Abstract.

Motor imagery corresponds to a subliminal activation of the motor system, a system which appears to be involved, not only in producing movements, but also in imagining actions, recognizing tools, learning by observation or even understanding the behavior of other people. The use of techniques for mapping brain activity and probing cortical excitability, as well as observation of brain lesioned patients during imaging tasks, provide new insights into the covert aspects of motor activity.

Introduction.
Motor imagery is now becoming a hot topic in the field of cognitive neuroscience. A major conceptual advance in the last three or four years has been to decouple the mental phenomenon of simulating an action from the conscious representation of that action. New methodologies are now being developed, where subjects have to go through the process of mental simulation before they can give a response on the feasibility of a movement. Objective cues, like pattern of responses or response time can then be correlated with neural events observed during this mental activity.

The two sides of motor imagery: explicit and implicit motor images.
Whereas the term of motor image classically refers to explicit or conscious representation of an action (imagine yourself running or raising your hand), the same concept also includes other, implicit or unconscious, aspects of the same phenomenon. One example of an implicit motor imagery is provided by an unpublished experiment by Frak, Paulignan and Jeannerod : subjects were shown a glass of water from above with an indication of where the thumb and the index finger should contact it. Subjects, without performing the action, had to judge (by pressing on different keys) whether the action of raising the glass and pouring the water in another container would be easy, difficult, or impossible for each set of finger positions. Their pattern of responses followed the limitations that the geometry of the upper limb would have imposed on real motor performance, which implies that, although they received no instruction to do so, they unconsciously simulated the movement before giving the response. Their response times (within 1500-2000 ms) increased with the estimated difficulty of the task (Figure 1). This result is in line with other observations showing that recognition of the handedness of a visually presented hand depends on covert sensorimotor processes (1,2**,3,4,5). As a rule, these experiments tend to converge on a common finding, namely that the time to give the response reflects the degree of mental rotation needed to bring one’s hand in a position adequate for achieving the task. This implicit process is the motor counterpart of the classical mental rotations or displacements used for giving responses about visual objects. However, unlike 3-D shapes, which can be rotated at the same rate in any direction, rotation of one's hand is limited by biomechanics of the arm joints (6). According to Parsons and Fox (7), response times thus reflect biomechanically compatible trajectories, at the same rate as for executed movements. Parsons et al (2**) further showed that, in the case of handedness recognition, these sensorimotor processes are also constrained by the neural structures controlling the side of the hand to be recognized: In two split-brain subjects, they found that handedness recognition was almost impossible when an image of a right hand was visually presented to the right hemisphere, whereas recognition was normal when it was presented to  the left hemisphere, which controls the movements of the right hand. Interestingly, in normal subjects, response times also reflect a better recognition of handedness by the hemisphere contralateral to the presented hand (3)
 Similar effects on response times have been observed for making judgements on how to use hand held objects or tools : according to Tucker and Ellis (8), such objects automatically potentiate the actions they afford. Our implicit knowledge about actions influences the way we cognitively process the visual world (9). Finally, another form of implicit mental operation has been recently added to the realm of motor imagery: this is the cognitive process related to recognizing and understanding actions observed from other individuals. It has often been suggested that an action can only be understood to the extent that it can performed by the observer (10,11), the general idea of "knowing by simulated doing” (12). Thus, observing an action would activate within the observer the same mechanisms that would be activated were that action intended or imagined by the observer. In turn, this implicit representation in the brain of how movements are generated influences interpretation of observed actions. Hence the benefits of observing actions of others for understanding their behavior and for learning new motor skills (observational learning) .

How similar are a simulated and an executed action?
This new approach to motor imagery, which focuses on the vehicle (the brain mechanisms involved) rather than on the content of motor images, was critical for getting out of experiments where the only data were based on subjective reports. If a motor image bears any relationship to the action it simulates, then properties pertaining to the action should be expressed in the image. This similarity is clearly illustrated by results obtained using the mental chronometry paradigm. It is known, for example, that simulated actions take the same time as really executed ones (for review of earlier work, see 13). Sirigu et al (14**) have reached the same conclusion by instructing subjects to mentally move their index finger between two imaginary targets. Mental movement was paced by a metronome beating at an increasing rate. Subjects had to indicate when they could no longer follow the metronome rate with their mental movement. The metronome frequency at which this occurred was very close from the break frequency for actual movements. In addition, the break frequency was a function of the size of the imaginary targets between which the subject had to move, hence replicating the classical speed-accuracy tradeoff (the so-called Fitts law) observed with real alternating movements. This result suggests that simulated movements follow regularities which are known to rule motor behavior (see also 15).
 Motor simulation thus relies, at least partly, on mechanisms common with motor execution. Positive evidence concerning this point arises from three main experimental sources which will be detailed in this and the next sections, i.e., studies based on brain metabolism, on brain lesioned patients and on changes in brain excitability. Earlier brain mapping experiments using PET had partly answered the question of activation of cortical and subcortical motor structures during motor imagery. Indeed, these experiments were most useful in demonstrating the existence of a consistent cortical network involved in the generation of motor imagery (for a review of work prior to 1995, see 16). This network involves structures directly concerned with motor execution, such as premotor cortex, lateral cerebellum, basal ganglia ; it also involves areas concerned with action planning, such as dorsolateral prefrontal cortex, inferior frontal cortex and posterior parietal cortex (7,17*,18,19**). Comparison with motor execution reveals that the cortical zones activated during imagery only partly overlap with execution zones. In SMA, for example, motor imagery involves pre-SMA (19**,20), a region anterior to SMA proper to which activation during execution is limited. A similar dissociation also exists in parietal cortex (20). Finally, inferior fontal cortex, which is activated during mental simulation and suppressed during execution has been suggested to be the site for motor inhibition during mental simulation (19**).
 The effects of brain lesions are also good indicators of the role of some of these brain sites in controlling motor imagery. Basically, the results show that conditions affecting the motor system leave intact the ability to generate motor imagery. Sirigu et al (21) showed that a patient with hemiparesis due to cortical degeneration limited to primary motor cortex was still able to generate motor imagery with his affected hand, although the mentally simulated movements were slowed down to the same extent as executed movements. In both conditions, however, the speed accuracy tradeoff was preserved (i.e., movements were faster for easier tasks). Similarly, mentally simulated movements have been shown to be slowed down in Parkinson patients in the same way as executed movements (22). Parietal lesions, by contrast, seem to alter the ability to generate imagery. In a patient with a unilateral posterior parietal lesion, Sirigu et al (14**) observed a dissociation between execution and simulation of tapping movements with the contralesional arm. Executed movements, although slower than with the ipsilesional arm, retained the speed accuracy tradeoff, whereas the simulated movements did not.

Motor imagery as a subliminal activation of the motor system.
The main debate on neural mechanisms of motor imagery now focuses on the degree of involvement of motor pathways, and particularly, primary motor cortex. The most recent studies diverge in their results whether they used PET or fMRI. The extensive PET study by Deiber et al (19**) failed to find a significant activation of primary motor cortex and lateral cerebellum during motor imagery of finger movements. fMRI studies, by contrast, unambiguously demonstrate that pixels activated during contraction of a muscle (an intrinsic hand muscle, for example) are also activated during imagery of a movement involving the same muscle (23*). Porro et al (24**) carefully demonstrated this point in comparing fMRI signal intensity during a control task (visual imagery) and two “motor” tasks, motor performance and imagery of repetitive finger opposition movements. fMRI signals were increased during both motor tasks at the presumed site of primary motor cortex, in the posterior portion of precentral gyrus. Porro et al were also able to determine that pixels activated during both motor performance and motor imagery represent a large fraction of the whole population of pixels activated during motor performance (note that the question of whether these pixels contain corticospinal cells cannot be answered with such methods). It remains that primary motor cortex activation reported during motor imagery amounts only about 30% of the level observed during execution (23*,24**). For an indirect confirmation of these results using quantified electroencephalography and neuromagnetic techniques, see 25,26, 27.
 The role of primary motor pathways in motor imagery can also be analyzed using a more direct measurement of corticospinal excitability. Transcranial magnetic stimulation (TMS) of motor cortex was used to trigger motor evoked potentials (MEPs) in arm muscles during simulated arm movements. MEPs were found to be increased, but only in that or those muscles involved in the imagined movements. Accordingly, MEPs were selectively increased in a wrist flexor when the subject mentally simulated wrist flexion, whereas MEPs in the antagonist extensor muscle remained unchanged. In addition, other types of imagery (e.g., visual) did not affect MEPs in any of the recorded muscles (28*,29*,30) (Figure 2). A logical consequence of increased motor cortex excitability is that it should propagate down to the motoneuron level. This is still a controversial, issue, however. Bonnet et al (31) found increased spinal reflexes during a mentally simulated isometric foot pressure. Increase was more marked for the limb used for pressing than for the contralateral limb. In addition, T-reflexes were more increased than H-reflexes. By contrast, Hashimoto and Rothwell (29*) found no significant increase in upper limb H-reflexes during simulated wrist movements (see also 32). In neither study a significant increase in EMG activity was observed in the involved muscles during mental simulation. The discrepancy between these results may only be apparent, however, due to the broad difference in the situations where spinal reflexes were tested : changes in excitability should be less marked at the upper limb level during a wrist movement than at the lower limb level during a postural activity like foot pressure.

Conclusion.
Future research on motor imagery should follow two main directions. First, it will be important to determine the exact nature of the subliminal activation of the motor pathways involved in this process, and, more specifically, to determine whether it actually corresponds to an “endogenous” activation of motor structures. This will require a complete description of the state of the motor system, which seems difficult with the presently available techniques : for example, standard EMG recordings may miss the activity of deep muscular fibers ; the degree of activity of muscle spindle afferents remains unknown, etc (33). This approach should bridge the gap between the study of cognitive phenomena such as the covert states of action generation, thought to be possible only in man, and the detailed study of the underlying neural mechanisms, accessible only in animals. Recent experimental results in monkeys indicate that this possibility is at hand. Neuron discharges in both parietal and premotor cortices have been shown to map the pattern of an action, even when that action will not be produced by the animal (see 34,35).
 Second, at a more applied level, motor imagery should reveal a potent tool for probing and possibly improving the functionning of the motor system. The fact that the motor pathways are globally activated during motor imagery represents a rationale for rehearsing effects observed during motor learning (36,37) and opens new possibilities for rehabilitating patients with motor impairments.
 

Aknowledgements

Work supported by the Biomed 2 Program.

References and recommended reading

1 Parsons LM, Fox PT, Downs JH, Glass T, Hirsch TB, Martin CC, Jerabek PA, Lancaster JL: Use of implicit motor imagery for visual shape discrimination as revealed by PET. Nature 1995, 375:54-58.

2** Parsons LM, Gabrieli JDE, Phelps EA, Gazzaniga MS: Cerebrally lateralized mental representations of hand shape and movement. J. Neurosci 1998, 18:6539-6548.
In this paper, two split brain subjects had to make handedness judgements for drawings representing right or left hand shapes shown at different orientations, and presented to one visual hemifield (and thus to one hemisphere). Response accuracy was good when the side of the hand shown to the hemisphere corresponded to the side of the hand controlled by that hemisphere, but was severely degraded for the opposite combination. Hence the authors conclusion that mentally simulating one's action and recognizing handedness both depend on lateralized mechanisms.

3 Johnson SH: Cerebral organization of motor imagery. Controlateral control of grip selection in mentally represented prehension. Psychol Sci 1998, 9:219-222.

4 Gentilucci M, Daprati E, Gangitano M: Implicit visual analysis in handedness recognition. Consciousness and Cognition 1998

5 Wohlschläger A, Wohlschläger A: Mental and manual rotation. J Exper Psychol Hum Perc Perf 1998, 24:397-412.

6 Wexler M, Kosslyn SM, Betrthoz A: Motor processes in mental rotation. Cognition 1998, 68:77-94.

7 Parsons LM, Fox PT: The neural basis of implicit movements used in recognising hand shape. Cog Neuropsychol 1998, 15:583-615.

8 Tucker M, Ellis R: On the relations between seen objects and components of potential actions. J Exper Psycol Hum Perc Perf 1998, 24:830-846.

9 de'Sperati C, Stucchi N: Recognizing the motion of a graspable object is guided by handedness. Neuroreport 1997, 8:2761-2765.

10 Jeannerod M: To act or not to act. Perspectives on representation of actions. Quart J Exper Psychol 1999, 52A:1-29.

11 Gallese V, Goldman A: Mirror neurons and the simulation theory of mind-reading. Trends Cog Sci 1998, 12:493-501.

12 Schwartz DL, Black T: Inferences through imagined actions. Knowing by simulated doing. J Exper Psychol Hum Perc Perf 1999, 25:116-136.

13 Jeannerod M: Mental imagery in the motor context. Neuropsychologia 1995, 33:1419-1432.

14** Sirigu A, Duhamel J-R, Cohen L, Pillon B, Dubois B, Agid Y: The mental representation of hand movements after parietal cortex damage. Science 1996, 273:1564-156.
The contribution of this paper to the problem of motor imagery is twofold. First, it introduces an objective method for assessing the timing of motor imagery, based on the speed-accuracy tradeoff. Second, it reports on motor imagery abilities in several patients with focal lesions in parietal and motor cortices. Further discussion on these results is available in Reference 34.

15 Decety J, Jeannerod M: Mentally simulated movements in virtual reality. Does Fitts law hold in motor imagery ? Behav Brain Res 1996, 72:127-134.

16 Jeannerod M, Decety J: Mental motor imagery : A window into the representational stages of action. Current Opinion in Neurolobiology 1995 5:727-732.

17* Grafton ST, Arbib MA, Fadiga L, Rizzolatti G: Localization of grasp representations in humans by positron emission tomography. 2. Observation compared with imagination. Exper Brain Res 1996, 112:103-111.
This PET study compared cortical activation during motor imagery of grasping movements and during observation of the same movements performed by an actor. In the latter condition, SMA, lateral area 6, area 45 in the inferior frontal gyrus and area 40 in the posterior parietal lobe were involved.

18 Mellet E, Petit L, Mazoyer B, Denis M, Tzourio N: Reopening the mental imagery debate. Lessons from functional anatomy. Neuroimage 1998, 8:129-139.

19** Deiber MP, Ibanez V, Honda M, Sadato N, Raman R, Hallett M: Cerebral processes related to visuomotor imagery and generation of simple finger movements studied with positron emission tomography. Neuroimage 1998, 7:73-85.
This PET study stresses the fact that neural networks activated during motor imagery and motor performance are distinct. In the condition 'imagine only', activation was oberved in the contralateral inferoparietal cortex, preSMA, anterior cingulate, premotor cortex and dorsolateral prefrontal cortex. In the condition 'imagine, then move', an additional activation was found in contralateral sensorimotor cortex and lateral cerebellum. In addition, in the latter condition, activity was decreased in the inferior prefrontal cortex, suggesting that this area is involved in motor inhibition during motor imagery.

20 Gerardin  E, Sirigu A, Lehéricy S, Poline JP, Agid Y, Le Bihan D: Two dissociable neuronal networks for real and imaged hand movements. An fMRI study (submitted).

21 Sirigu A, Cohen L, Duhamel J-R, Pillon B, Dubois B, Agid Y, Pierrot-Deseiligny C: Congruent unilateral impairments for real and imagined hand movements. NeuroReport, 1995, 6:997-1001.

22 Dominey P, Decety J, Broussolle E, Chazot G, Jeannerod M: Motor imagery of a lateralized sequential task is assymetrically slowed in hemi-Parkinson patients. Neuropsychologia 1995, 33:727-741.

23* Roth M, Decety J, Raybaudi M, Massarelli R, Delon-Martin C, Segebarth C, Morand S, Gemignani A, Decorps M, Jeannerod M: Possible involvement of primary motor cortex in mentally simulated movement. A functional magnetic resonance imaging study. Neuroreport 1996, 7:1280-1284.
One of  the early papers using fMRI to demonstrate the involvement of primary motor cortex in motor simulation. Normal subjects performed (actually or mentally) a sequence of finger opposition movements (the so-called Luria task) during the scans. In 4/6 subjects the primary motor cortex contralateral to the imagined movement was activated at a rate which correponded to about 30% of activation during motor performance

24** Porro CA, Francescato MP, Cettolo V, Diamond ME, Baraldi P, Zuiani C, Bazzocchi M, di Prampero PE: Primary motor and sensory cortex activation during motor performance and motor imagery. A functional magnetic resonance study. J Neurosci 1996, 16:7688-7698.
A very careful study demonstrating the implication of motor cortex in mental simulation of finger opposition movements. Activation obtained during motor imagery was less that during motor performance, but significantly more than during mentally imagining a static visual scene. In addition, the pixels activated during both motor imagery and motor performance appear to represent a large fraction of all pixels activated during motor performance.

25 Cochin S, Barthelemy C, Roux S, Martineau J: Observation and execution of movement. Similarirtities demonstrated by quatified electroencephalography. Eur J Neurosci 1999, 11:1839-1842.

26 Schnitzler A, Salenius S, Salmelin R Jousmäki V, Hari R: Involvement of primary motor cortex in motor imagery; A neuromagnetic study. Neuroimage 1997, 6:201-208.

27 Lang W, Cheyne D, Hollinger P, Gerschlager W, Lindinger G: Electric and magnetic fields accompanying internal simulation of movement. Cog Brain Res 1998, 3:125-129.

28* Fadiga L, Buccino G, Craighero L, Fogassi L, Gallese V, Pavesi G: Corticospinal excitability is specifically modulated by motor imagery. A magnetic stimulation study. Neuropsychologia 1999, 37:147-158.
An elegant study of corticospinal excitability during motor imagery of arm and finger movements, using transcranial magnetic stimulation (TMS). Motor evoked potentials were larger for those muscles which were involved in the imagined movement. In addition, TMS applied to the right motor cortex increased corticospinal excitability on the left side only, whereas TMS applied to the left motor cortex increased exctibility bilaterally.

29* Hashimoto R, Rothwell JC: Dynamic changes in corticospinal excitability during motor imagery. Exper Brain Res 1999, 125:75-81.
This study reported the effects of trnascranial magnetic stimulation 5TMS) of motor cortex during motor imagery of wrist flexion-extension movements. In 5/9 subjects, motor evoked potentials were larger in the flexor muscle during imagined flexion than during imagined extension, and vice-versa. The H-reflex was not modified during motor imagery, which suggests that the motor system can generate these commands without recourse to sensory feedback.

30 Kasai T, Kawai S, Kawanishi M, Yahagi S: Evidence for facilitation of motor evoked potentials (MEPs) induced by motor imagery. Brain Res 1997, 744:147-150

31 Bonnet M, Decety J, Requin J, Jeannerod M: Mental simulation of an action modulates the excitability of spinal reflex pathways in man. Cog Brain Res 1997, 5:221-228.

32 Yahagi S, Shimura Y, Kasai T: An increase in cortical excitability with no change in spinal excitability during motor imagery. Percept Mot Skills 1996, 83:288-290.

33 Gandevia SC, Wilson LR, Inglis JT, Burke D: Mental rehearsal of motor tasks recruits alpha motoneurons, but fails to recruit human fusimotor neurons selectively. J Physiol 1997, 505:259-266.

34 Crammond DJ: Motor imagery: Never in your wildest dream. Trends Neurosci 1997, 20:54-57.

35 Rizzolatti G, Fadiga L, Gallese V, Fogassi L: Premotor cortex and the recognition of motor actions. Cog Brain Res 1996, 3:131-141.

36 Roure R, Collet C Deschaumes-Molinaro C, Dittmar A, Raqa H, Delhomme G, Vernet-Maury E: Autonomic nervous system responses correlate with mental rehearsal in volleyball training. Eur J Appl Physiol 1998, 78:99-108.

37 Yagüez L, Nagel D, Hoffman H, Canavan AGM, Wist E, Hömberg V: A mental route to mental learning. Improving trajectorial kinematics through imagery training. Behav Brain Res 1998, 90:95-106.
 
 

Legend for Figure 1

The subjects were seated in front of a horizontally placed 15" monitor. A cylindrical container filled with water (5cm high, 3cm in diameter) was placed at the center of the monitor screen at a distance of 50 cm from the body plane. Subjects were asked to lift the plastic cylinder and pour the water into another container using a precision grip formed by the thumb and index fingers of their right hand. They were also asked to carefully observe the axis defined by the contact points of the fingers on the cylinder surface, along which the force was applied during the grasp (the opposition axis). After several repetitions of this task, the cylinder was removed and the computer monitor was used to display the target stimuli. In each trial an image of the upper surface of the cylinder (a circle) was presented for 5 seconds at the same location where the real cylinder was placed during the preliminary run. Each circle was marked with two contact points with defined an opposition axis at various orientations with respect to the frontal plane. Subject’s task consisted, when shown a stimulus, to answer as quickly as possible whether the previously experienced action of grasping the cylinder full of water and pouring the water would be possible with the fingers placed according to the opposition axis indicated on the circle. No actual movement was allowed. The subjects had to rate the level of feasibility of the grasp with 3 levels (“easy”, “difficult”, “impossible") by pressing one of three keys with their left hand. Response time was measured.
Responses times (Figure 1A) were longer for the grasps judged to be more difficult due to the orientation of opposition axis, as shown by Figure 1B. (Unpublished data from Frak, Paulignan and Jeannerod).
 
 

Legend for Figure 2.
Subjects were instructed to imagine forearm flexion-extension movements with their right arm. The course of the mental simulation of the movement was paced by a frequency modulated sound. Transcranial magnetic stimulation (TMS) was applied to motor cortex on one side and the motor evoked potentials (MEPs) were recorded from the contralateral flexor muscle (biceps brachialis). The figure shows a typical example from one subject. A, B: control experiment where TMS was applied during visual imagery of a luminous bar shrinking or expanding. No change is observed in the flexor MEPs. C: Flexor MEPs recorded following TMS applied during the extension phase of the motor imagery. D : MEPs recorded following TMS during the flexion phase. These MEPs are selectively and significantly increased, as shown in part E (black bar : Z-score for MEP area of flexor muscle during imagined flexion; grey bar, during imagined extension). From 28*, with permission.