Longitudinal changes of motor cortical excitability and transcallosal inhibition after subcortical stroke.

Clinical Neurophysiology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph

Longitudinal changes of motor cortical excitability and transcallosal inhibition after subcortical stroke Utako Takechi a,⇑, Kaoru Matsunaga b, Ryoji Nakanishi b, Hiroaki Yamanaga b, Nobuki Murayama c, Kosuke Mafune d, Sadatoshi Tsuji a a

Department of Neurology, University of Occupational and Environmental Health, School of Medicine, Japan Department of Neurology and Rehabilitation Medicine, Kumamoto Kinoh Hospital, Kumamoto, Japan Department of Human and Environmental Informatics, Graduate School of Science and Technology, Kumamoto University, Japan d Department of Mental Health, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Japan b c

a r t i c l e

i n f o

Article history: Available online xxxx Keywords: Stroke Transcranial magnetic stimulation Motor-evoked potential Intracortical inhibition Transcallosal inhibition

h i g h l i g h t s Continuing clinical improvement was observed over a 1-year period following subcortical stroke. Overall excitability in the unaffected hemisphere was increased at the post-acute period following

stroke, which may have resulted in enhanced transcallosal inhibition to the affected hemisphere. It is unclear whether there was a causal relationship between the enhanced transcallosal inhibition

and the extent of clinical recovery.

a b s t r a c t Objective: A general lack of longitudinal studies on interhemispheric interactions following stroke led us to use transcranial magnetic stimulation (TMS) to examine changes in corticospinal/intracortical excitability and transcallosal inhibition over a 1-year period following subcortical stroke. Methods: We measured TMS parameters such as motor threshold (MT), short-interval intracortical inhibition (SICI), and ipsilateral silent period (iSP) and evaluated clinical scores at three time-points (T1, T2, and T3) in 24 patients and 25 age-matched healthy subjects. Results: At T1, we observed reduced MTs and SICIs with prolonged iSPs in the unaffected hemisphere (UH). In contrast, increased MTs and reduced SICIs were observed in the affected hemisphere (AH). These abnormalities gradually reduced and no MEP response to TMS at T1 predicted a worse prognosis. The prolonged iSP at T1 was associated with more severe impairments, but it did not necessarily predict a worse prognosis after 1 year. Conclusions: UH excitability was increased at the post-acute time-period, which may have resulted in enhanced transcallosal inhibition to the AH. However, it is unclear whether there was a causal relationship between the enhanced transcallosal inhibition and the extent of clinical recovery. Significance: This is the first study to demonstrate changes in transcallosal inhibition over a longitudinal period following stroke. Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

1. Introduction Motor circuit reorganization in the cerebral cortex is known to contribute to recovery following stroke. This reorganization can be ⇑ Corresponding author. Address: Department of Neurology, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu-city, Fukuoka 807-8555, Japan. Tel.: +81 93 693 7438; fax: +81 936939842. E-mail address: [email protected] (U. Takechi).

examined by transcranial magnetic stimulation (TMS) using measures of corticospinal and intracortical excitability (Liepert et al., 2000; Shimizu et al., 2002; Liepert et al., 2005; Talelli et al., 2006; Wittenberg et al., 2007; Bütefisch et al., 2008; Manganotti et al., 2008; Swayne et al., 2008; Takeuchi et al., 2010). TMS measures such as motor threshold (MT) and recruitment curve (RC) reflect corticospinal excitability, whereas intracortical excitability is reflected by cortical silent period (cSP), and short-interval intracortical inhibition (SICI) and intracortical facilitation (ICF), which is

http://dx.doi.org/10.1016/j.clinph.2014.01.034 1388-2457/Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Takechi U et al. Longitudinal changes of motor cortical excitability and transcallosal inhibition after subcortical stroke. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.01.034

2

U. Takechi et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

assessed by the paired-pulse TMS technique (Kujirai et al., 1993). Previous studies have demonstrated that reduced corticospinal excitability in the affected hemisphere could result not only from direct cortical damage but also from disruption of the corticospinal connection. In contrast, some studies have shown that decreased SICIs in both hemispheres might be associated with functional recovery after stroke (Liepert et al., 2000; Shimizu et al., 2002; Liepert et al., 2005; Wittenberg et al., 2007; Bütefisch et al., 2008; Manganotti et al., 2008). Most of these studies, however, were conducted over a short time-period following stroke (within 6 months). It has been proposed that activity of the motor cortex (M1) in the unaffected hemisphere (UH) and/or the ipsilateral motor pathway of patients following stroke may play a role in functional recovery. This theory is based on an observation that the M1 in the UH of these patients was activated in association with movements of the paretic hand (Chollet et al., 1991; Cao et al., 1998; Marshall et al., 2000; Gerloff et al., 2006). However, there is some evidence that this is not the case. For example, the magnitude of M1 activation in the UH due to paretic hand movement does not correlate with functional recovery (Cramer et al., 1997; Ward et al., 2003). Additionally, ipsilateral motor pathways from the UH to the paretic hand are more commonly detected in patients with poor motor recovery (Netz et al., 1997; Gerloff et al., 2006). Lastly, a TMS study showed that disrupted M1 activation in the UH failed to delay simple reaction times in the paretic hand of patients with chronic stroke (Werhahn et al., 2003). Therefore, these findings support the hypothesis that a rescue of motor function in patients affected by chronic stroke relies predominantly on reorganization in the affected hemisphere (AH) (Werhahn et al., 2003; Ward et al., 2006). However, more recent imaging studies have demonstrated that the increased activation of the UH including the M1 was seen during movement of the affected hand in fully recovered chronic stroke patients (Butefisch et al., 2005; Schaechter and Perdue, 2008). Additionally, Lotze et al. (2006) demonstrated that rTMS over the dorsal premotor cortex, M1 and the superior parietal lobe of the UH during complex hand movements resulted in significant interference with recovered performance in chronic stroke patients. Thus, these studies support the idea that the UH is beneficial for some aspects of effectively recovered motor behavior after stroke (Lotze et al., 2006). It has recently been suggested that interhemispheric imbalance is strongly related to motor function of the affected hand in patients with chronic stroke (Hummel et al., 2006; Nowak et al., 2009). Interhemispheric interactions between the two motor cortices can be assessed by paired-pulse TMS using measures such as interhemispheric inhibition (IHI) (Ferbert et al., 1992) or by ipsilateral silent period (iSP) (Meyer et al., 1995). Both IHI and iSP are believed to involve transcallosal pathways (Ferbert et al., 1992). Murase et al. (2004) studied IHI from the UH to the AH while patients with chronic subcortical stroke were told to generate a voluntary movement in their paretic hand. IHI at rest was comparable between patients and healthy controls, however, closer to movement onset, IHI turned into facilitation in controls but not in patients (Murase et al., 2004). These results suggest that an enhanced IHI from the UH to the AH might interfere with motor function of the affected hand. Another study demonstrated that 1-Hz rTMS (suppressive rTMS) over the UH can reduce the iSP (transcallosal inhibition from the UH to the AH) and improve motor function of the affected hand in patients with chronic stroke (Takeuchi et al., 2005). This group suggested that M1 in the UH inhibits M1 in the AH via an abnormally enhanced transcallosal inhibition, and that disruption of this abnormal transcallosal inhibition by 1-Hz rTMS causes paradoxical functional facilitation of the affected hand. Based on these findings, it has been suggested that enhanced transcallosal inhibition from the UH to the AH could

adversely influence motor recovery in some patients with subcortical stroke (Murase et al., 2004). However, there have been no longitudinal studies on interhemispheric interactions following stroke. One question addressed in the current study is how reorganization of bilateral motor circuits and interhemispheric connections occurs over the long course of recovery following stroke. A related question is how neurophysiological measures correlate with clinical scores. We hypothesize that there is excessive transcallosal inhibition from the UH to the AH associated with the disinhibited M1 of the UH, which could hamper recovery of motor symptoms, especially in the post-acute phase. Furthermore, we hypothesize that this interhemispheric imbalance gradually improves with clinical recovery. Therefore, in the current study we assessed longitudinal changes of transcallosal inhibition (using iSP) as well as corticospinal (using MT and RC) and intracortical (using cSP, SICI, and ICF) excitability by TMS and their clinical correlations from post-acute to chronic time-periods (up to 1 year) following subcortical stroke. Patients with subcortical stroke were chosen on the grounds that their cerebral cortex and transcallosal pathways were preserved, and we considered that dynamic functional changes could occur in these patients.

2. Subjects and methods 2.1. Subjects A total of 24 patients (16 men and eight women) were recruited from the recovery rehabilitation ward. Ages of the patients ranged between 39 and 81 with a mean age of 63.6. Fourteen patients had hemorrhagic stroke while 10 had ischemic stroke; clinical characteristics of these patients are shown in Tables 1 and 2. In order to be included in the study, patients had to have experienced their first-ever subcortical stroke within the last 40 days, resulting in upper limb weakness and a single monohemispheric lesion as found by magnetic resonance imaging (MRI). Patients were

Table 1 Patient characteristics. Pt

Age

Sex

Affected hand

Lesion

Etiology

T1

T2

T3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

64 76 69 81 50 72 75 56 52 55 51 39 73 72 77 66 50 58 62 60 72 72 67 58

M M F F F M F M M M M M M M F F M F M M M F M M

R R L R R L R L R R R R R R R R R R L R R L R L

P CR CR CRDWM T P IC P P CRDWM CR CR CR CR P T T T P T P T CR T

Hemorrhage Lacunar Lacunar Atheroma Hemorrhage Hemorrhage Lacunar Hemorrhage Hemorrhage Atheroma Lacunar Lacunar Lacunar Lacunar Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage Lacunar Hemorrhage

30 28 18 23 24 28 33 23 22 40 18 18 18 31 28 31 23 13 18 35 40 34 32 33

163 91 81 58 135 69 104 93 71 172 53 123 54 59 104 116 93 97 116 140 89 118 137 117

367 372 360 365 359 365 363 367 366 368 362 369 365 367 363 367 366 378 362 371 369 370 361 361

Pt indicates patient; M, male; F, female; Rt, right; Lt, left; P, putamen; CR, corona radiata; T, thalamus; IC, internal capsule; DWM, deep white matter, T1, T2, T3, days post-stroke.


3


excluded from the study if they had a history of epilepsy or other neurological diseases, were active users of drugs targeting the CNS, had a cardiac pacemaker or could not undergo TMS due to other contraindications, were pregnant, were terminally ill, and/ or possessed a degree of dementia that left the patient uncooperative. Subcortical stroke location was defined as a lesion below the level of the corpus callosum, sparing the cerebral cortex. Lesions were also regarded as ‘‘subcortical’’ if they involved the deep white matter inferior to the corpus callosum, including the internal capsule, thalamus, and basal ganglia. Hemorrhagic stroke was classified into two groups (thalamic or putaminal) on the basis of the hemorrhage location (Table 1). All patients received multidisciplinary post-stroke care appropriate to their clinical needs. An agematched control group of 25 healthy subjects (17 men and eight women between the ages of 54 and 83, with a mean age of 63) also participated in this study. The control subjects reported no history of epilepsy or other neurological diseases and were not taking any regular medication. Full written informed consent was obtained from all subjects in accordance with the Declaration of Helsinki, and the study was approved by the Research Ethics Committee of Kumamoto Kinoh Hospital, Japan. 2.2. Behavioral evaluation For each session, upper limb function was evaluated by assessing hand grasping power and performance in the Jebsen–Taylor Hand Function Test (JHFT) (Jebsen et al., 1969) and Fugl–Meyer test (Gladstone et al., 2002). Hand grasping power, which was performed using a Hand Dynamometer T.K.K. 5401 (Takei-Kiki, Japan), was evaluated twice for each hand with the higher of the two scores being selected. The Hand Dynamometer T.K.K. 5401 was sensitive to forces greater than 5 kg, and a grip of 5 kg or less was referred to as 0 kg. The JHFT includes seven tasks including writing, turning over cards, picking up small objects, simulated feeding, stacking checkers, picking up large objects, and picking up large heavy objects, and is scored based on the time it takes to complete each task (Jebsen et al., 1969). In the current study, some patients were unable to complete one or more tasks due to

relative weakness; therefore, the results of each task were classified into three categories and scored as 0, 1, or 2 (0: could not complete, 1: completed with the time over mean + 3 SD of the control group, 2: completed within the time of mean + 3 SD of the control group). The National Institutes of Health Stroke Scale (NIHSS) (Lyden et al., 1994) and the modified Rankin Scale (MRS) (van Swieten et al., 1988) were also evaluated for more global aspects of clinical recovery.

2.3. Transcranial magnetic stimulation TMS was performed using two MAGSTIM 200 stimulators (Magstim, Dyfed, UK) connected via a Y-cable to a figure-eight coil with an external wing diameter of 9 cm. During stimulation, the coil was held with the handle pointing posterolaterally at an angle of 45 degrees to midline, so that the current induced in the brain was anterior to the central sulcus (Matsunaga et al., 2005). Surface electromyographic (EMG) recordings in a belly-to-tendon montage were made bilaterally from the first dorsal interosseous muscles (FDI). The position at which the stimulation produced optimal MEPs in the contralateral FDI was identified on each side. The raw EMG signal was then amplified, band-pass filtered (30–1000 Hz) by a Synax 1200 (NEC, Japan), and sampled at a rate of 5000 Hz on a laboratory computer for off-line analysis (Signal Software, Cambridge Electronic Design, Cambridge, UK). For the right and left hands, the following parameters of corticospinal and intracortical excitability and transcallosal inhibition were assessed: resting motor threshold (rMT), active motor threshold (aMT), RC, cSP, iSP, SICI, and ICF (Table 3).

2.3.1. Motor threshold The rMT was defined as the lowest stimulation intensity required to evoke an MEP in the resting FDI >50 lV in five out of 10 trials. The aMT was assessed during slight (10–15% maximum) tonic contraction of the FDI and was defined as the lowest stimulation intensity required to evoke an MEP >200 lV in five out of 10 trials (Matsunaga et al., 2005). If no MEP could be evoked at

Table 2 Clinical scores in each patient. Patient

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

NIHSS

MRS

Hand grasping power (Kg)

JHFT

Fugl-Meyer

T1

T2

T3

T1

T2

T3

T1

T2

T3

T1

T2

T3

T1

T2

T3

11 9 2 5 10 4 11 14 5 18 5 7 4 3 18 18 11 7 11 14 6 10 8 8

8 4 2 3 8 2 10 3 3 11 3 4 3 3 12 5 2 3 7 11 4 6 6 4

8 3 1 1 8 2 10 3 3 10 2 2 2 0 12 3 2 1 7 7 2 5 5 3

5 3 2 4 2 4 4 4 3 5 3 4 3 2 5 5 4 3 4 5 3 5 4 4

4 2 1 2 1 2 3 2 2 4 2 3 2 1 4 3 3 2 3 4 2 4 3 3

3 1 1 1 1 2 3 2 2 3 2 2 1 1 4 2 2 1 3 3 1 3 3 3

0 0 20.7 12 0 19.7 0 17.3 26.8 0 5.9 17.9 19 18.5 0 0 8.4 10.1 0 6.4 8.8 0 0 6

0 0 23.5 15.6 0 14.8 0 24.3 40.6 0 13.4 12.1 21.2 21.2 0 5.5 13.3 12.1 0 7.7 14.7 0 0 6.8

7.3 11.8 24.1 17.3 5.2 21 0 35 41.2 0 5.9 24.3 23.5 29 0 10 17.5 15.9 0 13 22.6 0 5.6 12.5

7 7 14 14 7 14 7 12 15 7 14 7 17 20 7 7 7 14 7 7 12 7 7 9

7 10 18 14 7 15 7 18 21 7 14 14 19 20 7 14 14 17 7 14 17 8 7 14

7 14 20 16 9 18 7 21 21 7 17 19 21 21 7 16 15 20 7 13 21 9 7 15

4 27 65 53 5 51 3 43 59 2 26 26 53 61 6 2 41 55 6 15 28 3 16 47

11 40 65 59 40 63 13 64 66 17 51 57 58 61 11 50 61 59 6 23 66 33 20 59

16 46 65 60 44 64 14 65 66 18 54 60 58 62 11 51 62 60 6 24 66 34 22 60

NIHSS indicates National Institutes of Health Stroke Scale; MRS, Modified rankin scale; JHFT, Jebsen–Taylor hand function test.


(24) (24) (24) (24) (24) (24) (24) (14) (24) (24) (24) (24) (24) (24) (24) (14) 0.58 ± 0.03 95.6 ± 10.4 99.8 ± 11.0 133.4 ± 14.8 136.8 ± 11.7 159.3 ± 14.8 94.08 ± 4.86 53.64 ± 7.50 (13) (13) (13) (13) (13) (13) (14) (24) 0.68 ± 0.06 98.0 ± 14.6 89.5 ± 11.2 128.0 ± 20.2 146.7 ± 26.5 143.0 ± 25.2 144.43 ± 14.2 25.33 ± 1.68 (13) (13) (13) (13) (13) (13) (14) (24) 0.63 ± 0.05 85.5 ± 6.7 89.6 ± 5.9 104.5 ± 7.3 145.0 ± 18.2 121.8 ± 15.5 198.36 ± 15.6 28.33 ± 1.39 (13) (13) (13) (13) (13) (13) (14) (24) (25) (25) (25) (25) (25) (25) (25) (25) 0.58 ± 0.05 59.7 ± 5.2 78.6 ± 10.1 102.3 ± 8.7 164.9 ± 13.0 154.8 ± 10.9 110.80 ± 5.84 19.40 ± 7.91 cSP (ms) iSP (ms)

ICF (% control MEP)

2 ms 3 ms 4 ms 10 ms 15 ms RC gradient Paired-pulse TMS Control MEP size (mV) SICI (% control MEP)

100% 110% 120% 130% 140%

(25) (25) (25) (25) (25) (25) (25) (25)

0.50 ± 0.06 106.8 ± 20.4 107.6 ± 13.7 126.9 ± 13.8 143.2 ± 16.9 146.6 ± 16.3 217.36 ± 17.2 27.38 ± 2.04

(24) (24) (11) (11) (11) (11) (11) (11) (24) (24) (11) (11) (11) (11) (11) (11)

19/24 21/24 24/24 65.5 ± 4.68 54.5 ± 4.95 0.12 ± 0.02 0.24 ± 0.05 0.32 ± 0.07 0.88 ± 0.21 1.09 ± 0.26 0.51 ± 0.13 14/24 14/24 24/24 73.6 ± 5.08 64.2 ± 5.85 0.16 ± 0.03 0.71 ± 0.18 1.13 ± 0.27 1.55 ± 0.38 1.90 ± 0.35 0.86 ± 0.19 25/25 25/25 25/25 54.2 ± 2.03 42.1 ± 1.54 0.14 ± 0.02 0.33 ± 0.06 0.59 ± 0.09 1.12 ± 0.18 1.78 ± 0.22 0.81 ± 0.11 Rate of recordable MEP at rest Rate of recordable MEP during VC Rate of recordable iSP rMT (%) aMT (%) RC (mV)

VC indicates voluntary contraction. The number of subjects is shown in parentheses. Values from the right FDI are shown in normal subjects. iSP in AH indicates values from the affected to unaffected hemispheres; iSP in UH, values from the unaffected to affected hemispheres.

0.68 ± 0.04 60.6 ± 6.8 72.4 ± 7.5 93.7 ± 8.4 151.3 ± 10.5 149.8 ± 13.6 114.63 ± 8.08 26.14 ± 3.54

(24) (24) (23) (23) (23) (23) (23) (23) (24) (24) (23) (23) (23) (23) (23) (23) (24) (24) (11) (11) (11) (11) (11) (11)

0.61 ± 0.04 66.4 ± 6.0 80.9 ± 5.6 109.9 ± 7.9 159.0 ± 16.2 171.0 ± 24.9 117.13 ± 7.07 34.07 ± 4.86

24/24 24/24 24/24 51.6 ± 2.14 41.8 ± 1.40 0.13 ± 0.01 0.48 ± 0.07 1.12 ± 0.17 1.71 ± 0.22 2.13 ± 0.25 1.04 ± 0.13 24/24 24/24 19/24 51.2 ± 2.45 39.5 ± 1.93 0.12 ± 0.01 0.33 ± 0.04 0.79 ± 0.10 1.30 ± 0.17 1.96 ± 0.20 0.93 ± 0.11

T3 T2

24/24 24/24 14/24 47.0 ± 2.01 37.7 ± 1.66 0.15 ± 0.01 0.41 ± 0.05 0.99 ± 0.14 1.75 ± 0.25 2.57 ± 0.29 1.24 ± 0.16 21/24 24/24 24/24 59.4 ± 4.09 48.9 ± 4.14 0.10 ± 0.01 0.33 ± 0.08 0.51 ± 0.12 1.13 ± 0.26 1.80 ± 0.31 0.83 ± 0.16

Unaffected hemishpere

T1 T3 T2 T1

Affected hemishpere

Patients Normal subjects Table 3 TMS measures.

(24) (24) (24) (24) (24) (24) (24) (14)


(24) (24) (23) (23) (23) (23) (23) (23)

4

maximum stimulator output, then the threshold was described as 100%. 2.3.2. Recruitment curve In order to make the RC, MEPs were recorded with 10% interval stimulus intensities ranging between 100% and 140% of the rMT for each hemisphere (Chen et al., 1998). Ten trials were recorded at each stimulus intensity. 2.3.3. Short-interval intracortical inhibition and intracortical facilitation SICI and ICF were assessed by paired-pulse TMS in the resting condition with a subthreshold conditioning stimulus (CS) preceding the suprathreshold test stimulus (TS) (Kujirai et al., 1993). The CS was fixed at 80% of the aMT, and the TS was adjusted to evoke an unconditioned MEP of approximately 0.5 mV amplitude in the contralateral FDI. A TS was given either on its own or was preceded by a CS at interstimulus intervals (ISIs) of 2, 3, 4, 10, and 15 ms, so that there were six conditions. SICI was assessed at ISIs of 2, 3, and 4 ms, whereas ICF was assessed at 10 and 15 ms. These conditions were randomly intermixed and presented 10 times each, so that a total of 60 trials were performed in each experiment. 2.3.4. Contralateral and ipsilateral silent period We recorded both cSP and iSP following stimulation of each hemisphere during a unilateral maximum tonic contraction of the target muscle. At first, we stimulated the ipsilesional M1 at stimulus intensities that were 120% of the rMT to evoke the cSP from the affected FDI. Next, the contralesional M1 was stimulated at 150% of the rMT of the contralesional hemisphere to evoke the iSP from the affected FDI (Takeuchi et al., 2010). Thereafter, we recorded the cSP and iSP from the unaffected FDI with the same procedure. The durations of the cSP and iSP were assessed in the trace obtained from averaging the 10 single-rectified EMGs. The cSP duration was measured from the TMS onset to the first point when continuous EMG activity returned after a period of complete EMG suppression following the MEP (Triggs et al., 1992; Classen et al., 1997). The iSP was quantified by the period of relative EMG suppression after the TMS stimulus; that is, when EMG activity dropped below the background EMG activity (Trompetto et al., 2004; Takeuchi et al., 2005). ISP onset was defined as the point after the TMS at which EMG activity became constantly (greater than 10 ms in duration) under the mean amplitude of the background EMG activity preceding the TMS stimulus (mean EMG). The iSP end was defined as the first point after iSP onset at which the level of EMG activity regained the mean EMG. Lastly, the iSP duration was defined as iSP end – iSP onset (Trompetto et al., 2004; Takeuchi et al., 2005). The iSP is thought to be evoked through the transcallosal pathway stemming from the stimulated to the unstimulated motor cortex; therefore, the iSP is most likely an indicator of transcallosal inhibition (Boroojerdi et al., 1996; Trompetto et al., 2004; Avanzino et al., 2007). 2.4. Study design Behavioral and TMS assessments were performed at three timepoints: T1, at the beginning of the hospitalization period for rehabilitation; T2, at the end of the hospitalization period; and T3, 1 year after stroke onset. We chose T1 and T2 as the beginning and end of hospitalization in order to focus on the effect of inpatient rehabilitation services on changes in clinical and physiological measures. We chose 1 year after stroke onset to evaluate functional status over an extended time-period following the end of inpatient rehabilitation. This last time point is different from previous studies that have only assessed within half a year of


NIHSS

rMT (n = 24) aMT (n = 24) RC 110% (n = 11) RC 120% (n = 11) RC 130% (n = 11) RC 140% (n = 11) RC gradient (n = 11) SICI 2 ms (n = 13) SICI 3 ms (n = 13) SICI 4 ms (n = 13) ICF 10 ms (n = 13) ICF 15 ms (n = 13) cSP (n = 14) iSP (n = 14)

r p r p r p r p r p r p r p r p r p r p r p r p r p r p

MRS

Hand grasping (Kg)

JHFT

Fugl-Meyer

T1

T2

T3

T1

T2

T3

T1

T2

T3

T1

T2

T3

T1

T2

T3

.497 (.013) .524 (.009) .452 (.163) .466 (.149) .703 (.016) .493 (.123) .521 (.101) .477 (.100) .558 (.048) .488 (.090) .160 (.602) .080 (.795) .567 (.035) .726 (.003)

.483 (.017) .570 (.004) .278 (.408) .153 (.653) .005 (.989) .034 (.922) .086 (.801) .308 (.307) .183 (.549) .469 (.106) .428 (.145) .264 (.383) .481 (.081) .714 (.004)

.450 (.027) .542 (.006) .383 (.245) .065 (.848) .589 (.057) .252 (.454) .262 (.437) .017 (.956) .200 (.512) .246 (.418) .440 (.132) .023 (.941) .101 (.732) .373 (.189)

.347 (.097) .380 (.067) .557 (.075) .562 (.072) .491 (.125) .359 (.279) .411 (.210) .321 (.285) .266 (.379) .604 (.029) .426 (.147) .233 (.444) .599 (.024) .304 (.290)

.402 (.051) .426 (.038) .140 (.682) .085 (.804) .249 (.460) .424 (.194) .384 (.244) .025 (.935) .248 (.413) .072 (.814) .283 (.349) .192 (.529) .586 (.028) .602 (.023)

.484 (.017) .518 (.010) .307 (.359) .045 (.895) .568 (.068) .266 (.429) .241 (.475) .198 (.517) .204 (.504) .356 (.233) .499 (.083) .040 (.898) .410 (.145) .377 (.184)

.712 (.000) .656 (.001) .127 (.709) .136 (.689) .418 (.201) .245 (.467) .345 (.298) .394 (.183) .189 (.537) .121 (.694) .003 (.993) .264 (.383) .091 (.758) .814 (.000)

.529 (.008) .545 (.006) .087 (.800) .169 (.620) .159 (.640) .251 (.457) -.260 (.441) .262 (.388) .206 (.499) .501 (.081) .369 (.215) .010 (.975) .651 (.012) .689 (.006)

.442 (.031) .547 (.006) .291 (.385) .391 (.235) .100 (.770) .100 (.770) .127 (.709) .358 (.230) .264 (.384) .527 (.064) .648 (.017) .687 (.010) .437 (.118) .260 (.370)

.675 (.000) .671 (.000) .350 (.292) .308 (.357) .499 (.118) .317 (.342) .336 (.313) .458 (.116) .400 (.175) .194 (.526) .045 (.884) .160 (.602) .481 (.081) .697 (.006)

.621 (.001) .640 (.001) .084 (.806) .075 (.827) .042 (.902) .266 (.429) .145 (.671) .193 (.527) .126 (.682) .137 (.656) .263 (.386) .031 (.920) .710 (.004) .420 (.135)

.410 (.047) .532 (.007) .117 (.732) .318 (.341) .023 (.946) .145 (.671) .210 (.535) .263 (.386) .037 (.905) .422 (.151) .539 (.058) .553 (.050) .306 (.287) .108 (.712)

.684 (.000) .659 (.000) .410 (.210) .506 (.113) .465 (.150) .351 (.290) .487 (.128) .371 (.213) .248 (.414) .289 (.338) .195 (.523) .063 (.837) .004 (.988) .716 (.004)

.540 (.006) .562 (.004) .234 (.488) .179 (.598) .400 (.223) .018 (.957) .239 (.479) .206 (.499) .241 (.428) .288 (.341) .409 (.165) .169 (.581) .553 (.040) .451 (.106)

.388 (.061) .564 (.004) .115 (.736) .520 (.101) .276 (.412) .405 (.217) .437 (.179) .368 (.217) .393 (.184) .438 (.135) .662 (.014) .604 (.029) .246 (.397) .257 (.374)


NIHSS indicates National Institutes of Health Stroke Scale; MRS, Modified Rankin scale; JHFT, Jebsen-Taylor Hand Function Test. Note: rMT, aMT, cSP, RC, SICI, and ICF are values in the affected hemispheres. iSP are values from the unaffected hemisphere. Correlation coefficients (r) are shown for the plots of each TMS measure against the clinical scores shown. P values are shown in brackets. Significant correlations are shown in bold.

5


Table 4 Correlation coefficients between TMS measures and clinical scores at three time points.

6


stroke onset. Thus, the days after T1 and T2 varied among patients (Table 1) because they were heterogeneous with regard to impairment (Table 2). Moreover, patients with severe impairments needed a longer time to start intensive rehabilitation under stable mental and physical conditions, and needed a longer hospitalization period for rehabilitation in order to maximize recovery efforts; this wide range of impairment allowed correlations of functional status to be examined. Most patients, except for those with only mild motor dysfunction at discharge, have continued to receive regular outpatient rehabilitation services since discharge. 2.5. Data analysis and statistics When determining SICI and ICF, the mean peak-to-peak amplitude of conditioned MEPs at each ISI was expressed as the percentage of the mean amplitude of unconditioned MEPs. One- or twoway repeated-measures analysis of variance (ANOVA) were performed for behavioral and TMS measures. The RC was evaluated by a two-way repeated-measures ANOVA with ‘‘hemisphere’’ and ‘‘intensity’’ as factors for each time point and ‘‘intensity’’ and ‘‘time’’ as factors for each hemisphere. Differences from the control group were tested using a two-factor mixed-ANOVA, and a Bonferroni correction was used as a post-hoc test for multiple comparisons when the results for ANOVAs were significant. In order to evaluate differences in longitudinal changes of clinical scores between two patient groups, i.e., patients with (n = 14) or without (n = 10) MEPs recorded from the affected hand at T1, a two-factor mixed-ANOVA with ‘‘patient group’’ and ‘‘time’’ as factors was used. Data for the right FDI in the control group were used as control values because there was no significant difference between the left and right sides for any measure, except for RC (120%, 130%), which was lower in the right FDI compared to the left FDI. In addi-

tion, JHFT data for the left hand of the control group was used as the control value. Differences between the UH and AH were tested using a Student’s paired t-test, whereas differences from the control group were tested using unpaired t-tests. Mann-Whitney’s U test was used to assess the difference between hemorrhagic and ischemic stroke. Spearman’s correlation coefficient was used to (1) study the possible correlation between TMS measures and clinical scores at each time point (Table 4) and (2) test the possible correlation between longitudinal changes of TMS measures and clinical scores from T1 to T2 and from T1 to T3. For this purpose, we used data for the interval changes of both TMS measures and clinical scores (subtraction of T1 values from T2 values [T2 – T1] or from T3 values [T3 – T1]) (Table 5). We also performed a simple linear regression analysis (Supplementary Tables S1, S2 and S3) in order to investigate whether TMS measures at T1 or T2 could predict subsequent clinical recovery (clinical scores at T2 or T3). The gradient of the line of best fit for the RC (RC gradient) at each time point in each patient was also obtained using the least squares method. This RC gradient value was also used to evaluate the possible correlation between the RC and clinical scores. Significance was assumed at 5% for all statistical analyses. 3. Results 3.1. Clinical scores and recordable MEP responses Data of clinical scores and TMS measures are shown in Tables 2 and 3. A gradual improvement in clinical scores T1 to T3 was observed in all patients (Fig. 1): NIHSS (one-way ANOVA: F2,46 = 41.116, P < 0.001, post hoc: T1 vs. T2, P < 0.001; T2 vs. T3,

Table 5 Correlation coefficients between longitudinal changes of TMS measures and clinical scores from T1 to T2 (T2 – T1) or from T1 to T3 (T3 – T1). T2 – T1

rMT(n = 24) aMT(n = 24) RC110%(n = 11) RC120%(n = 11) RC130%(n = 11) RC140%(n = 11) RC gradient (n = 11) SICI 2 ms(n = 13) SICI 3 ms(n = 13) SICI 4 ms(n = 13) ICF 10 ms(n = 13) ICF 15 ms(n = 13) cSP (n = 14) iSP (n = 14)

r p r p r p r p r p r p r p r p r p r p r p r p r p r p

T3 – T1

NIHSS

MRS

HG

JHFT

Fugl-Meyer

NIHSS

MRS

HG

JHFT

Fugl-Meyer

.013 (.951) .085 (.693) .276 (.412) .397 (.226) .449 (.166) .360 (.277) .575 (.064) .431 (.142) .603 (.029) .575 (.040) .109 (.723) .310 (.302) .199 (.495) .351 (.219)

.236 (.266) .361 (.083) .581 (.061) .452 (.163) .387 (.239) .129 (.705) .387 (.239) .022 (.942) .134 (.663) .178 (.560) .089 (.772) .223 (.465) .130 (.658) .152 (.604)

.039 (.857) .128 (.550) .464 (.151) .373 (.259) .027 (.937) .027 (.937) .064 (.853) .479 (.098) .016 (.957) .000 (1.000) .132 (.668) .467 (.108) .238 (.412) .404 (.152)

.084 (.698) .188 (.380) .229 (.499) .476 (.139) .041 (.904) .059 (.862) .279 (.406) .008 (.979) .437 (.135) .365 (.220) .053 (.865) .266 (.381) .475 (.086) .062 (.832)

.001 (.995) .085 (.694) .484 (.131) .553 (.078) .132 (.698) .201 (.554) .498 (.119) .240 (.430) .138 (.654) .240 (.430) .085 (.781) .212 (.487) .165 (.574) .697 (.006)

.028 (.897) .326 (.121) .513 (.106) .226 (.504) .669 (.024) .753 (.007) .235 (.486) .023 (.941) .424 (.149) .286 (.343) .318 (.289) .435 (.137) .288 (.318) .199 (.496)

.126 (.556) .310 (.140) .294 (.380) .392 (.233) .691 (.018) .530 (.094) .544 (.083) .325 (.278) .216 (.479) .391 (.186) .216 (.479) .301 (.317) .135 (.645) .103 (.727)

.126 (.558) .309 (.142) .118 (.729) .045 (.894) .018 (.958) .409 (.212) .273 (.417) .181 (.553) .033 (.915) .096 (.754) .495 (.086) .170 (.578) .301 (.296) .424 (.131)

.223 (.296) .351 (.093) .268 (.425) .163 (.632) .436 (.180) .072 (.834) .388 (.238) .356 (.233) .006 (.985) .191 (.532) .492 (.087) .336 (.262) .491 (.075) .534 (.049)

.079 (.712) .033 (.879) .435 (.181) .284 (.398) .192 (.571) .197 (.562) .540 (.086) .425 (.148) .030 (.922) .262 (.386) .074 (.809) .055 (.858) .254 (.382) .739 (.003)

NIHSS indicates National Institutes of Health Stroke Scale; MRS, Modified Rankin scale; HG, Hand grasping power; JHFT, Jebsen-Taylor Hand Function TestNote: rMT, aMT, cSP, RC, SICI, and ICF are values in the affected hemispheres. iSP are values from the unaffected hemisphere. Correlation coefficients (r) are shown for the plots of each TMS measure against the clinical scores shown. P values are shown in brackets. Significant correlations are shown in bold.


7


a NIHSS

b MRS

14

5

**

**

**

**

4

10

scores

scores

12

8 6

3 2

4 1

2

0

0 T1

T2

T1

T3

T3

d JHFT

25

e Fugl-Meyer

20

**

20

70

** **

**

60

**

15

**

15

scores

50

scores


c Hand grasping

T2

10

10

30 20

5

5

40

10 0

0

0 T1

T2

T3

T1

T2

T3

T1

T2

T3

Fig. 1. Longitudinal changes of clinical scores (mean ± SE) for all patients. There were significant improvements in all clinical scores. Asterisks indicate significant differences among the three time-points (Bonferroni, ⁄⁄P < 0.01).

P < 0.001), MRS (one-way ANOVA: F2,46 = 129.786, P < 0.001, post hoc: T1 vs. T2, P < 0.001; T2 vs. T3, P < 0.001), hand-grasping power (one-way ANOVA: F2,46 = 21.621, P < 0.001, post hoc: T1 vs. T3, P < 0.001; T2 vs. T3, P = 0.001), JHFT (one-way ANOVA: F2,46 = 28.002, P < 0.001, post hoc: T1 vs. T2, P < 0.001; T2 vs. T3, P < 0.001), and Fugl-Meyer test (one-way ANOVA: F2,46 = 31.769, P < 0.001, post hoc: T1 vs. T2, P < 0.001; T2 vs. T3, P < 0.01). At T1, MEPs could be recorded from the affected hand by stimulation of the AH in only 14 patients both at rest and during voluntary contraction. This was the case even when the maximum stimulator output was used. In contrast, clear MEPs could be recorded from the unaffected hand by stimulation of the UH in all patients (Table 3). Importantly, the number of patients with recordable MEPs from the affected hand gradually increased from T1 to T3, and at T3, MEPs could be recorded from the affected hand during voluntary contraction in all patients (Table 3). A comparison of longitudinal changes in clinical scores (Fig. 2) between patients with MEPs (n = 14) vs. patients without MEPs (n = 10) recorded from the affected at T1, showed significant interactions between ‘‘patient group’’ ‘‘time’’ in both hand grasping power (mixed ANOVA: F2,44 = 6.572, P = 0.003) and JHFT (mixed ANOVA: F2,44 = 8.999, P = 0.003) (Fig. 2c and d). This was because hand grasping power and JHFT scores gradually improved from T1 to T3 in the 14 patients with MEPs (one-way ANOVA: hand grasping; F2,26 = 30.759, P < 0.001, JHFT; F2,26 = 52.048, P < 0.001), but not in the 10 patients without MEPs (Fig. 2c and d). In contrast, no significant interactions between ‘‘patient group’’ ‘‘time’’ were observed for NIHSS, MRS, and Fugl–Meyer test (Fig. 2a, b, e). This was because both patients with and without MEPs showed improvement in these three clinical scores with time.

3.2. Motor threshold There were significant ‘‘time’’ ‘‘hemisphere’’ interactions for both rMT (n = 24, two-way ANOVA: F2,46 = 20.762, P < 0.001) (Fig. 3a and c) and aMT (n = 24, two-way ANOVA: F2,46 = 18.495, P < 0.001) (Fig. 3b and d). This was because at T1, both MTs were significantly elevated in the AH (unpaired-t test: rMT, P = 0.001; aMT, P = 0.001) (Fig. 3a and b) but were marginally reduced in the UH (unpaired-t test: rMT, P = 0.015; aMT, P = 0.057) (Fig. 3c and d), compared with the control value. Both MTs for the AH normalized at T3, and rMT for the UH normalized at T2 (n = 24, AH: rMT; one-way ANOVA, F2,46 = 11.465, P = 0.001, aMT; one-way ANOVA: F2,46 = 11.937, P = 0.001, UH: rMT; one-way ANOVA, F2,46 = 16.267, P < 0.001, aMT; one-way ANOVA: F2,46 = 9.577, P = 0.002) (Fig. 3a–d). Furthermore, both MTs were significantly higher for the AH than for the UH at T1 and T2 (n = 24, Bonferroni: rMT: T1, P < 0.001; T2, P = 0.005; aMT: T1, P < 0.001; T2, P = 0.006), but not at T3 (rMT: P = 0.064, aMT: P = 0.091). 3.3. Recruitment curve (RC) The RC could be made for all five stimulus intensities and the RC gradient could be calculated for the UH in 23 patients and for the AH in 11 patients from T1 through to T3 (Table 3); in the other patients, one or more stimulus intensities were above the maximum intensity of the stimulator. The 11 patients for whom the RC could be obtained from both hemispheres showed no significant ‘‘intensity’’ ‘‘hemisphere’’ interactions at any of the time-points examined (Fig. 4a, b, c). Additionally, the interaction between ‘‘time’’ ‘‘intensity’’ did not reach significance across the three


8


a NIHSS 14

b MRS **

*

**

5

**

12 4

scores

scores

10 8 6

3 2

4 1

2

**

0 T1

** T2

10 paents with 㧪㧝㧜㧜㧑 14 paents with 㧨㧝㧜㧜㧑

T3

T1

T2

d JHFT

*

**

absent MEPs at T1 present MEPs at T1

T3

e Fugl-Meyer **

20

**

70

**

**

60 20

15

15

scores

50

scores


**

0

c Hand grasping 25

**

10

10

40

**

**

30 20

5 5

10 0

0 T1

T2

T3

0 T1

T2

T3

T1

T2

T3

Fig. 2. Comparison of the longitudinal changes in clinical scores (mean ± SE) between two patient groups: 14 patients with vs. 10 patients without MEPs obtained from the affected hand by stimulation of the AH at T1. There were significant interactions between ‘‘patient group’’ ‘‘time’’ for both hand grasping power (c) (mixed ANOVA: F2,44 = 8.999, P = 0.003) and JHFT scores (mixed ANOVA: F2,44 = 6.572, P = 0.003) (d): JHFT scores and hand grasping power significantly improved from T1 to T3 in the 14 patients with MEPs, whereas neither significantly improved from T1 through to T3 in the 10 patients without MEPs. In contrast, there were no significant interactions between ‘‘patient group’’ ‘‘time’’ for NIHSS (a), MRS (b), and Fugl-Meyer test scores (e): for both patient groups, these three clinical scores significantly improved with time. Asterisks indicate significant differences among the three time-points (Bonferroni, ⁄P < 0.05, ⁄⁄P < 0.01).

time points for both the AH (Fig. 4a, b, c) and UH (Fig. 4d). However, compared with control values, patients affected by subcortical stroke showed a significant ‘‘intensity’’ ‘‘group’’ interaction in the UH at T1 (mixed ANOVA: F4, 176 = 3.744, P = 0.036), but not at T2 and T3 (Fig. 4d). This was because at T1, the RC gradient in the UH of patients was steeper than the gradient for controls (Table 3, Fig. 4d). In the AH, there were no significant interactions at any of the time-points examined. In two patients, the RC for the AH could not be obtained with the five stimulus intensities at T1, owing to raised MTs, but it could be obtained at T2 and T3. In these two patients, the RC in the UH was much steeper than that in the AH at both T2 and T3 (Fig. 4e and f). 3.4. Paired-pulse TMS Paired-pulse TMS was successfully performed over the UH of all 24 patients at T1 through to T3. However, paired-pulse TMS over the AH was successfully performed in only 13 patients through T1 to T3 because MEPs of 0.5 mV could not be evoked in other patients owing to the elevated MTs. The amplitude of control MEPs evoked by TS alone at T1 was 0.50 ± 0.06 mV (mean ± SE) in the AH and 0.58 ± 0.03 mV in the UH. Control MEP amplitudes of the AH and UH did not significantly differ from T1 through to T3. There was no significant interaction between ‘‘ISI’’ ‘‘hemisphere’’ at any time-point for the 13 patients in which paired-pulse TMS was successfully performed. However, there was a significant ‘‘time’’ ‘‘ISI’’ interaction (n = 24, two-way repeated ANOVA: F8,184 = 2.207, P = 0.029) in the UH (Fig. 5b), but not in the AH (Fig. 5a). A mixed ANOVA revealed significant interactions only at T1 in both the UH and AH (mixed-ANOVA: UH; T1,

F4,188 = 5.027, P = 0.003, post hoc Bonferroni: ISI 2 ms, P = 0.003; mixed-ANOVA: AH; T1, F4,144 = 4.178, P = 0.011, post hoc Bonferroni: ISI 2 ms, P = 0.006): SICI was significantly decreased only at an ISI of 2 ms in both the UH and AH at T1, whereas at T2 and T3, SICI in both the UH and AH was similar to or closer to that of values seen in controls (Fig. 5a and b). There was no significant change in the ICF at any time point in both the UH and AH (Fig. 5a and b).

3.5. Contralateral silent period We observed a significant ‘‘time’’ ‘‘hemisphere’’ interaction in the 14 patients for whom cSPs could be obtained from both hemispheres (n = 14, two-way ANOVA: F2,26 = 7.567, P = 0.013, Fig. 6a). This could be explained by the fact that in the AH, cSPs were prolonged at T1 and gradually became less sustained with time (n = 14, one-way ANOVA: F2,26 = 7.469, P = 0.012), whereas in the UH, cSPs were slightly shortened at T1 (significant in 24 patients [Table 3] but not significant in 14 [Fig. 6a], compared with the control value, n = 14, one-way ANOVA: F2,26 = 4.839, P = 0.016, Fig. 6a, n = 24, one-way ANOVA: F2,46 = 10.659, P < 0.001, Table 3). The duration of the cSP was significantly longer in the AH than in the UH at all time-points (n = 14, post hoc Bonferroni: T1, P = 0.002; T2, P < 0.001; T3, P = 0.029). In addition, compared to the control value, the cSP for the AH was significantly prolonged at all timepoints (n = 14, unpaired t-test: T1, P < 0.001; T2, P < 0.001; T3, P = 0.042), whereas the cSP for the UH was shortened at T1 (significant for 24 patients, unpaired t-test: T1, P = 0.033 [Table 3] but not significant in 14 patients [Fig. 6a]) but not at T2 and T3.


9


rMT in AH (n=24)

**

100

**

b

**

100

** Motor threshold (%)

Motor threshold (%)

a

aMT in AH (n=24)

*

50

*

*

** * 50

control 0

0 control

T1

T2

T3

T1 control

T1

T2

T3 T2 T3

rMT in UH (n=24)

d 100

100

**

Motor threshold (%)

Motor threshold (%)

c

aMT in UH (n=24)

**

*

50

*

**

50

0

0 control

T1

T2

T3

control

T1

T2

T3

Fig. 3. Results (mean ± SE) of resting (rMT) and active (aMT) motor thresholds for the control group and at three time-points for the affected (AH) and unaffected (UH) hemispheres of all 24 patients. If no MEP could be evoked at maximum stimulator output, then the threshold was described as 100%. The graphs show group mean values for rMT (a) and aMT (b) in the AH and rMT (c) and aMT (d) in the UH. For both rMT and aMT, there were significant ‘‘hemisphere’’ ‘‘time’’ interactions (two-way ANOVA: rMT; F2,46 = 20.762, P < 0.001, aMT; F2,46 = 18.495, P < 0.001). At T1, MTs were significantly elevated in the AH (a, b) but were reduced in the UH (c, d) (marginally [P = 0.057] for aMT): both MTs in the AH normalized at T3 (a, b), and rMT in the UH normalized at T2 (c). Asterisks indicate significant differences among the three time-points within patients (a, b, c, d), and asterisks just above the solid and striped bars (T1 or T2) indicate significant differences between control and patient groups (a, b, c) (paired or unpaired t-tests, ⁄P < 0.05, ⁄⁄P < 0.01).

3.6. Ipsilateral silent period An EMG example of the iSP in the UH from a single patient is shown in Fig. 7; the iSP was prolonged at T1, but became shorter at T2 and T3 in the UH (Fig. 7). In the 14 patients for whom the iSP could be obtained from both hemispheres, there was a significant ‘‘time’’ ‘‘hemisphere’’ interaction (n = 14, two-way ANOVA: F2,26 = 10.190, P = 0.001, Fig. 6b). This significant interaction could be explained by the fact that iSPs in the UH were prolonged at T1 and progressively became shorter at T2 and T3 (n = 14, oneway ANOVA: F2, 26 = 10.976, P < 0.001, Fig. 6b), whereas in the AH, iSPs did not significantly change from T1 to T3. The duration of the iSP was significantly longer in the UH than in the AH at T1 (n = 14, post hoc Bonferroni: T1, P = 0.001, Fig. 6b); however, the difference was not significant at either T2 or T3. Compared with the control value, iSPs in the UH were significantly prolonged at T1 and T2 (n = 14, unpaired t-test: T1, P < 0.001; T2, P = 0.011, Fig. 6b), but not at T3. In the AH, iSPs were significantly longer in the patient group than in the control group at all time-points (n = 24, unpaired t-test: T1, P = 0.003; T2, P < 0.001; T3, P = 0.014 [Table 3]; however, this finding was significant in the 14 patients only at T2, Fig. 6b). 3.7. Correlations between TMS measures and clinical scores at each time-point (Table 4) Significant correlations were observed between MTs for the AH and all clinical scores from T1 through to T3. In addition, many

significant correlations were observed between iSPs in the UH and all clinical scores at T1 or T2. These findings suggest that more severely impaired patients possessed longer iSPs in the UH at T1 or T2. Among other measures, the cSP in the AH, SICI (3 ms), ICF (10 ms, 15 ms), and RC (130%) were correlated with a number of clinical scores at a few time-points (Table 4). 3.8. Correlations between both longitudinal changes of TMS measures and clinical scores from T1 to T2 or from T1 to T3 (Table 5) Significant negative correlations were observed between changes in the iSP from the UH and upper limb function (JHFT and Fugl–Meyer test) from T1 to T2 and from T1 to T3. A shortened iSP in such intervals was indicative of better recovery of upper limb function. Changes in some TMS measures such as SICI (3 ms, 4 ms) and RC (130% and 140%) were correlated with changes in more global clinical scores (NIHSS or MRC) (Table 5). 3.9. Association between TMS measures at T1 or T2 and clinical scores at T2 or T3 (Supplementary Tables S1, S2 and S3) With simple linear regression analysis, significant relationships were found between MTs at T1 and all clinical scores at T2 or T3, and MTs at T2 and all clinical scores at T3 (R2 = 0.212–0.659, P < 0.05) (Supplementary Table S1, S2 and S3). Higher MTs at T1 or T2 predicted a worse prognosis of subsequent clinical recovery. The following is a breakdown of the association between TMS measures and time-points: ICF 15 ms


10


T1 (n=11)

3

2

1

0

3

2

1

110%

120%

130%

140%

d 4

MEP amplitude (mV)

control T1 T2 T3

2

1

120%

130%

0 120%

130%

1

100%

4 3 2

140%

130%

140%

T3 ( n=2 ) AH UH

5 4 3 2 1 0

100%

Smulus intensity ( % of rest MT )

120%

6

AH UH

5

110%


f

T2 (n=2)

0 110%

2

140%

1

100%

3


6

3

110%

e

UH (n=23)

AH UH

0 100%


MEP amplitude (mV)

AH UH

0 100%

T3 (n=11)

c4 MEP amplitude (mV)

AH UH

MEP amplitude (mV)

MEP amplitude (mV)

4

T2 (n=11)

b4

MEP amplitude (mV)

a

110%

120%

130%

140%

100%


110%

120%

130%

140%


Fig. 4. Results (mean ± SE) for the recruitment curve (RC). For only 11 patients, the RC could be made with all five stimulus intensities from both hemispheres at T1 through T3. (a), (b), (c) show comparisons of the RC between the AH and UH at T1 (a), T2 (b), and T3 (c) in these 11 patients. There were no significant ‘‘intensity’’ ‘‘hemisphere’’ interactions at any of the time-points investigated; the slope of the RC was not significantly different between the UH and AH at T1 (a), T2 (b), and T3 (c). (d) shows a comparison of the RC between the control group and the UH of 23 patients for whom the RC could be made with all five stimulus intensities, at three time-points. The interaction between ‘‘time’’ ‘‘intensity’’ did not reach significance across the three time-points for the UH of patients. However, compared with the control value, the RC of the UH showed a significant ‘‘intensity’’ ‘‘group’’ interaction at T1 (mixed ANOVA: F4, 176 = 3.744, P = 0.036) but not at T2 and T3; only at T1, the RC slope of the UH was steeper than the control value (d). (e), (f) show comparisons of the RC between both hemispheres at T2 (e) and T3 (f) in two patients for whom the RC for the AH could be obtained with all five stimulus intensities at T2 and T3, but not at T1 owing to the raised motor thresholds. In these two patients, the RC of the UH was much steeper than that of the AH at T2 (e) and T3 (f).

a 2002

2002

** 1001 control T1 T2

Response size (% control MEP)

Response size (% control MEP)

UH (n=24)

b

AH (n=13)

** 1

100 control T1 T2 T3

T3 0

0 2ms

3ms

4ms

10ms

15ms

Intersmulus interval ( ms )

2ms

3ms

4ms

10ms

15ms

Intersmulus interval ( ms )

Fig. 5. Results (mean ± SE) of SICI (ISI = 2, 3, 4 ms) and ICF (ISI = 10, 15 ms) produced by paired-pulse TMS in the control group and at the three time-points in the AH (n = 13) (a) and UH (n = 24) (b) of patients. There was a significant ‘‘time’’ ‘‘ISI’’ interaction (two-way repeated ANOVA: F8,184 = 2.207, P = 0.029) for the UH (b), but not for the AH (a). Compared with the control values, a mixed ANOVA revealed significant interactions only at T1 in both the UH and AH (mixed-ANOVA: UH; T1, F4,188 = 5.027, P = 0.003, post hoc Bonferroni: ISI 2 ms; P = 0.003, mixed-ANOVA: AH; T1, F4,144 = 4.178, P = 0.011, post hoc Bonferroni: ISI 2 ms; P = 0.006): SICI was significantly decreased only at ISI of 2 ms in both the UH and AH at T1, whereas at T2 and T3, SICI in both hemispheres was similar to or closer to that of the control value (a, b). There was no significant change in ICF at any time-point for both the UH and AH (a, b). Asterisks indicate significant differences at 2 ms between the control group and patients at T1 (Bonferroni, ⁄⁄P < 0.01).

at T1 and HG at T3 (R2 = 0.340, P = 0.034) (Supplementary Table S2), cSP at T2 and HG (R2 = 0.517, P = 0.004) or JHFT at T3 (R2 = 0.520, P = 0.004), iSP in the UH at T2 (R2 = 0.489, P = 0.005) and NIHSS or FM at T3 (R2 = 0.629, P = 0.001) (Supplementary

Table S3). These findings suggest that a prolonged iSP in the UH at T1 did not necessarily predict a worse prognosis in terms of clinical recovery; however, a protracted iSP in the UH at T2 predicted worse prognosis of NIHSS and FM at T3.



3.10. Differences between hemorrhagic and ischemic stroke There were no significant differences between hemorrhagic and ischemic stroke in days after stroke-onset and clinical scores. In addition, there was no significant interaction between ‘‘stroke type’’ ‘‘time’’ in all TMS measures of each hemisphere, such as rMT, aMT, SICI/ICF, cSP, and iSP. Finally, there was no significant interaction between ‘‘stroke type’’ ‘‘time’’ in all clinical scores, such as NIHSS, MRS, JHFT, hand grasping power, and Fugl–Meyer test. 4. Discussion In the current study, we focused on the effect of subcortical stroke lesions (regardless of hemorrhagic vs. ischemic etiologies) on longitudinal changes in physiological and clinical measures. The time intervals between T1 and T2 varied across patients and this may be a limitation of this study since the recovery of motor impairment is a function of time. Therefore, this study cannot evaluate the recovery of physiological and clinical measures as a function of time within six months after the stroke. However, we were able to measure at roughly the same time for all patients between T1 and T3; therefore, we were able to assess the physiological and the clinical measures and their correlations as a function of time, for two time points (T1 and T3). The present study demonstrated continuing clinical improvement both during and after inpatient rehabilitation service administration following subcortical stroke (Fig. 1). Interestingly, over a 1-year period, global clinical scores (NIHSS and MRC) of both patients with and without MEPs obtained from the affected hand at the post-acute phase improved with time, whereas scores for specific upper limb function (hand grasping power and JHFT) improved to a greater extent in patients with MEPs than in patients without MEPs (Fig. 2c and d). This suggests that having a recordable MEP from the affected hand at the post-acute phase is an indicator for a better prognosis of the affected upper limb function after a 1-year period. The present study also demonstrated dynamic changes of cortical physiology and, most importantly, this study was the first to assess longitudinal changes of transcallosal inhibition. We assert the importance of the change of each physiological measure over the one-year period in terms of how a motor output to the paretic hand is generated. 4.1. Corticospinal excitability As expected, at T1 both resting and active MTs were higher in the AH than MTs obtained from the UH and the control group. While both rMTs and aMTs in the AH were still higher at T2, they decreased with time and had normalized by 1 year. These data demonstrate that MTs can recover even in chronic phases. More surprisingly, rMTs in the UH at T1 was significantly decreased compared to the control value and returned to normal at T2. To our knowledge, no previous report has shown decreased MTs in the UH. Thus, our data are the first to reveal corticospinal hyperexcitability of the UH during the post-acute phase. In the 11 patients for whom the RC could be generated using all five intensities from both hemispheres, the RC did not differ between the two hemispheres and did not change significantly from T1 though to T3. This suggests that the RC was not affected in patients with only mild impairments. However, the clear hemispheric difference of the RC in two patients with increased MTs (Fig. 4e and f) suggests that in patients with more severe impairments, corticospinal hypoexcitability of the AH tends to persist over at least a 1year period. As previously mentioned, this may be associated with

11

poor prognosis of affected upper limb function in patients without MEPs (Fig. 2c and d). Interestingly, at T1 the RC for the UH was steeper than it was for control values (see Table 3 and Fig. 4d). This may constitute additional evidence of corticospinal hyperexcitability of the UH during the post-acute phase. 4.2. Intracortical excitability In contrast to the ICF which remains consistently normal, the SICI has been reported to be reduced in the AH as well as in the UH, especially during the acute phase following both cortical and subcortical stroke (Bütefisch et al., 2003, 2008; Manganotti et al., 2008), and to our knowledge, only one study has reported a converse finding in the AH (Wittenberg et al., 2007). There is little information regarding SICI changes over time in the chronic stage. Our data are in line with most previous reports; however, SICIs (ISI = 2 ms) for the AH tended to remain reduced for up to 1 year (Fig. 5a), whereas SICIs (ISI = 2 ms) for the UH had normalized immediately after the intensive rehabilitation period (Fig. 5b). Both the SICI and ICF are thought to occur due to activation of cortical circuits. The SICI is associated with GABAA activity, while the mechanisms mediating ICF are still unclear; one candidate neuron type possibly responsible for ICF is the glutamatergic interneuron (Ziemann et al., 1996; Chen, 2004; Ziemann, 2004; Di Lazzaro et al., 2005). It has been speculated that reduced activity of GABAergic transmission may promote long-term changes in synaptic efficacy (Hess and Donoghue, 1996). Thus, a reduced SICI may not only have an immediate role in facilitating motor outputs from the AH, but may also be of strategic importance in allowing functional reorganization in both hemispheres. Manganotti et al. (2008) reported reduced SICIs in both the AH and UH early after stroke. These authors found that in patients with good recovery, the reduced SICI in the UH returned to normal by 3 months, but remained abnormal in patients with poor recovery (Manganotti et al., 2008). Thus, a reduced SICI for the UH may be one of the initial motor system responses to the lesion in order to facilitate functional reorganization. One lingering issue in the field of stroke research is whether a reduced SICI is caused by the lesion, whether it reflects a compensatory mechanism as mentioned above, or whether it is a combination of both. Alternatively, the reduction in SICI could be epiphenomenal, or it could play both facilitative and inhibitory roles in recovery. Especially in the AH, the reduction in SICI could be caused by lesions developing in ascending modulatory systems of the M1 after subcortical stroke. If this is the case, slow normalization of the reduced SICI in the AH of our patients would just be reflective of the recovery process of lesions. Another possible reason for the reduced SICI in the AH is associated with the stimulation strength, which was higher for the AH than for the UH in order to obtain a similar MEP size (0.5 mV). Therefore, it is likely that this increased stimulation strength involved more cortical neuron excitation. Taken all of these issues into account, we cannot exclude the possibility that the reduced SICI in our patients was attributable to either one, or a combination, of these different mechanisms. A prolonged cSP for the AH has been demonstrated in the acute phase of subcortical stroke (Haug et al., 1992; Classen et al., 1997; Ahonen et al., 1998; Liepert et al., 2000, 2005), with a consistently normal cSP for the UH (Cicinelli et al., 1997; Liepert et al., 2000). Our results are in line with those of previous studies. The cSP is thought to be mediated by GABAB cortical neurons (Werhahn et al., 1999) and abnormalities in cSPs of our patients could partly reflect changes in the excitability of inhibitory neurons. However, the implication of changing cSPs should be considered cautiously because the cSP duration depends on the stimulus intensity (Kim-



a

b

Contralateral silent period (n=14)

300

**

Silent period (msec )

250 AH 200

**

Ipsilateral silent period (n=14)

**

*

**

**

60

*

UA

*

150

**

* **

50

** **

100 50

Silent period (msec )

12

AH 40

UH

*

30

**

20 10

0

0

control control

T1

T2

T3

control control

T1

T2

T3

Fig. 6. Results (mean ± SE) of contralateral and ipsilateral silent periods (cSP and iSP) for the control group and at the three time-points for the AH and UH of patients. Both cSPs and iSPs could be obtained from both hemispheres for 14 patients. For the cSP in these 14 patients (a), there was a significant ‘‘time’’ ‘‘hemisphere’’ interaction (twoway ANOVA: F2,26 = 7.567, P = 0.013). Compared with the control value, cSPs for the AH were significantly prolonged at all time-points (a). For the iSP in these 14 patients (b), there was a significant ‘‘hemisphere’’ ‘‘time’’ interaction (two-way ANOVA: F2,26 = 10.190, P = 0.001). Compared with the control value, iSPs in the UH were significantly prolonged at T1 and T2 (b). At T2, iSPs in the AH were also significantly more prolonged in patients than they were in controls (b). Asterisks indicate significant differences among the three time-points or between the AH and UH within patients, and asterisks just above the solid and striped bars indicate significant differences between control and patient groups (paired or unpaired t-tests, ⁄P < 0.05, ⁄⁄P < 0.01).

smulaon

mV

T1 0 mV

T2 0 mV

0.5 mV

T3

20 msec

0 Fig. 7. An EMG example of iSPs in the UH from a single patient is shown. The waveforms are averages of 10 rectified trials. The dotted line represents the mean background EMG amplitude before the TMS stimulus. The duration of the iSP was prolonged at T1, but became shorter at T2 and T3.

iskidis et al., 2005). The intensity used was expressed as a percent of the MT, and any change in the cSP over time could then reflect MT changes. Therefore, cSP changes in our patients might have resulted from MT changes rather than from changes in inhibitory cortical neurons. 4.3. Transcallosal inhibition Transcallosal inhibition can be measured either as an iSP, demonstrated by single-pulse TMS during voluntary tonic contraction (Ferbert et al., 1992), or as inhibition of the MEP evoked by a

TMS pulse over the M1 of the contralateral hemisphere, using the paired-pulse TMS technique (IHI) (Ferbert et al., 1992; Chen et al., 2003). Both of these measures have been shown to be absent in some patients with callosal lesions or agenesis of the corpus callosum; thus, iSPs and IHI are believed to involve transcallosal pathways (Ferbert et al., 1992). IHI is thought to consist of two phases: an early phase with short ISIs of 7–10 ms (S-IHI) and a later phase with long ISIs >15 ms (L-IHI) (Talelli et al., 2006). Previous studies have suggested that S-IHI and L-ISI constitute physiologically distinct forms of IHI (Chen et al., 2003; Kukaswadia et al., 2005; Irlbacher et al., 2007; Ni et al., 2009) and similar circuits mediate L-IHI and iSP (Chen et al., 2003; Giovannelli et al., 2009). For example, target muscle activation decreases S-IHI but has little effect on LIHI (Chen et al., 2003). In addition, Murase et al. (2004) studied S-IHI at an ISI of 10 ms by paired-pulse TMS using a simple reaction time paradigm and found that closer to movement onset, S-IHI turned into facilitation in healthy volunteers. Thus, voluntary contraction influences S-IHI. In contrast, the strength of target muscle contractions has little effect on the iSP duration (Ferbert et al., 1992). Therefore, iSP measurements deal more directly than S-IHI with interhemispheric inhibition of voluntary motor cortical output (Giovannelli et al., 2009). Transcallosal inhibition from the UH to the AH is more difficult to assess because it requires recordable MEPs or voluntary EMG activity in the affected hand. Thus, it is possible to assess transcallosal inhibition from the UH to the AH only in patients with mild impairments of the affected hand. Our study showed that the duration of the iSP from the UH in the 14 patients that had mild impairments (the same 14 patients with recordable MEPs at T1) was much longer than those recorded from the AH or the control value at T1. Additionally, these iSP durations became shorter with time. It should also be noted that iSP duration increases with increasing stimulus intensity (Ferbert et al., 1992). Thus, if our results for the iSP in the UH at T1 only reflected the rMT change, iSPs should have been shortened because rMTs in the UH were lower than rMTs in controls and in the AH at T1 (Fig. 3). Therefore, the prolongation of the iSP for the UH in our patients does not reflect the MT changes, but it does reflect enhanced transcallosal inhibition from the UH to the AH. The duration of the iSP from the AH was longer than the control value, but this could result from the MT elevation



rather than the enhanced transcallosal inhibition. Thus, we cannot draw a conclusion from our data on whether excitability of transcallosal inhibition from the AH to the UH was pathologically changed, which could even be decreased from the normal level. However, we can conclude that transcallosal inhibition from the UH to the AH was pathologically enhanced and stronger than the inhibition in the opposite direction, leading to interhemispheric imbalance, especially during the post-acute phase. Previous studies have shown a consistently normal IHI from the AH to the UH in patients with lesions below the centrum semiovale, whereas in some cases of cortical stroke, inhibition can be reduced (Boroojerdi et al., 1996; Shimizu et al., 2002). Moreover, previous data suggest that the IHI from the UH to the AH is preserved if the lesion spares the transcallosal fibers (Bütefisch et al., 2008). Our data are consistent with these previous reports, and only one study (Bütefisch et al., 2008) has reported a reduction in IHI from the AH to the UH; however, this study also reported that the IHI from the UH to the AH was normal in patients with subcortical strokes (only five patients with slight weakness: their Motricity index (76–92) indicated a weakness level of MMT 4 or more) (Bütefisch et al., 2008). The discrepancy between this previous study and our data could be explained by the difference in the degree of weakness (more weakness 5.9–26.8 Kg of grasping power in our 14 patients with recordable iSPs, see Table 2) and the method for studying transcallosal inhibition (S-IHI vs. iSP). Additionally, the decrease in IHI from the AH to the UH reported in the Bütefisch’s study was mainly due to data collected from one patient (Bütefisch et al., 2008). Lastly, it may be difficult to detect abnormally increased IHIs (transcallosal inhibition) with the paired-pulse TMS method in the resting condition. This difficulty could be due to a floor effect, and if this were the case then the method used in our study (iSP measurement) would have rectified this issue. Nevertheless, our data are consistent with the theory of interhemispheric imbalance, since transcallosal inhibition from the UH to the AH was stronger than the inhibition recorded in the opposite direction. Taking the findings from the Bütefisch at al. study into account, if we were able to obtain iSPs from all patients, including those with severe impairments, then it is possible that iSPs might have become more prolonged (since normal IHIs were observed in patients with slight impairment in the Bütefisch at al. study vs. longer iSPs in our patients with more severe impairments). Additionally, the iSP may be more prolonged due to negative correlations with affected upper limb function at T1, as discussed below.

4.4. Clinical correlations Clinical correlations were found mainly for MTs and iSPs. The correlations of the upper limb function with MTs persisted for up to 1 year, gradually becoming weaker, which is in agreement with previous work (Manganotti et al., 2008; Swayne et al., 2008). The correlations for cSPs in the AH may result from MT changes as mentioned above, but more importantly, iSPs from the UH negatively correlated with affected upper limb function at T1. These correlations became less at T2 and disappeared at T3 (Table 4). In addition, shortened iSPs over an extended period of time were associated with better recovery (JHFT and Fugl-Meyer test) (Table 5). Finally, a simple linear regression analysis revealed that prolonged iSPs in the UH at T1 did not predict a worse prognosis of clinical recovery; however, as expected, elevated MTs were an indicator of a worse prognosis (Supplementary Tables S1 and S2). These results suggest that in the 14 patients with more mild impairment, enhanced transcallosal inhibition from the UH to the AH may have temporally interfered with motor function of the affected upper limb during the post-acute phase, and that motor

13

function gradually recovered as the enhanced transcallosal inhibition returned to normal levels. However, in 41% (n = 10) of the patients who had no MEP response from the affected hand to TMS at the post-acute phase, correlation studies between TMS measures (intracortical excitability and transcallosal inhibition) and clinical scores could not be explored. These severely impaired patients without MEPs had a worse prognosis of the affected upper limb function after a 1-year period (Fig. 2). Thus, it seems that the best correlations between TMS measures and clinical recovery occur in either the presence or absence of the MEP response. This is consistent with previous reports (Escudero et al., 1998; Pennisi et al., 1999) and is also associated with earlier findings that ipsilateral motor pathways from the UH to the paretic hand are more commonly detected in patients with poor motor recovery (Netz et al., 1997; Gerloff et al., 2006). Thus, in patients with severe motor impairments, we were unable to draw any conclusions as to whether abnormal transcallosal inhibition was associated with motor recovery. 4.5. Interhemispheric imbalance and motor recovery The reduced rMT and SICI in the UH observed in our study led to overall cortical hyperexcitability at the post-acute phase, which may have resulted in enhanced transcallosal inhibition on the AH in patients with milder impairments. In addition, the enhanced inhibition we observed could have interfered with recovery of corticospinal output from the AH. Recently, it has been suggested that interhemispheric competition after stroke is strongly related to motor function of the affected hand (Hummel et al., 2006; Nowak et al., 2009). For example, one of the previously mentioned studies used a simple reaction time paradigm to show that S-IHI at rest was comparable between patients and controls; however, closer to movement onset, S-IHI turned into facilitation in controls but not in patients (Murase et al., 2004). Interestingly, the amount of S-IHI was positively correlated with the severity of the motor impairment of the affected hand (Murase et al., 2004). This group suggested that enhanced S-IHI from the UH to the AH interfered with motor function in the affected hand. In addition, Takeuchi et al. (2005) demonstrated that 1-Hz rTMS (suppressive rTMS) over the UH was enough to reduce iSPs (transcallosal inhibition from the UH to the AH) and improved motor function of the affected hand of patients with chronic stroke. This group suggested that M1 in the UH inhibited M1 in the AH via an abnormally enhanced transcallosal inhibition and that a disruption of this abnormal transcallosal inhibition by 1-Hz rTMS was sufficient to induce improvements in motor function. Thus, one implication of interhemispheric imbalance after stroke is as follows (Nowak et al., 2009): deficits in transcallosal inhibition from the AH to the UH increases excitability in the UH, which exerts excessive transcallosal inhibition back the AH, leading to interference of motor function or recovery of the affected hand. Our data support this idea, in that abnormally enhanced transcallosal inhibition from the UH may have hampered motor recovery during the post-acute phase. Following this period, enhanced transcallosal inhibition gradually improved, leading to continuing clinical recovery. However, the idea of abnormal interhemispheric interactions and its related effects on the AH and UH is a controversial topic. For example, Bütefisch et al. (2008) reported that in sub-acute stroke patients, SICI was decreased in both the AH and UH regardless of cortical or subcortical infarction. Additionally, Bütefisch et al. (2008) reported that IHIs were abnormally decreased from the AH to the UH but that IHIs from the UH to the AH were normal. Interestingly, these changes occurred in patients with excellent recovery. Thus, the conclusion of this previous study was that a decreased SICI in the UH, which did not result in excessive IHI from the UH to the AH with suppression of M1 excitability in the AH,


14


may represent an adaptive process which supports recovery. One reason for the discrepancy between studies on interhemispheric interactions may depend on the difference of the measurement condition (resting vs active). Bütefisch et al. (2008) used S-IHIs in the resting condition, whereas other studies measured S-IHIs (Murase et al., 2004) or iSPs (Takeuchi et al., 2005) in the active condition. In the current study, we used the iSP measurement. Although the mechanisms underlying S-IHI and iSP are different (as mentioned above), it may be crucial to take measurements in the active condition in order to evaluate abnormal transcallosal inhibition of voluntary cortical motor output with movements of the paretic hand. The contribution of the M1 in the UH to motor recovery in patients with stroke is also controversial. Increased activity of M1 in the UH could not only be harmful, but could also be beneficial to recovery (Bütefisch et al., 2005; Gerloff et al., 2006; Lotze et al., 2006; Schaechter and Perdue, 2008). Previous studies performed on patients 1–6 weeks post-stroke have demonstrated that a reduced SICI in the UH is related to recovery (Manganotti et al., 2002; Bütefisch et al., 2003, 2008). In addition, a previous multimodal imaging study of well-recovered patients after capsular stroke concluded as follows; (1) effective recovery is based on enhanced utilization of resources of both the AH and UH, (2) basic corticospinal commands arise from the AH without recruiting ipsilateral motor pathways from the UH, and (3) increased UH activity probably facilitates control of recovered motor function by operating at a higher-order processing level (Gerloff et al., 2006). The present study demonstrated that enhanced transcallosal inhibition from the UH to the AH, which was associated with hyperexcitability of M1 in the UH, correlated with the severity of the motor impairment of the affected upper arm at the post-acute phase, but did not predict a poor recovery at the 1-year period. Thus, it is possible that enhanced transcallosal inhibition from the UH to the AH may have hampered the affected upper arm function temporarily at the post-acute phase, whereas the co-existing UH hyperexcitability facilitated motor recovery by operating at a higher-order processing level. Alternatively, the enhanced transcallosal inhibition could be the result of the extent of injury to the brain. The normalization over a 1-year period would reflect the recovery process of the lesions. Thus, it is difficult to definitively conclude that the enhanced transcallosal inhibition and the extent of clinical recovery were causally related in this study. Further studies on large patient groups with varying levels of impairment and at different stages are needed to clarify the role of the increased UH activity and interhemispheric imbalance on recovery following subcortical stroke. Conflict of interest All funding sources supporting this work are acknowledged. The authors will disclose to the editor any pertinent financial interests associated with the manufacture of any drug or product described in the manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.clinph.2014.01. 034. References Ahonen JP, Jehkonen M, Dastidar P, Molnar G, Hakkinen V. Cortical silent period evoked by transcranial magnetic stimulation in ischemic stroke. Electroencephalogr Clin Neurophysiol 1998;109:224–9.

Avanzino L, Teo JT, Rothwell JC. Intracortical circuits modulate transcallosal inhibition in humans. J Physiol 2007;583:99–114. Boroojerdi B, Diefenbach K, Ferbert A. Transcallosal inhibition in cortical and subcortical cerebral vascular lesions. J Neurol Sci 1996;144:160–70. Bütefisch CM, Netz J, Wessling M, Seitz RJ, Homberg V. Remote changes in cortical excitability after stroke. Brain 2003;126:470–81. Bütefisch CM, Kleiser R, Korber B, Muller K, Wittsack HJ, Homberg V, et al. Recruitment of contralesional motor cortex in stroke patients with recovery of hand function. Neurology 2005;64:1067–9. Bütefisch CM, Wessling M, Netz J, Seitz RJ, Homberg V. Relationship between interhemispheric inhibition and motor cortex excitability in subacute stroke patients. Neurorehabil Neural Repair 2008;22:4–21. Cao Y, D’Olhaberriague L, Vikingstad EM, Levine SR, Welch KM. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke 1998;29:112–22. Chen R, Tam A, Butefisch C, Corwell B, Ziemann U, Rothwell JC, et al. Intracortical inhibition and facilitation in different representations of the human motor cortex. J Neurophysiol 1998;80:2870–81. Chen R, Yung D, Li JY. Organization of ipsilateral excitatory and inhibitory pathways in the human motor cortex. J Neurophysiol 2003;89:1256–64. Chen R. Interactions between inhibitory and excitatory circuits in the human motor cortex. Exp Brain Res 2004;154:1–10. Chollet F, DiPiero V, Wise RJ, Brooks DJ, Dolan RJ, Frackowiak RS. The functional anatomy of motor recovery after stroke in humans: a study with positron emission tomography. Ann Neurol 1991;29:63–71. Cicinelli P, Traversa R, Rossini PM. Post-stroke reorganization of brain motor output to the hand: a 2–4 month follow-up with focal magnetic transcranial stimulation. Electroencephalogr Clin Neurophysiol 1997;105:438–50. Classen J, Schnitzler A, Binkofski F, Werhahn KJ, Kim YS, Kessler KR, et al. The motor syndrome associated with exaggerated inhibition within the primary motor cortex of patients with hemiparetic. Brain 1997;120:605–19. Cramer SC, Nelles G, Benson RR, Kaplan JD, Parker RA, Kwong KK, et al. A functional MRI study of subjects recovered from hemiparetic stroke. Stroke 1997;28:2518–27. Di Lazzaro V, Oliviero A, Saturno E, Dileone M, Pilato F, Nardone R, et al. Effects of lorazepam on short latency afferent inhibition and short latency intracortical inhibition in humans. J Physiol 2005;564:661–8. Escudero JV, Sancho J, Bautista D, Escudero M, Lopez-Trigo J. Prognostic value of motor evoked potential obtained by transcranial magnetic brain stimulation in motor function recovery in patients with acute ischemic stroke. Stroke 1998;29:1854–9. Ferbert A, Priori A, Rothwell JC, Day BL, Colebatch JG, Marsden CD. Interhemispheric inhibition of the human motor cortex. J Physiol 1992;453:525–46. Gerloff C, Bushara K, Sailer A, Wassermann EM, Chen R, Matsuoka T, et al. Multimodal imaging of brain reorganization in motor areas of the contralesional hemisphere of well recovered patients after capsular stroke. Brain 2006;129:791–808. Giovannelli F, Borgheresi A, Balestrieri F, Zaccara G, Viggiano MP, Cincotta M, et al. Modulation of interhemispheric inhibition by volitional motor activity: an ipsilateral silent period study. J Physiol 2009;587:5393–410. Gladstone DJ, Danells CJ, Black SE. The fugl-meyer assessment of motor recovery after stroke: a critical review of its measurement properties. Neurorehabil Neural Repair 2002;16:232–40. Haug BA, Schonle PW, Knobloch C, Kohne M. Silent period measurement revives as a valuable diagnostic tool with transcranial magnetic stimulation. Electroencephalogr Clin Neurophysiol 1992;85:158–60. Hess G, Donoghue JP. Long-term potentiation and long-term depression of horizontal connections in rat motor cortex. Acta Neurobiol Exp (Wars) 1996;56:397–405. Hummel FC, Voller B, Celnik P, Floel A, Giraux P, Gerloff C, et al. Effects of brain polarization on reaction times and pinch force in chronic stroke. BMC Neurosci 2006;7:73. Irlbacher K, Brocke J, Mechow JV, Brandt SA. Effects of GABA(A) and GABA(B) agonists on interhemispheric inhibition in man. Clin Neurophysiol 2007;118:308–16. Jebsen RH, Taylor N, Trieschmann RB, Trotter MJ, Howard LA. An objective and standardized test of hand function. Arch Phys Med Rehabil 1969;50:311–9. Kimiskidis VK, Papagiannopoulos S, Sotirakoglou K, Kazis DA, Kazis A, Mills KR. Silent period to transcranial magnetic stimulation: construction and properties of stimulus-response curves in healthy volunteers. Exp Brain Res 2005;163:21–31. Kujirai T, Caramia MD, Rothwell JC, Day BL, Thompson PD, Ferbert A, et al. Corticocortical inhibition in human motor cortex. J Physiol 1993;471:501–19. Kukaswadia S, Wagle-Shukla A, Morgante F, Gunraj C, Chen R. Interactions between long latency afferent inhibition and interhemispheric inhibitions in the human motor cortex. J Physiol 2005;563:915–24. Liepert J, Storch P, Fritsch A, Weiller C. Motor cortex disinhibition in acute stroke. Clin Neurophysiol 2000;111:671–6. Liepert J, Restemeyer C, Kucinski T, Zittel S, Weiller C. Motor strokes: the lesion location determines motor excitability changes. Stroke 2005;36:2648–53. Lotze M, Markert J, Sauseng P, Hoppe J, Plewnia C, Gerloff C. The role of multiple contralesional motor areas for complex hand movements after internal capsular lesion. J Neurosci 2006;26:6096–102. Lyden P, Brott T, Tilley B, Welch KM, Mascha EJ, Levine S, et al. Improved reliability of the NIH stroke scale using video training. NINDS TPA stroke study group. Stroke 1994;25:2220–6.


U. Takechi et al. / Clinical Neurophysiology xxx (2014) xxx–xxx Manganotti P, Patuzzo S, Cortese F, Palermo A, Smania N, Fiaschi A. Motor disinhibition in affected and unaffected hemisphere in the early period of recovery after stroke. Clin Neurophysiol 2002;113:936–43. Manganotti P, Acler M, Zanette GP, Smania N, Fiaschi A. Motor cortical disinhibition during early and late recovery after stroke. Neurorehabil Neural Repair 2008;22:396–403. Marshall RS, Perera GM, Lazar RM, Krakauer JW, Constantine RC, DeLaPaz RL. Evolution of cortical activation during recovery from corticospinal tract infarction. Stroke 2000;31:656–61. Matsunaga K, Maruyama A, Fujiwara T, Nakanishi R, Tsuji S, Rothwell JC. Increased corticospinal excitability after 5 Hz rTMS over the human supplementary motor area. J Physiol 2005;562:295–306. Meyer BU, Roricht S, Grafin von Einsiedel H, Kruggel F, Weindl A. Inhibitory and excitatory interhemispheric transfers between motor cortical areas in normal humans and patients with abnormalities of the corpus callosum. Brain 1995;118:429–40. Murase N, Duque J, Mazzocchio R, Cohen LG. Influence of interhemispheric interactions on motor function in chronic stroke. Ann Neurol 2004;55:400–9. Netz J, Lammers T, Homberg V. Reorganization of motor output in the non-affected hemisphere after stroke. Brain 1997;120:1579–86. Ni Z, Gunraj C, Nelson AJ, Yeh IJ, Castillo G, Hoque T, et al. Two phases of interhemispheric inhibition between motor related cortical areas and the primary motor cortex in human. Cereb Cortex 2009;19:1654–65. Nowak DA, Grefkes C, Ameli M, Fink GR. Interhemispheric competition after stroke: brain stimulation to enhance recovery of function of the affected hand. Neurorehabil Neural Repair 2009;23:641–56. Pennisi G, Rapisarda G, Bella R, Calabrese V, Maertens De Noordhout A, Delwaide PJ. Absence of response to early transcranial magnetic stimulation in ischemic stroke patients: prognostic value for hand motor recovery. Stroke 1999;30:2666–70. Schaechter JD, Perdue KL. Enhanced cortical activation in the contralesional hemisphere of chronic stroke patients in response to motor skill challenge. Cereb Cortex 2008;18:638–47. Shimizu T, Hosaki A, Hino T, Sato M, Komori T, Hirai S, et al. Motor cortical disinhibition in the unaffected hemisphere after unilateral cortical stroke. Brain 2002;125:1896–907. Swayne OB, Rothwell JC, Ward NS, Greenwood RJ. Stages of motor output reorganization after hemispheric stroke suggested by longitudinal studies of cortical physiology. Cereb Cortex 2008;18:1909–22.

15

Takeuchi N, Chuma T, Matsuo Y, Watanabe I, Ikoma K. Repetitive transcranial magnetic stimulation of contralesional primary motor cortex improves hand function after stroke. Stroke 2005;36:2681–6. Takeuchi N, Tada T, Toshima M, Ikoma K. Correlation of motor function with transcallosal and intracortical inhibition after stroke. J Rehabil Med 2010;42:962–6. Talelli P, Greenwood RJ, Rothwell JC. Arm function after stroke: neurophysiological correlates and recovery mechanisms assessed by transcranial magnetic stimulation. Clin Neurophysiol 2006;117:1641–59. Triggs WJ, Macdonell RA, Cros D, Chiappa KH, Shahani BT, Day BJ. Motor inhibition and excitation are independent effects of magnetic cortical stimulation. Ann Neurol 1992;32:345–51. Trompetto C, Bove M, Marinelli L, Avanzino L, Buccolieri A, Abbruzzese G. Suppression of the transcallosal motor output: a transcranial magnetic stimulation study in healthy subjects. Exp Brain Res 2004;158:133–40. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19:604–7. Ward NS, Brown MM, Thompson AJ, Frackowiak RS. Neural correlates of motor recovery after stroke: a longitudinal fMRI study. Brain 2003;126:2476–96. Ward NS, Newton JM, Swayne OB, Lee L, Thompson AJ, Greenwood RJ, et al. Motor system activation after subcortical stroke depends on corticospinal system integrity. Brain 2006;129:809–19. Werhahn KJ, Kunesch E, Noachtar S, Benecke R, Classen J. Differential effects on motorcortical inhibition induced by blockade of GABA uptake in humans. J Physiol 1999;517:591–7. Werhahn KJ, Conforto AB, Kadom N, Hallett M, Cohen LG. Contribution of the ipsilateral motor cortex to recovery after chronic stroke. Ann Neurol 2003;54:464–72. Wittenberg GF, Bastings EP, Fowlkes AM, Morgan TM, Good DC, Pons TP. Dynamic course of intracortical TMS paired-pulse responses during recovery of motor function after stroke. Neurorehabil Neural Repair 2007;21:568–73. Ziemann U, Lonnecker S, Steinhoff BJ, Paulus W. Effects of antiepileptic drugs on motor cortex excitability in humans: a transcranial magnetic stimulation study. Ann Neurol 1996;40:367–78. Ziemann U. TMS and drugs. Clin Neurophysiol 2004;115:1717–29.


Long-term effects of contralesional rTMS in severe stroke: safety, cortical excitability, and relationship with transcallosal motor fibers.

Transcallosal connectivity of the human cortical motor network.

Amyotrophic lateral sclerosis affects cortical and subcortical activity underlying motor inhibition and action monitoring.

Motor Recovery After Subcortical Stroke Depends on Modulation of Extant Motor Networks.

Contralesional Cortical Structural Reorganization Contributes to Motor Recovery after Sub-Cortical Stroke: A Longitudinal Voxel-Based Morphometry Study.

Compensatory motor network connectivity is associated with motor sequence learning after subcortical stroke.

Time-dependent changes in motor cortical excitability by electrical stimulation combined with voluntary drive.

Cortical excitability changes distinguish the motor neuron disease phenotypes from hereditary spastic paraplegia.

Cortical Excitability Measured with nTMS and MEG during Stroke Recovery.

Transcallosal effects of a cortical epileptiform focus.

Rapid Changes in Cortical and Subcortical Brain Regions after Early Bilateral Enucleation in the Mouse.

Correction: Rapid Changes in Cortical and Subcortical Brain Regions after Early Bilateral Enucleation in the Mouse.

Enhancement of Cortical Excitability and Lower Limb Motor Function in Patients With Stroke by Transcranial Direct Current Stimulation.

Cortical activation changes and improved motor function in stroke patients after focal spasticity therapy--an interventional study applying repeated fMRI.

Altered topological properties of the cortical motor-related network in patients with subcortical stroke revealed by graph theoretical analysis.

Changes of motor-cortical oscillations associated with motor learning.

Motor recovery and cortical plasticity after functional electrical stimulation in a rat model of focal stroke.

No changes in functional connectivity during motor recovery beyond 5 weeks after stroke; A longitudinal resting-state fMRI study.

Cortical excitability changes over time in progressive multiple sclerosis.

Motor Recovery of the Affected Hand in Subacute Stroke Correlates with Changes of Contralesional Cortical Hand Motor Representation.

Predicting Modulation in Corticomotor Excitability and in Transcallosal Inhibition in Response to Anodal Transcranial Direct Current Stimulation.

Cortical dynamics and subcortical signatures of motor-language coupling in Parkinson's disease.

Changes in corticomotor excitability and intracortical inhibition of the primary motor cortex forearm area induced by anodal tDCS.

Cumulative effects of anodal and priming cathodal tDCS on pegboard test performance and motor cortical excitability.