RESEARCH ARTICLE Cathy M. Stinear �� Winston D. Byblow Maarten Steyvers �� Oron Levin Stephan P. Swinnen Kinesthetic, but not visual, motor imagery modulates corticomotor excitability Received: 9 February 2005/ Accepted: 24 May 2005/Published online: 3 August 2005 �� Springer-Verlag 2005 Abstract The hypothesis that motor imagery and actual movement involve overlapping neural structures in the central nervous system is supported by multiple lines of evidence. The aim of this study was to examine the modulation of corticomotor excitability during two types of strategies for motor imagery: Kinesthetic Motor Imagery (KMI) and Visual Motor Imagery (VMI) in a phasic thumb movement task. Transcranial magnetic stimulation (TMS) was applied over the contralateral motor cortex (M1) to elicit motor evoked potentials (MEPs) in the dominant abductor pollicis brevis (APB) and abductor digiti minimi (ADM). In a separate experiment, transcutaneous electrical stimuli were delivered to the median nerve at the dominant wrist, to elicit F-waves from APB. Imagined task performance was paced with a 1 Hz auditory metronome, and stimuli were delivered either 50 ms before (ON phase), or 450 ms after (OFF phase), the metronome beeps. Recordings were also made during two control condi- tions: Rest, and a Visual Static Imagery (VSI) condition. Significant MEP amplitude facilitation occurred only in APB, and only during the ON phase of KMI. F-wave persistence and amplitude were unaffected by imagery. These results demonstrate that kinesthetic, but not vi- sual, motor imagery modulates corticomotor excitabil- ity, primarily at the supraspinal level. These findings have implications for the definition of motor imagery, and for its therapeutic applications. Keywords Motor cortex �� Motor imagery �� Visual imagery �� Human Introduction The involvement of premotor, supplementary motor, cingulate and parietal cortical areas during motor imag- ery and actual movement is supported by multiple lines of evidence from PET (Deiber et al. 1998 Jackson et al. 2003), and fMRI studies (Porro et al. 1996 Lotze et al. 1999 Gerardin et al. 2000 Porro et al. 2000 Ehrsson et al. 2003 Hanakawa et al. 2003 Kuhtz-Buschbeck et al. 2003 Dechent et al. 2004 Meister et al. 2004). However, there is conflicting evidence regarding the involvement of primary motor cortex (M1) during motor imagery. Contralateral M1 activity has been observed during imagined unilateral upper limb movement using EEG (Beisteiner et al. 1995 Pfurtscheller and Neuper 1997 Caldara et al. 2004 Mattia et al. 2004) and fMRI (Roth et al. 1996 Lotze et al. 1999 Jancke et al. 2001 Ehrsson et al. 2003 Lotze et al. 2003 Nair et al. 2003 Ross et al. 2003), while bilateral M1 activity has been observed during motor imagery of tongue protrusion (Spiegler et al. 2004) and a golf swing (Ross et al. 2003). In contrast, other studies using EEG (Romero et al. 2000), PET (Decety et al. 1994 Stephan et al. 1995), and fMRI (Gerardin et al. 2000 Hanakawa et al. 2003 Meister et al. 2004) have not shown any significant M1 activity during motor imagery of upper limb movement. The question of whether M1 excitability is altered during motor imagery has also been examined using transcranial magnetic stimulation (TMS). During motor imagery there is an increase in corticomotor excitability C. M. Stinear (&) �� W. D. Byblow Human Motor Control Laboratory, Department Sport & Exercise Science, University of Auckland, Private Bag 92019, Auckland, New Zealand E-mail: c.stinear@auckland.ac.nz Tel.: +64-9-3737-599, ext 86844 Fax: +64-9-3737-043 M. Steyvers �� O. Levin �� S. P. Swinnen Motor Control Laboratory, Department of Kinesiology, Group Biomedical Sciences, Katholieke Universiteit Leuven, Leuven, Belgium M. Steyvers Division of Physical Therapy, Department of Health Sciences, Hogeschool Antwerpen, Belgium Exp Brain Res (2006) 168: 157���164 DOI 10.1007/s00221-005-0078-y

that is both muscle-specific and temporally modulated, mirroring the modulation in excitability that is observed during actual task performance (Rossi et al. 1998 Yahagi and Kasai 1998 Fadiga et al. 1999 Hashimoto and Rothwell 1999 Facchini et al. 2002 Clark et al. 2003 Stinear and Byblow 2003). Based on this body of evidence, we have concluded that M1 is engaged by motor imagery in a manner that mimics its involvement in actual task performance. The extent to which M1 may or may not be engaged during motor imagery can be confounded by two fac- tors. First, of the numerous investigations of motor imagery, very few have reported an analysis of EMG recorded during data acquisition (Hashimoto and Rothwell 1999 Romero et al. 2000 Jackson et al. 2003 Stinear and Byblow 2003 Caldara et al. 2004 Stinear and Byblow 2004). This is a fundamental issue because motor imagery is no longer imagery when the muscu- lature is activated. Motor imagery is a cognitive process that engages a variety of supraspinal structures, without resulting in any outflow from the spinal motorneuron pool. This precise neurophysiological definition is in contrast to the widely accepted notion that motor imagery simply produces no overt joint movement. This study therefore aimed to explore changes in corticomo- tor excitability at supraspinal and spinal levels, in the absence of any outflow from the spinal motorneuron pool. Second, differences between outcomes of studies to date may also have arisen due to differences in motor imagery strategies. Generally, motor imagery strategies can be divided into kinesthetic motor imagery (KMI) and visual motor imagery (VMI) (Hall et al. 1985). KMI involves imagining the feeling that actual task perfor- mance produces. VMI involves imagining seeing your- self performing the task. Few studies of motor imagery explicitly report the instructions given to subjects, so it is di���cult to determine whether subjects were biased to- wards using one type of strategy or the other. To date, no studies have compared the modulation of M1 acti- vation or corticomotor excitability during KMI and VMI of the same movement. The aim of the present study, therefore, was to ex- plore changes in corticomotor excitability during KMI and VMI of the same motor task ��� phasic thumb abduction, while maintaining careful control of back- ground EMG activity. We hypothesised that KMI would produce significant modulation of the excitability of the thenar muscle representation in M1, compared to VMI of the same task and two control conditions (Rest and Visual Static Imagery). Specifically, we predicted (a) that there would be clear facilitation of responses re- corded from abductor pollicis brevis (APB) during KMI of phasic thumb movement, but not during VMI of the same movement (b) that this facilitation would only occur at the time of imagined movements (ON phase), and not between them (OFF phase) and (c) that any changes in response amplitude during KMI would be confined to APB, and not be observed in a control muscle (abductor digiti minimi, ADM). Furthermore, we predicted that spinal motorneuron excitability, as reflected by APB F-wave persistence and amplitude, would not be affected by any type of imagery. Materials and methods Participants Ten right-handed subjects participated in the TMS experiment (6 female, mean age 32 years, range 26��� 42 years), and ten right-handed subjects participated in the F-wave experiment (7 female, mean age 33 years, range 20���53 years). All subjects were right-handed, as assessed using the Edinburgh Handedness Inventory (Oldfield 1971), (mean laterality score 91.4, range 72.2��� 100). The study was approved by the local ethical committees of KU Leuven and the University of Auckland, and subjects gave their written informed consent, in accordance with the Declaration of Helsinki. Subjects had no neurological symptoms or deficits, and no history of neurological illness. Preparation Subjects were seated comfortably with their dominant hand and forearm resting on a pillow on their lap in a semi-supinated position. For the TMS experiment, surface electromyography (EMG) was recorded from their dominant abductor pollicis brevis (APB) and abductor digiti minimi (ADM) using a pair of 10 mm diameter electrodes (Blue Sensor, Medicotest, Den- mark), following standard skin preparation techniques (Sethi and Thompson 1989). The preamplified signals (gain 80 dB) were bandpass filtered (15 Hz���1 kHz), sampled at 5 kHz (CED Power 1401, Cambridge Electronic Design, Cambridge, UK) and stored for offline analysis. For the F-wave experiment, surface EMG was recorded from the dominant APB using a pair of 12 mm diameter surface Ag���AgCl Hydrospot electrodes (Physiometrix Inc., MA, USA), following standard skin preparation techniques (Sethi and Thompson 1989). Signals were amplified by two Grass P511AC EMG amplifiers (Grass Instrument Division, RI, USA). The EMG data were bandpass filtered (30 Hz���1 kHz), sampled at 4 kHz with a 12-bit Mac- Lab A/D acquisition system and software, and stored for offline analysis. Single pulse transcranial magnetic stimuli were delivered using a Dantec MagLite r-25 stimulator (Medtronic, Skovlunde, Denmark), via a figure-of-eight coil (MC-B70, outer diameter 50 mm, biphasic pulse (width 280 ls)). The coil was positioned over the left motor cortex, at the optimal site for producing responses in the resting APB muscle. This site was marked to en- sure consistent coil placement. The coil was handheld and oriented so that the handle was directed posteriorly at a 45�� angle to the midline, so that the induced current 158