements (also called everyday tasks or functional tasks). Because MP consists of repetitive application of motor imagery, the suggested mechanism can only be valid if motor imagery of everyday movements is shown to rely on the cortical motor system.

Based on studies using simple movements, it could be hypothesized that everyday movements will recruit large parts of the motor system. In particular, premotor cortices should be activated as these have been reported to be involved in MI by virtually all previous studies. However, contrary to these findings a previous study investigating everyday movements reported virtually no activation of lateral or medial premotor cortices during MI of stance, walking, and running (Jahn et al., 2004). This initial evidence on MI of everyday movements casts severe doubts on the proposed mechanism of MP efficacy and the development of theoretical considerations driving MP-based interventions.

Thus, the first aim of this study was to show that MI of everyday movements relies on the cortical (pre)motor system. Based on previous evidence derived from simple movements and on more theoretical accounts of MI (Jeannerod, 1994), we expected MI of everyday movements to activate mainly lateral and medial premotor cortices. However, if the finding of Jahn et al. (2004) holds true for all everyday movements, lateral premotor cortices may not be involved in MI of the currently used movements as well.

The complexity of everyday movements imposes a number of challenges for their investigation. For example, in MI studies on simple finger movements participants typically practice the task prior to scanning to ensure a comparable level of movement familiarity across participants. For everyday movements, such as swimming, this is not feasible and a comparable level of familiarity can therefore not be ensured. In the present experiment, we dealt with this problem by including a wide variety of different everyday movements, following the rationale that this should balance the effects of familiarity and hence ensure that a roughly comparable level of familiarity is achieved across movements and participants.

We further reasoned that a wide variety of movements would enhance the ecological validity of the current approach, not in the least because MP training employs a range of different movements. We therefore presumed that the current data were more likely to resemble the real-world application of MP if a wide variety of movements was employed and, therefore, result in knowledge which actually has the potential to facilitate the optimization of rehabilitation and training methods.

The advantages of using a variety of movements are countered by new methodological hurdles. Most critical here is the question of whether different everyday movements result in a comparable activation pattern or in very different ones. In other words, the specificity of the cortical activation patterns of different everyday movements is unknown. On the one hand, it is plausible to assume that most everyday movements involve so many different muscles that the cortical areas involved in MI overlap considerably (cf., also Schieber and Hibbard, 1993). In that case, specificity can be assumed to be low, and consequently the pooling of different movements may have no effect on the detection and identification of cortical areas related to MI. On the other hand, cortical activation during MI of different movements has been shown to map onto the homuncular organization in the sensorimotor system (Ehrsson et al., 2003, Stippich et al., 2002). This suggests that such different movements as swimming and eating a meal may activate very specific and distinct cortical areas. In that case, pooling such movements is likely to decrease statistical power considerably and may even prevent identification of MI-related cortical areas.

Based on these considerations, a further aim of the present study was to characterize the activation specificity of everyday movements. This question is not only of theoretical interest, but also of practical relevance because rather similar activation patterns would allow future studies to pool different movements, while different patterns suggest that different movements should be treated separately. To answer this question, we included two MI conditions which differed with respect to the limbs involved in the imagined movements, and tested whether these two conditions show differential cortical activation patterns. In essence, we employed movements confined to the upper extremities (e.g., eating a meal) and movements involving the whole body (e.g., swimming).