- Email: [email protected]
Control of hand movements after striatocapsular stroke: high-resolution temporal analysis of the function of ipsilateral activation
Clinical Neurophysiology 114 (2003) 1468–1476 www.elsevier.com/locate/clinph
Control of hand movements after striatocapsular stroke: high-resolution temporal analysis of the function of ipsilateral activation Rolf Verleger*, Sven Adam, Michael Rose1, Clemens Vollmer, Bernd Wauschkuhn, Detlef Ko¨mpf Department of Neurology, Medical University of Lu¨beck, Ratzeburger Allee 160, D 23538 Luebeck, Germany Accepted 15 April 2003
Abstract Objective: Hemiparesis due to infarction of the middle cerebral artery has become an increasingly important focus of research on cortical plasticity. Positron emission tomography and functional magnetic resonance imaging studies in such patients found involvement of the hemisphere ipsilateral to the affected hand related to movements of this hand. To understand the function of this ipsilateral activation, the present study investigated movement-related electroencephalogram (EEG) potentials in patients and healthy control subjects to measure timing of ipsi- and contralateral activation relative to movement onset. Methods: Thirteen patients were investigated in their chronic stage. Their pyramidal tracts were affected by infarctions of the middle cerebral artery at striatocapsular level. EEG potentials were recorded from 26 scalp electrodes while patients were pressing a key with their right or left index finger within a warned choice – response task. Results: Beginning 200 ms before responses of the affected hand, there was normal contralateral preponderance of EEG negativity. Briefly after response onset, however, the other unaffected hemisphere, ipsilateral to the responding hand, became additionally active. This pattern did not occur with responses made by the unaffected hand nor in healthy participants. Conclusions: The timing of the onset of ipsilateral activity precludes its role in response initiation. Rather, this activity may indicate reflex-like activation of the unaffected motor system to compensate for possible failure of the affected hand. q 2003 International Federation of Clinical Neurophysiology. Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Hemiparesis; Stroke; Event-related potentials; Cortical reorganization; Plasticity
1. Introduction The primary motor cortex is the principal unit of the brain for executing hand movements, based on input provided by other cortical, subcortical and cerebellar structures (Rizzolatti et al., 1998; Gilbert, 2001). In either hemisphere, output of these areas is sent via the pyramidal tract. Infarction (or hemorrhage) in the middle cerebral artery can impair this function, by damaging either the primary motor cortex or the subcortical pyramidal tract at the corona radiata and the capsula interna. The present study focussed on patients with such striatocapsular, subcortical damage. Under these circumstances, though in * Corresponding author. Tel.: þ49-451-500-2916; fax: þ 49-451-5002489. E-mail address: [email protected] (R. Verleger). 1 Now at: Department of Neurology, University Clinic of Hamburg, Hamburg, Germany.
principle unaffected, the cortical motor areas have to deal with the problem of blocked or restricted efferent conduction. Therefore, this condition has become a model case in research on cortical plasticity (Liepert and Weiller, 1999; Seitz and Azari, 1999; Chen et al., 2002). An issue of particular interest is the role taken by the motor cortex ipsilateral to the affected hand when patients move this hand. Evidence for ipsilateral involvement, in addition to contralateral activation, was reported in positron emission tomography (PET) (Weiller et al., 1993; Seitz et al., 1999) and functional magnetic resonance imaging (fMRI) studies (Cramer et al., 1997; Cao et al., 1998; Carey et al., 2002; but see Small et al., 2002) with a focus of activity in the ipsilateral premotor cortex (Cramer et al., 1999; Nelles et al., 1999; Seitz et al., 1999). What may be the function of such ipsilateral activation? One source of evidence is its long-term time-course, related to recovery. There have been reports both on increasing
1388-2457/03/$30.00 q 2003 International Federation of Clinical Neurophysiology. Published by Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S1388-2457(03)00125-1
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
(Nelles et al., 1999) and decreasing (Marshall et al., 2000; Carey et al., 2002) ipsilateral activation over time. Decrease of ipsilateral activation accompanying recovery would suggest that its continuing existence is maladaptive. Indeed, the size of patients’ electromyogram (EMG) activity on the affected hand evoked by transcranial magnetic stimulation of the ipsilateral motor cortex has been reported to be associated with poor recovery (Turton et al., 1996; Netz et al., 1997). Another source of evidence, exploited in the present study, is the precise timing of ipsilateral activity: Does it take place before, simultaneously with, or after contralateral activity? Measurement based on blood flow (PET and fMRI), powerful as they are for localizing activity, have not been able to provide answers on this question. What is needed is the detailed temporal resolution provided by electrophysiology. Therefore, multi-channel electroencephalogram (EEG) was measured in a task, in which participants had to produce responses with their right or left hands. Of major interest was the time-course of EEG potentials time-locked to response onset, to investigate whether any additional activity of the side ipsilateral to the affected hand occurred in the patients, more than when moving their unaffected hand, and more than the control group, and if so, how this ipsilateral activity relates in time both to contralateral activity and to the overt response. The earlier the ipsilateral activation occurs, both in relation to contralateral activity and to movement start, the more plausible would be its supportive function in initiating movements. Previous EEG studies in such patients did not exploit the EEG’s capacity for temporal resolution to this degree (Kitamura et al., 1996; Green et al., 1999; Kopp et al., 1999; Platz et al., 2000). Participants were cued to respond with their affected and unaffected hands (right or left hand, respectively) in random order over trials, and responses had to be produced with varying degrees of force, in order to enhance task demands to a level that would indeed be difficult to the patients and, therefore, would increase the likelihood that differences between control of both hands would be revealed. To control for differences in preparatory activity, EEG potentials were also measured in the interval between cue and imperative stimulus (IS).
2. Methods 2.1. Participants Thirteen patients and 8 age-matched healthy subjects participated. Patients were selected from the files of our department, by these criteria: an ischemic event should have taken place within the area supplied by the middle cerebral artery affecting the pyramidal tract at subcortical level; there should still be mild hemiparesis due to this event; there should be no or only very mild somatosensory deficits and no other ischemic event or additional neurological disease
1469
should have occurred. Clinical data of the patients are detailed in Table 1. Of the 13 patients, 11 had been affected by infarction, two by hemorrhage. The pyramidal tract was affected at corona radiata in 5 patients, capsula interna in 3 patients (including the two hemorrhagic patients), and in both areas in 5 patients. Additional areas affected were nucleus lentiformis ðn ¼ 4Þ, nucleus caudatus ðn ¼ 3Þ, putamen ðn ¼ 1Þ, and thalamus ðn ¼ 1Þ. These patients were in a chronic, stable state, the stroke dating back 5 years on average (20 months – 9 years). The left hemisphere was affected in 10 patients, the right hemisphere in 3 patients. Thus, since one right hemisphere patient was left handed before the lesion, the dominant hemisphere was affected in 11 patients, consecutively, 3 of these patients at least partially changed their hand dominance from the affected to the healthy side, as assessed by the Edinburgh Handedness Inventory, in spite of the mild degree of paresis. At the time of our study, the patients’ mean age was 65 years (57 – 71 years), and the mean degree of paresis of the affected hand as determined by standard neurological examination (Medical Research Council scale, extended by intermediate values) was 4– 5, i.e. very mild (ranging from 3– 4 to 4– 5). The 11 infarction patients received anticoagulating medication. The 8 control subjects were 4 men and 4 women, their mean age was 62 years (53 – 73 years), all were right handed. All participants gave their informed consent according to the declaration of Helsinki. They had normal or corrected-to-normal visual acuity and did not take medication known to affect the central nervous system. The study was approved by the ethical committee of the Medical University at Lu¨beck. 2.2. Procedure and stimuli Participants had to use their right or left hands with defined forces, randomly varying over trials, within defined temporal limits. Responses were made to an imperative signal which appeared 2 s after a cue that indicated hand and force. The devices for force measurement were two forcesensitive keys, one for each index finger. The keys were affixed at the front edge of the armrests of the experimental chair. Subjects’ forearms lay on the armrests, in order to prevent subjects from exerting additional force by moving their arms or their body. Either key was connected to an electronic force sensor whose voltage output increased proportionally to the pressure exerted on the key. This force output was visualized on the computer screen by two green bars, one for either hand, moving upwards from a zero position proportionally to the exerted force (see Verleger et al., 1999, for details). In addition to this immediate visual feedback, differential auditory feedback was provided for correct, incorrect, and premature responses. Responses were accepted as correct if force was exerted on the key by the indicated hand only. The window allowed for responding was 0 –1.9 s after the imperative signal.
1470
Table 1 Clinical data Age (years)
Sex
Infarct or hemorrhage
Lesion site
#1 #2 #3 #4
69 64 62 57
m m m f
infarct infarct infarct infarct
cr cr? cr cr and ci
#5 #6 #7 #8 #9 #10 #11 #12 #13
64 67 64 60 64 71 57 70 71
f m f m f m m f f
infarct infarct infarct hemorrhage infarct infarct hemorrhage infarct infarct
cr cr cr ci cr cr ci cr ci
Summary
65 (57–71)
7m, 6f
11 infarcts, 2 hemorrhage
5 cr, 3 ci, 5 cr and ci
Other lesion sites
nucl. caudatus? nucl. caudatus, nucl. lentiformis
and ci and ci and ci
putamen, nucl. lentiformis nucl. lentiformis thalamus
and ci nucl. caudatus, nucl. lentiformis 4 £ nucl. lentiformis, 3 £ nucl. caudatus, 1 £ putamen, 1 £ thalamus
Length £ width £ height of lesion (cm)
Time passed
Side of lesion
Hand preference before lesion
Affected hand
Hand preference changed
Degree of paresis
2 £ 1 £ 1.8 not known 1 £ 1 £ 1.6 4 £ 2 £ 3.0
1y8 3y6 5y8 6y2
left right right left
right right left right
right left left right
no – no no
4– 5 3– 4 4 4– 5
1 £ 1 £ 1.6 2 £ 1 £ 2.4 2 £ 2 £ 2.4 2.5 £ 4 £ 6.4 1.5 £ 1 £ 3.2 1.5 £ 1 £ 2.4 2 £ 2 £ 3.0 1.5 £ 1 £ 1.6 5 £ 2 £ 4.0
5y1m 4y8m 4y2m 4y0m 3y8m 4y5m 6y1m 8 y 10 m 4y2m
left left left left left left left left right
right right right right right right right right right
right right right right right right right right left
no yes yes no no no yes no –
4 4 4– 5 4– 5 4– 5 4– 5 4– 5 4– 5 4– 5
2.2 £ 1.6 £ 2.8
4y9m (1 y 8 m – 8 y 10 m)
10 left, 3 right
12 right, 1 left
10 right, 3 left
3 changes
4– 5 (3–4 to 4– 5)
m m m m
M, male; F, female; cr, corona radiata; ci, capsula interna; y, years; m, months. MRI (computed tomography (CT)) scans were made at the acute stage. Length and width of lesion were measured in the MRI (CT) slice with the largest extent of lesion; height is given by the number of slices in which the lesion was visible (Patient 2’s MRI scan got lost before this precise measurement could be performed. Details about his lesion sites were taken from his clinical files). Careful neurological examination was performed immediately before EEG recording to evaluate the degree of paresis and any signs of impairment not due to the known infarction. ‘Hand preference before lesion’ was asked in retrospect by applying the items of the Edinburgh Handedness Inventory. No entry was made in the column ‘preference changed’ if the affected hand was not the preferred one. Summary: means and ranges are indicated for ‘age’ and ‘time passed’, median and range for ‘degree of paresis’ (scale ranging from 0, plegia, to 5, no impairment), mean of each dimension separately (length, width, height) for length £ width £ height of lesion.
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
Code
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
The session consisted of 5 repetitions of two blocks, spontaneous force and graded force. Working through the entire session of 5 repetitions £ two blocks took about 75 min, plus about 10 min for instructions and practice at the beginning. For this report, data were pooled across 150 trials for either hand, from the unimanual spontaneous force and the 30 and 60% graded force trials, where participants had to press the key with the right or left hand, applying 100, 60, or 30% of their maximum convenient force, respectively. Subjects looked at the 17 inches screen from about 1.2 m. In each trial, two events occurred on the screen, cue and IS. The cue was the appearance of two blue bars, with the left bar assigned to the left hand and the right bar to the right hand. IS was the turn of the colour of the bars from blue to yellow, 2 s after cue onset. As soon as participants exerted force on the keys, the momentarily applied force was displayed by a green bar for either hand, of same size as the blue/yellow bars. Four seconds after IS, the next trial started with a new cue. In spontaneous force, the two blue bars appeared side by side. Simultaneously, an upward-pointing light-blue arrow appeared below the left, right, or both bars, forming the cue. Participants had to press the key(s) denoted by the arrow(s) briskly and forcefully when the blue bars had turned yellow (IS). The 3 tasks (right, left, both hands, 10 trials in each of the 5 repetitions across the session) alternated in random order. In graded force, two vertical scales appeared simultaneously with the blue bars, at the outer side of either bar. Either scale consisted of a vertical line separated in 4 equal parts by 5 horizontal tick marks at its outer side. Zero percent was marked at the lowest mark, 50% at the middle one, and 100% at the upper mark. The scale of 100%was the maximum force calculated from the preceding spontaneous force condition. The left and the right blue bar appeared at heights of 0, 30, or 60% of their corresponding scale, varying over trials, this way denoting which force had to be exerted after IS by which hand. Half the trials were unimanual tasks. Participants had to try and cover the yellow bar(s) by the corresponding green bar(s) after IS, by applying graded force. 2.3. Recording, data processing, and analysis For data analysis, all patients were considered to have a left-sided lesion and, therefore, right hand paresis. To this end, hand-force recordings and lateral EEG recording sites were reversed for the 3 patients who had the lesion in their right hemispheres. 2.4. EEG potentials EEG was recorded from 26 Ag/AgCl scalp electrodes (Fp1, Fp2, F3, Fz, F4, FC5, FC1, FCz, FC2, FC6, C3, C1, Cz, C2, C4, CP5, CP1, CP2, CP6, P7, P3, Pz, P4,
1471
P8, O1, O2) referred to an electrode at the nose tip. Vertical and horizontal electrooculogram (EOG) were recorded bipolarly from above vs. below the left eye and from the outer canthi of the eyes, for recognizing ocular artifacts. EEG and EOG were amplified (band pass 0.03– 35 Hz), digitized with 100 Hz per channel and digitally stored. Data were edited in several steps. First, trials with amplifier blocking (zero lines) were rejected from further analysis. Then, the impact of vertical EOG deflections on each EEG channel was estimated (by linear regression in epochs of large EOG activity) and these artifacts were removed from the EEG by subtraction (Verleger et al., 1982), this way keeping the affected epochs for further analysis. Third, in the same way, the impact of horizontal EOG deflections was estimated and removed (Anderer et al., 1992). Fourth, trials with other artifacts were rejected from further analysis (large amplitudes, fast shifts, and slow drifts). Finally, trials with premature, missing (. 1500 ms), or wrong side responses were rejected. The mean number (range) of remaining artifact-free and correctly responded unimanual trials was 85 (29 – 109) and 85 (40 – 112) for healthy (left) hand and affected (right) hand trials in the patients, and 114 (94 –128) and 119 (79 – 132) for left and right hand trials in the control group. The group difference was significant (F1;19 ¼ 13:8, P ¼ 0:001) reflecting both more erroneous responses and more EEG artifacts in the patients’ rejected data. These data were averaged separately for either hands and each participant in two ways, time-locked to the stimuli and response. Response-locked averages covered the time from 400 ms before response onset to 400 ms afterwards. Parameters quantified were mean amplitudes before response onset (2 220 to 2 180 ms), at response onset (2 20 to þ 20 ms), and after response onset (þ 160 to þ 200 ms). Stimuluslocked averages covered the entire 4 s recording of the trials from 100 ms before cue onset to 1.9 s after IS onset. Parameters quantified were contingent negative variation (CNV; mean amplitude 90 –0 ms before IS onset) and movement-accompanying negativity (MAN; most negative peak 350 – 1500 ms after IS). Time-courses of both response-locked and stimulus-locked potentials were analyzed mainly from C3 and C4 (the scalp sites overlying the motor cortex). Topographic maps were done of the scalp distribution of potentials for the mentioned parameters, using spherical spline interpolation. Analyses of variance (ANOVA) had the factors Group, Hand, and Recordings (two when analyzing C3 and C4, otherwise 22). For comparing contra- and ipsilateral topography, ANOVA had the factors Hemisphere (contralateral vs. ipsilateral) and recordings (10: F3, FC5, FC1, C3, C1, CP5, CP1, P7, P3, O1 on the left; F4, FC6, FC2, C4, C2, CP6, CP2, P8, P4, O2 on the right). When repeated-measurement factors had more than two levels, degrees of freedom were corrected by Huynh –Feldt’s 1 In this case, the corrected P values will be reported together with the original degrees of freedom.
1472
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
2.5. Force The voltage output of both force-sensitive keys was sampled with 100 Hz from 0.1 s before to 3.9 s after the cue (1.9 s after IS) and stored on hard disk. Data were converted to Newton by reference to calibration with defined weights and further analyzed with self-developed software. (Newton is the unit of force, 1 N is the force exerted by a mass of 102 g on its basis). First, 3 types of errors were separately counted, each calculated as number of error trials divided by the total number of trials in which the involved hand could commit the error: premature responses (i.e. before the imperative signal), missing responses, and wrong side. ‘Wrong side’ was counted for any pressure exceeding 1 N on the wrong key even if the stronger response was on the correct side. Four parameters of the force curves were determined separately for either hand in each errorless single trial and were averaged across trials. (1) Response time (RT): the time elapsed from IS onset until force increased above 1 N in spontaneous force and above 9% of maximum force in graded force. (2) Movement time (MT): the time from RT until force maximum in spontaneous force and until the moment of best fit in graded force. (3) Slope: the force difference between RT þ 200 ms and RT, to measure the initial impulse exerted on the keys. (4) Force: in spontaneous force, the amplitude of the force maximum and in graded force, the best fit, defined as the smallest deviation from target force, are expressed as percentage of maximum force. These parameters were analyzed by ANOVA with 3 factors, Group (patients vs. control group) and, as repeatedmeasurement factor, Hand (left vs. right ¼ healthy vs. affected in the patients).
3. Results 3.1. Behavior The mild degree of palsy of patients’ affected hands (range 3.5 –4.5) was well visible in the force measures. As Fig. 1 shows, (upper panels) maximum force was weaker when patients pressed the keys with the affected hand than with the unaffected hand (Hand £ Group: F1;19 ¼ 6:77, P ¼ 0:02; effect of Hand in the patients: F1;12 ¼ 4:23, P ¼ 0:06), the initial slope of force was less steep with this hand (Hand £ Group: F1;19 ¼ 10:69, P ¼ 0:004; effect of Hand in the patients: F1;12 ¼ 11:80, P ¼ 0:005), and (not shown by Fig. 1) deviation from target force was larger with this hand (Hand £ Group: F1;19 ¼ 7:33, P ¼ 0:01; effect of Hand in the patients: F1;12 ¼ 4:39, P ¼ 0:06). In contrast, temporal parameters of the response did not differentiate between groups, i.e. patients were neither reliably slower in initiating their response (RT) nor in executing their response (MT) (main effect of group:
Fig. 1. Response-locked EEG potentials. Grand means across all participants, time-locked to response onset. Depicted are response forces and EEG potentials, averaged separately for left and right hand trials. X-axis is in milliseconds relative to response onset. Upper panels: response force (gray denoting the affected right hand, black the unaffected left hand). Yaxis is in Newton. Second and third panels: EEG potentials recorded from sites overlying the left and right motor cortex, C3 (gray, affected left side) and C4 (black, unaffected right side), cf. the schematic head on the right. Yaxis is in microvolts (EEG), with negative values plotted upwards. Baseline for these data is the 100 ms interval before the cue, like in stimulus-locked potentials (Fig. 2). Lower panels: differences contralateral 2 ipsilateral. Black lines denote the C4 –C3 difference from key-presses with the unaffected hand (i.e. the differences between the C4 and C3 waveshapes in the third panels from above), gray lines denote the C3 –C4 difference from key-presses with the affected hand (i.e. the differences between the C3 and C4 waveshapes in the second panels from above). The arrows mark the time point when the contra-ipsilateral difference returned to zero, after movement onset in the patients’ recordings.
F1;19 ¼ 2:28, P ¼ 0:15 for RT; F1;19 ¼ 0:22, P ¼ 0:64 for MT). Nor was there a differential effect for the patients’ affected hand (Hand £ Group: F1;19 ¼ 0:05, P ¼ 0:83 for RT; F1;19 ¼ 1:43, P ¼ 0:25 for MT). Finally, there were some differences between groups in errors: patients had more responses missing ðRT . 1500 msÞ than the control group (9 vs. 4% on average), in particular with their affected hand (10% in the affected, 8% in the unaffected hand; Hand £ Group: F1;19 ¼ 6:76, P ¼ 0:02). Patients also pressed the key on the wrong side more often (6.5 vs. 2%; Group: F1;19 ¼ 4:97, P ¼ 0:04) without reliable differences between hands. There was no group difference in the number of premature responses (about 2% of all trials). 3.2. Response-locked potentials The analysis of main interest was on the data time-locked to response onset, following the IS. Response-locked potentials recorded from scalp sites C3 and C4, situated above the left (affected) and right (healthy) motor cortex, are displayed in the middle panels of Fig. 1, as grand means of either group. Being referred to the epoch before the cue as their baseline, these response-locked waveshapes started on negative levels, more so in the control group than in the patients, due to the preceding larger CNV of the control group (Fig. 2 and analysis below). Of importance, for patients and control group alike, EEG became more
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
negative at the site overlying the contralateral than the ipsilateral motor cortex about 200 ms before movement onset, with this contra-ipsilateral difference increasing until movement onset. This is a standard result (e.g. Kutas and Donchin, 1980; Ball et al., 1999; Kunieda et al., 2000). These differences between contra- and ipsilateral waveshapes are more clearly seen in the difference waveshapes in the lower panels of Fig. 1 (similar to the ‘lateralized readiness potential’, Coles, 1989, where additionally data from both hands are pooled). These differences appeared to be larger in patients than in the control group, but this was not significant. After movement onset, however, a specific feature emerged when patients were using their affected hand: the contra-ipsilateral difference disappeared in this case. This difference in time-course, measured from onset of the contra – ipsilateral difference via response onset to 180 ms after response onset, was significant in the patients (time-course £ Hand £ Group F2;38 ¼ 3:4, P ¼ 0:06; timecourse £ Hand in the control group: F2;14 ¼ 0:2, n.s.; timecourse £ Hand in the patients F2;24 ¼ 5:8, P ¼ 0:02). This offset of the contra – ipsilateral difference might be due either to loss of contralateral or to increase of ipsilateral activation. Inspection of the time-courses of the contra- and ipsilateral recordings (middle panels of Fig. 1) does not provide a clear answer to this question. Yet, the latter alternative is strongly suggested by the maps of topographical distribution (Fig. 3a). In the control group, there was an exclusive contralateral focus both at movement onset and 180 ms afterwards, and the same was true when patients used their unaffected left hand. In contrast, when patients used their affected hand, there was likewise an exclusive contralateral focus at movement onset, but 180 ms after-
Fig. 2. Stimulus-locked EEG potentials. Grand means across all participants, time-locked to presentation of the stimuli. Depicted are EEG potentials, averaged separately for left and right hand trials, as time-course from electrodes overlying the left and right motor cortex, C3 (gray, affected left side) and C4 (black, unaffected right side), cf. the inserted schematic head. X-axis is in milliseconds from cue onset, thus cue is presented at 0 ms, IS at 2000 ms. Y-axis is in microvolts, with negative values plotted upwards. Note that response onset, the temporal reference for the data shown in Fig. 1, occurred in each trial within 2200–3500 ms on the time scale of the present figure, varying between trials.
1473
Fig. 3. Maps of topographic distributions of amplitudes. Scalp sites are viewed from the top. Actual recording sites are marked as white dots and indicated in the symbolized head. To optimize resolution maps were individually scaled from zero (blue) to 22 times the vector norm (square root of the sum of squared amplitudes; red). (a) Response-locked activity: values are displayed for the two time points at response onset and 180 ms afterwards (^20 ms), referred to the first 100 ms of the response-locked averages. Note contralateral foci in all maps and additional ipsilateral focus in patients’ right hand press at 180 ms. Negative maxima (first right hand, then left hand), control group: 23.4, 25.0, 24.8, 26.3 mV; patients: 27.8, 210.1, 25.7, 28.5 mV. (b) Stimulus-locked activity: values are displayed for CNV (mean of 90 – 0 ms before S2) and MAN (mean of 600 –800 ms after S2). Note contralateral foci of MAN and general lack of such asymmetry in CNV. Topography of CNV differed between groups, with relatively less activity in the patients at parietal sites and at the lateral central sites overlying both motor cortices. An additional topographic map has been inserted (patients’ mean amplitudes 600–800 ms after the cue indicating press with the affected hand) to clarify the nature of the tonic contralateral negativity seen in the upper-right waveshapes of Fig. 2 (probably artifactual, see text). Negative maxima (first right hand, then left hand), control group: 216.3, 23.5, 216.0, 24.8 mV; patients: 24.3, 29.6, 27.4, 29.0, 27.0 mV.
wards, this focus was accompanied by a second ipsilateral focus. Indeed, topography differed between left and right hemispheres within 3 of the patients’ 4 maps shown in Fig. 3a (topography £ hemisphere: F9;108 . 4:2, P , 0:001) reflecting the unilateral focus, but not within the þ 180 ms map for presses of the affected hand (F9;108 ¼ 1:2, P ¼ 0:32) because in this map, ipsi- and contralateral recordings had the same topography, with foci both at C3 and C4 overlying both the affected and unaffected sensorimotor cortex. Thus, these data suggest that there is indeed ipsilateral activity when patients move their affected hands but that this activity does not start before movement onset. 3.3. Stimulus-locked potentials To show the context in which the response-locked
1474
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
potentials were embedded, stimulus-locked potentials are presented in Fig. 2, spanning the entire trial. Following the visual potential, evoked by cue onset, and a phasic negative wave peaking at about 350 ms, negativity gradually increased until onset of the IS (CNV; cf. Walter et al., 1964; Verleger et al., 2000). Following the visual evoked potential evoked by IS, a broad, large, phasic negativity (MAN) spanned the time during which participants exerted force on the keys (Slobounov et al., 1998; Verleger et al., 1999). MAN was markedly larger at sites contralateral than ipsilateral to the key-pressing hand (F1;19 ¼ 18:8, P , 0:001). However, since responses occurred with varying latencies after IS, this stimulus-locked analysis cannot distinguish between parts of MAN that occurred before movement, thus probably reflect movement preparation, and parts of MAN that occurred during responding, thus might reflect both movement control and reafferent processing. It is due to this reason that the response-locked analysis had been performed. One might suspect that ipsilateral activation had already started before the time-window covered by the responselocked data. However, there was no hint in the stimuluslocked data for any increased ipsilateral activation when patients pressed with their affected hand, neither in CNV nor in MAN. The only feature specific to patients’ pressing with the affected hand occurred about 500– 1500 ms after the cue (1500 –500 ms before the IS) with increased contralateral rather than ipsilateral activity (difference of gray and black waveshapes in upper-right graph of Fig. 2). However, as the dichotomously configurated scalp map (Fig. 3b) suggests and inspection of the horizontal EOG recording confirmed (not shown), this feature was most probably an artifact due to incomplete removal of transmission of ocular potentials, because patients tended to look towards their affected hand. In any case, this was an increase of contralateral rather than ipsilateral activity. Patients generally had smaller CNV amplitudes than the control group (main effect of Group: F1;19 ¼ 4:5, P ¼ 0:048) without any differences between responses of affected and unaffected hands. Nor were there any differences in this reduced activation between the recordings from affected and unaffected hemispheres (Fig. 2), although other topographical differences occurred (group £ recording: F21;399 ¼ 3:0, P ¼ 0:005), with the patients’ amplitude reductions being most marked at posterior sites (e.g. main effect of Group at Pz: F1;19 ¼ 5:8, P ¼ 0:03) and at the two sites overlying the left and right motor cortex (C3: F1;19 ¼ 5:4, P ¼ 0:03; C4: F1;19 ¼ 5:0, P ¼ 0:04) whereas amplitudes did not differ significantly between groups at central mesial sites (e.g. Cz: F1;19 ¼ 3:0, P ¼ 0:10; FCz: F1;19 ¼ 0:7, P ¼ n.s.) (see maps in Fig. 3b).
4. Discussion We found evidence for activation of the hemisphere
ipsilateral to the affected hand, at recording sites overlying the motor areas, when patients were responding with this hand. This ipsilateral activation is likely to be the electrophysiological counterpart of the ipsilateral activation found in preceding studies on cerebral blood flow (Weiller et al., 1993; Cramer et al., 1997, 1999; Cao et al., 1998; Seitz et al., 1999; Nelles et al., 1999; Marshall et al., 2000; Carey et al., 2002). Of most importance, ipsilateral activation was found to prevail only after movement onset, starting approximately at movement onset, i.e. about 200 ms after regular activation of the contralateral side, and reaching its maximum another 200 ms later, whereby topographical foci became apparent at sites above the motor cortex of either hemispheres. This timing of ipsilateral activation provides valuable information about its function. Ipsilateral activation obviously occurred too late for initiating the response. Conversely, its timing rather can be taken to suggest that it was initiated by the response. Immediately after patients had initiated responses with their affected hand, the opposite motor system became activated. There are at least two alternative accounts for why this might have happened. Either this activation helped the contralateral, affected cortex in maintaining, modifying, or continuing the response. To do so, this activation must have access to the ipsilateral hand, be it directly, via some ipsilateral connection, or indirectly, via transcallosal connections to the opposite, contralateral motor cortex (cf. Ziemann et al., 1999, for a thorough discussion of possible ipsilateral pathways). However, this account of ipsilateral activation does not appear to be very probable, because the contralateral motor cortex was intact in our subcortically lesioned patients, thus did not necessarily need support, and because of the lacking role of ipsilateral activation in response initiation. If no support was needed for response initiation, why would it be needed for further control of the response? The alternative account is that this activation of the opposite hemisphere indeed means activation of the other, unaffected hand: this activation is rather contralateral than ipsilateral, referring not to the affected hand but to its unaffected counterpart. Such activation might be prophylactic and might make much sense in everyday life, where patients would put the unaffected hand in readiness as soon as they start using their affected hand, in order that the unaffected hand may come into play in case the affected hand fails, just like parents are in constant readiness when allowing their little child to handle some breakable object. One argument in favor of this alternative is that the focus of this ‘ipsilateral’ activation has been found to be anterior and lateral to the motor cortex, thus probably in the premotor cortex (Cramer et al., 1999; Nelles et al., 1999; Seitz et al., 1999), as it were one step before actually using the motor system. Another argument is that the amount of ipsilateral activation was found to be a predictor for poor recovery of the affected hand (Turton et al., 1996; Netz et al., 1997), which is plausible because, rational as such prophylactic activation of the unaffected hand might be, it might prevent patients from exhausting and
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
boosting the full potential of their affected hand. Thus, concurring with the notion of ‘learned non-use’ of the affected hand (Taub et al., 1998; Liepert et al., 2000) we interpret this ipsilateral activation to consist of learned reflectory preactivation of the motor system opposite to the affected hand. In this interpretation, the present data provide another rationale why treatments like ‘constraint-induced therapy’ might be effective (Taub et al., 1998; Liepert et al., 2000). The reflexlike habit of learned ‘overprotective’ activation of the healthy hand must be unbroken in order that the affected motor system can become more self-reliant and can make further progress. 4.1. Relation to previous EEG-potential studies Ipsilateral activation was also found in previous EEGpotential studies in small groups of patients or in single cases (Kitamura et al., 1996; Kopp et al., 1999; Green et al., 1999; cf. Verleger, 2002, for a comprehensive review), and very recently in a group of patients very similar to the one studied here (Gerloff et al., 2002). Evidence provided by these studies is extended by the present data in important ways because none of these studies made use of the fine-grained temporal resolution as given by the present approach. 4.2. Measurement of motor output Ipsilateral activation was here found under conditions where overt ‘mirror movements’, i.e. force exertion by the unaffected hand, did not occur. Any pressure by the uninvolved hand that exceeded 1 N led to rejection of the trial from analysis (6.5% of trials in the patients). This is certainly a more stringent control than the mere observation of patients’ behavior used by the previous studies on cerebral blood flow that found ipsilateral activation. As recently confirmed in a study on neurological patients, force-sensitive keys are an extremely precise and sensitive means to measure force output, in particular being sensitive to any accessory force exerted by the hand that is not cued to press (Franz and Miller, 2002). Confirming the sensitivity of force measurement, our patients differed in several parameters of their force output from the healthy control group. Since our instructions to participants focused on overt motor output, we did not control EMG activity of their finger-moving muscles. This might be done in future studies to clarify the question whether activation of the unaffected motor areas remains entirely prophylactic, without any peripheral excitation, or whether there is already some ‘leakage’ of this activation to peripheral muscle activity. Either outcome would be compatible with the present account of ipsilateral cortical activation. 4.3. Stimulus-locked EEG potentials The stimulus-locked data provided important additional information. In the absence of these data, it could have been argued that the patients’ late ipsilateral activation was only compensating for some earlier imbalance between contra- and
1475
ipsilateral sides, occurring already during response preparation. This, however, was not the case. We did find a difference in the slow preparatory negativity (CNV) between patients and control group, with CNV amplitudes being smaller in the patients particularly at parietal midline (replicating recent Bereitschaftspotential data, Platz et al., 2000) and above both motor cortices. This difference, however, was surprisingly unspecific, being present when preparing movements of both the affected and unaffected hand and being present at both C3 and C4, i.e. at sites overlying both the affected and unaffected motor cortex. Thus, there was no difference between contra- and ipsilateral side and between affected and unaffected side during response preparation. The paresis of the affected hand might have general impairing effects on patients’ manual mobility (Debaere et al., 2001), therefore, the generally lowered CNV above both motor cortices might be due to this general decrease of skillful use of the hands. Alternatively, keeping in mind that all amplitudes were measured relative to the baseline before the cue, the patients’ smaller increase might be smaller only relative to baseline. They might have maintained some tonic activation of their motor areas throughout the entire session such that they needed a lower increase of activation between cue and IS. Such tonic activation can only be measured with carefully prepared (and artifact-prone) DC recordings (e.g. Marshall et al., 1998) or with PET. Indeed, Weiller et al. (1993) reported relatively large activation of both motor systems in their patients, measured by PET. These latter considerations have an important implication for the main finding of the present study, the time-course of ipsilateral activity. It must be noted that this measure of ipsilateral activation is relative to the level at trial beginning. No statement can be made about the absolute amount of contra- and ipsilateral activation. Therefore, we do not know whether there is some tonic activation of the motor cortices throughout the entire session and whether this hypothetical tonic activation is larger in patients than in controls or vice versa, and whether it is larger contralaterally than ipsilaterally or vice versa. What can be said from the present data are statements about changes of contra- and ipsilateral activity in response to the requirement to use one or the other hand (stimulus-locked, evoked by the cue) and when actually preparing and executing the movement. What we found is that these relative changes when actually executing movements were larger on the ipsilateral side in patients when the movement is with the affected hand.
5. Conclusion In summary, extending the results of previous PET and fMRI studies, the present results for the first time show the time-course of contra- and ipsilateral activation. Starting after onset of contralateral activation and even after response onset, ipsilateral activation probably has no function in initiating movements of the affected hand.
1476
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
Acknowledgements Hardware and software for force measurements were installed and developed by Piotr Jas´kowski, software for experimental control, recording, and data analysis by Edmund Wascher, software for mapping topographical distributions by Peter Trillenberg. Manfred Kaps gave an important impulse for planning this study, and Andreas Moser and Matthias Nitschke provided helpful advice. M.R. was supported by DFG grant Ve 110/10-3 to R.V.
References Anderer P, Semlitsch HV, Saletu B, Barbanoj MJ. Artifact processing in topographic mapping of electroencephalographic activity in neuropsychopharmacology. Psychiatr Res Neuroimage 1992;45:79–93. Ball T, Schreiber A, Feige B, Wagner M, Lu¨cking CH, Kristeva-Feige R. The role of higher-order motor areas in voluntary movement as revealed by high-resolution EEG and fMRI. Neuroimage 1999;10:682–94. Cao Y, D’Olhaberriague L, Vikingstad EM, Levine SR, Welch KMA. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke 1998;29:112–22. Carey JR, Kimberley TJ, Lewis SM, Auerbach EJ, Dorsey L, Rundquist P, Ugurbil K. Analysis of fMRI and finger tracking training in subjects with chronic stroke. Brain 2002;125:773–88. Chen R, Cohen LG, Hallett M. Nervous system reorganization following injury. Neuroscience 2002;111:761–73. Coles MGH. Modern mind-brain reading: psychophysiology, physiology, and cognition. Psychophysiology 1989;26:251–69. Cramer SC, Nelles G, Benson RR, Kaplan JD, Parker RA, Kwong KK, Kennedy DN, Finklestein SP, Rosen BR. A functional MRI study of subjects recovered from hemiparetic stroke. Stroke 1997;28:2518–27. Cramer SC, Finklestein SP, Schaechter JD, Bush G, Rosen BR. Activation of distinct motor cortex regions during ipsilateral and contralateral finger movements. J Neurophysiol 1999;81:383 –7. Debaere F, Van Assche D, Kiekens C, Verschueren SMP, Swinnen SP. Coordination of upper and lower limb segments: deficits on the ipsilesional side after unilateral stroke. Exp Brain Res 2001;141:519–29. Franz EA, Miller J. Effects of response readiness on reaction time and force output in people with Parkinson’s disease. Brain 2002;125:1733–50. Gerloff C, Sailer A, Braun C, Schmitt C, Bushara K, Wittenberg GF, Chen R, Ishii K, Waldvogel D, Wassermann EM, Cohen LG, Hallett M. Plasticity after stroke: homologous areas with non-homologous function. Paper presented at the Fourth Workshop on Cortical Plasticity, March 15–17, 2002, Castle of Schwetzingen, Germany. Gilbert PFC. An outline of brain function. Cogn Brain Res 2001;12:61– 74. Green JB, Bialy Y, Sora E, Ricamato A. High-resolution EEG in poststroke hemiparesis can identify ipsilateral generators during motor tasks. Stroke 1999;30:2659 –65. Kitamura J-I, Shibasaki H, Takeuchi T. Cortical potentials preceding voluntary elbow movement in recovered hemiparesis. Electroenceph clin Neurophysiol 1996;98:149 –56. Kopp B, Kunkel A, Mu¨hlnickel W, Villringer K, Taub E, Flor H. Plasticity in the motor system related to therapy-induced improvement of movement after stroke. NeuroReport 1999;10:807– 10. Kunieda T, Ikeda A, Ohara S, Yazawa S, Nagamine T, Taki W, Hashimoto N, Shibasaki H. Different activation of presupplementary motor area, supplementary motor area proper, and primary sensorimotor area, depending on the movement repetition rate in humans. Exp Brain Res 2000;135:163–72. Kutas M, Donchin E. Preparation to respond as manifested by movementrelated potentials. Brain Res 1980;202:95–115.
Liepert J, Weiller C. Mapping plastic brain changes after acute lesions. Curr Opin Neurol 1999;12:709 –13. Liepert J, Bauder H, Miltner WHR, Taub E, Weiller C. Treatment-induced cortical reorganization after stroke in humans. Stroke 2000;31:1210 –6. Marshall L, Mo¨lle M, Fehm HL, Born J. Scalp recorded direct current brain potentials during human sleep. Eur J Neurosci 1998;10:1167–78. 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. Nelles G, Spiekermann G, Jueptner M, Leonhardt G, Mu¨ller S, Gerhard H, Diener HC. Evolution of functional reorganization in hemiplegic stroke: a serial positron emission tomographic activation study. Ann Neurol 1999;46:901–9. Netz J, Lammers T, Ho¨mberg V. Reorganization of motor output in the non-affected hemisphere after stroke. Brain 1997;120:1579– 86. Platz T, Kim IH, Pintschovius H, Winter T, Kieselbach A, Villringer K, Kurth R, Mauritz K-H. Multimodal EEG analysis in man suggests impairment-specific changes in movement-related electric brain activity after stroke. Brain 2000;123:2475–90. Rizzolatti G, Luppino G, Matelli M. The organization of the cortical motor system: new concepts. Electroenceph clin Neurophysiol 1998;106:283–96. Seitz RJ, Azari NP. Cerebral reorganization in man after acquired lesions. In: Stefan H, Andermann F, Chauvel P, Shorvon S, editors. Advances in neurology, vol. 81. Philadelphia, PA: Lippincott Williams and Wilkins; 1999. p. 37– 47. Seitz RJ, Knorr U, Azari NP, Herzog H, Freund H-J. Visual network activation in recovery from sensorimotor stroke. Restorat Neurol Neurosci 1999;14:25–33. Slobounov SM, Ray WJ, Simon RF. Movement-related potentials accompanying unilateral finger movements with special reference to rate of force development. Psychophysiology 1998;35:537 –48. Small SL, Hlustik P, Noll DC, Genovese C, Solodkin A. Cerebellar hemispheric activation ipsilateral to the paretic hand correlates with functional recovery after stroke. Brain 2002;125:1544–57. Taub E, Crago JE, Uswatte G. Constraint-induced movement therapy, a new approach to treatment in physical rehabilitation. Rehabil Psychol 1998;43:152–70. Turton A, Wroe S, Trepte N, Fraser C, Lemon RN. Contralateral and ipsilateral EMG responses to transcranial magnetic stimulation during recovery of arm and hand function after stroke. Electroenceph clin Neurophysiol 1996;101:316–28. Verleger R. Event-related EEG potential research in neurological patients. In: Zani A, Proverbio AM, editors. The cognitive electrophysiology of mind and brain. San Diego, CA: Academic Press; 2002. p. 309 –41. Verleger R, Gasser T, Mo¨cks J. Correction of EOG artifacts in event related potentials of the EEG: aspects of reliability and validity. Psychophysiology 1982;19:472– 80. Verleger R, Wascher E, Wauschkuhn B, Jas´kowski P, Allouni B, Trillenberg P, Wessel K. Consequences of altered cerebellar input for the cortical regulation of motor coordination, as reflected in EEG potentials. Exp Brain Res 1999;127:409 –22. Verleger R, Wauschkuhn B, Van der Lubbe RHJ, Jas´kowski P, Trillenberg P. Posterior and anterior contributions of hand-movement preparation to late CNV. J Psychophysiol 2000;14:69 –86. Walter WG, Cooper R, Aldridge VJ, McCallum WC, Winter AL. Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature 1964;203:380–4. Weiller C, Ramsay SC, Wise RJS, Friston KJ, Frackowiak RSJ. Individual patterns of functional reorganization in the human cerebral cortex after capsular infarction. Ann Neurol 1993;33:181–9. Ziemann U, Ishii K, Borgheresi A, Yaseen Z, Battaglia F, Hallett M, Cincotta M, Wassermann EM. Dissociation of the pathways mediating ipsilateral and contralateral motor-evoked potentials in human hand and arm muscles. J Physiol 1999;518:895–906.
Control of hand movements after striatocapsular stroke: high-resolution temporal analysis of the function of ipsilateral activation Rolf Verleger*, Sven Adam, Michael Rose1, Clemens Vollmer, Bernd Wauschkuhn, Detlef Ko¨mpf Department of Neurology, Medical University of Lu¨beck, Ratzeburger Allee 160, D 23538 Luebeck, Germany Accepted 15 April 2003
Abstract Objective: Hemiparesis due to infarction of the middle cerebral artery has become an increasingly important focus of research on cortical plasticity. Positron emission tomography and functional magnetic resonance imaging studies in such patients found involvement of the hemisphere ipsilateral to the affected hand related to movements of this hand. To understand the function of this ipsilateral activation, the present study investigated movement-related electroencephalogram (EEG) potentials in patients and healthy control subjects to measure timing of ipsi- and contralateral activation relative to movement onset. Methods: Thirteen patients were investigated in their chronic stage. Their pyramidal tracts were affected by infarctions of the middle cerebral artery at striatocapsular level. EEG potentials were recorded from 26 scalp electrodes while patients were pressing a key with their right or left index finger within a warned choice – response task. Results: Beginning 200 ms before responses of the affected hand, there was normal contralateral preponderance of EEG negativity. Briefly after response onset, however, the other unaffected hemisphere, ipsilateral to the responding hand, became additionally active. This pattern did not occur with responses made by the unaffected hand nor in healthy participants. Conclusions: The timing of the onset of ipsilateral activity precludes its role in response initiation. Rather, this activity may indicate reflex-like activation of the unaffected motor system to compensate for possible failure of the affected hand. q 2003 International Federation of Clinical Neurophysiology. Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Hemiparesis; Stroke; Event-related potentials; Cortical reorganization; Plasticity
1. Introduction The primary motor cortex is the principal unit of the brain for executing hand movements, based on input provided by other cortical, subcortical and cerebellar structures (Rizzolatti et al., 1998; Gilbert, 2001). In either hemisphere, output of these areas is sent via the pyramidal tract. Infarction (or hemorrhage) in the middle cerebral artery can impair this function, by damaging either the primary motor cortex or the subcortical pyramidal tract at the corona radiata and the capsula interna. The present study focussed on patients with such striatocapsular, subcortical damage. Under these circumstances, though in * Corresponding author. Tel.: þ49-451-500-2916; fax: þ 49-451-5002489. E-mail address: [email protected] (R. Verleger). 1 Now at: Department of Neurology, University Clinic of Hamburg, Hamburg, Germany.
principle unaffected, the cortical motor areas have to deal with the problem of blocked or restricted efferent conduction. Therefore, this condition has become a model case in research on cortical plasticity (Liepert and Weiller, 1999; Seitz and Azari, 1999; Chen et al., 2002). An issue of particular interest is the role taken by the motor cortex ipsilateral to the affected hand when patients move this hand. Evidence for ipsilateral involvement, in addition to contralateral activation, was reported in positron emission tomography (PET) (Weiller et al., 1993; Seitz et al., 1999) and functional magnetic resonance imaging (fMRI) studies (Cramer et al., 1997; Cao et al., 1998; Carey et al., 2002; but see Small et al., 2002) with a focus of activity in the ipsilateral premotor cortex (Cramer et al., 1999; Nelles et al., 1999; Seitz et al., 1999). What may be the function of such ipsilateral activation? One source of evidence is its long-term time-course, related to recovery. There have been reports both on increasing
1388-2457/03/$30.00 q 2003 International Federation of Clinical Neurophysiology. Published by Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S1388-2457(03)00125-1
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
(Nelles et al., 1999) and decreasing (Marshall et al., 2000; Carey et al., 2002) ipsilateral activation over time. Decrease of ipsilateral activation accompanying recovery would suggest that its continuing existence is maladaptive. Indeed, the size of patients’ electromyogram (EMG) activity on the affected hand evoked by transcranial magnetic stimulation of the ipsilateral motor cortex has been reported to be associated with poor recovery (Turton et al., 1996; Netz et al., 1997). Another source of evidence, exploited in the present study, is the precise timing of ipsilateral activity: Does it take place before, simultaneously with, or after contralateral activity? Measurement based on blood flow (PET and fMRI), powerful as they are for localizing activity, have not been able to provide answers on this question. What is needed is the detailed temporal resolution provided by electrophysiology. Therefore, multi-channel electroencephalogram (EEG) was measured in a task, in which participants had to produce responses with their right or left hands. Of major interest was the time-course of EEG potentials time-locked to response onset, to investigate whether any additional activity of the side ipsilateral to the affected hand occurred in the patients, more than when moving their unaffected hand, and more than the control group, and if so, how this ipsilateral activity relates in time both to contralateral activity and to the overt response. The earlier the ipsilateral activation occurs, both in relation to contralateral activity and to movement start, the more plausible would be its supportive function in initiating movements. Previous EEG studies in such patients did not exploit the EEG’s capacity for temporal resolution to this degree (Kitamura et al., 1996; Green et al., 1999; Kopp et al., 1999; Platz et al., 2000). Participants were cued to respond with their affected and unaffected hands (right or left hand, respectively) in random order over trials, and responses had to be produced with varying degrees of force, in order to enhance task demands to a level that would indeed be difficult to the patients and, therefore, would increase the likelihood that differences between control of both hands would be revealed. To control for differences in preparatory activity, EEG potentials were also measured in the interval between cue and imperative stimulus (IS).
2. Methods 2.1. Participants Thirteen patients and 8 age-matched healthy subjects participated. Patients were selected from the files of our department, by these criteria: an ischemic event should have taken place within the area supplied by the middle cerebral artery affecting the pyramidal tract at subcortical level; there should still be mild hemiparesis due to this event; there should be no or only very mild somatosensory deficits and no other ischemic event or additional neurological disease
1469
should have occurred. Clinical data of the patients are detailed in Table 1. Of the 13 patients, 11 had been affected by infarction, two by hemorrhage. The pyramidal tract was affected at corona radiata in 5 patients, capsula interna in 3 patients (including the two hemorrhagic patients), and in both areas in 5 patients. Additional areas affected were nucleus lentiformis ðn ¼ 4Þ, nucleus caudatus ðn ¼ 3Þ, putamen ðn ¼ 1Þ, and thalamus ðn ¼ 1Þ. These patients were in a chronic, stable state, the stroke dating back 5 years on average (20 months – 9 years). The left hemisphere was affected in 10 patients, the right hemisphere in 3 patients. Thus, since one right hemisphere patient was left handed before the lesion, the dominant hemisphere was affected in 11 patients, consecutively, 3 of these patients at least partially changed their hand dominance from the affected to the healthy side, as assessed by the Edinburgh Handedness Inventory, in spite of the mild degree of paresis. At the time of our study, the patients’ mean age was 65 years (57 – 71 years), and the mean degree of paresis of the affected hand as determined by standard neurological examination (Medical Research Council scale, extended by intermediate values) was 4– 5, i.e. very mild (ranging from 3– 4 to 4– 5). The 11 infarction patients received anticoagulating medication. The 8 control subjects were 4 men and 4 women, their mean age was 62 years (53 – 73 years), all were right handed. All participants gave their informed consent according to the declaration of Helsinki. They had normal or corrected-to-normal visual acuity and did not take medication known to affect the central nervous system. The study was approved by the ethical committee of the Medical University at Lu¨beck. 2.2. Procedure and stimuli Participants had to use their right or left hands with defined forces, randomly varying over trials, within defined temporal limits. Responses were made to an imperative signal which appeared 2 s after a cue that indicated hand and force. The devices for force measurement were two forcesensitive keys, one for each index finger. The keys were affixed at the front edge of the armrests of the experimental chair. Subjects’ forearms lay on the armrests, in order to prevent subjects from exerting additional force by moving their arms or their body. Either key was connected to an electronic force sensor whose voltage output increased proportionally to the pressure exerted on the key. This force output was visualized on the computer screen by two green bars, one for either hand, moving upwards from a zero position proportionally to the exerted force (see Verleger et al., 1999, for details). In addition to this immediate visual feedback, differential auditory feedback was provided for correct, incorrect, and premature responses. Responses were accepted as correct if force was exerted on the key by the indicated hand only. The window allowed for responding was 0 –1.9 s after the imperative signal.
1470
Table 1 Clinical data Age (years)
Sex
Infarct or hemorrhage
Lesion site
#1 #2 #3 #4
69 64 62 57
m m m f
infarct infarct infarct infarct
cr cr? cr cr and ci
#5 #6 #7 #8 #9 #10 #11 #12 #13
64 67 64 60 64 71 57 70 71
f m f m f m m f f
infarct infarct infarct hemorrhage infarct infarct hemorrhage infarct infarct
cr cr cr ci cr cr ci cr ci
Summary
65 (57–71)
7m, 6f
11 infarcts, 2 hemorrhage
5 cr, 3 ci, 5 cr and ci
Other lesion sites
nucl. caudatus? nucl. caudatus, nucl. lentiformis
and ci and ci and ci
putamen, nucl. lentiformis nucl. lentiformis thalamus
and ci nucl. caudatus, nucl. lentiformis 4 £ nucl. lentiformis, 3 £ nucl. caudatus, 1 £ putamen, 1 £ thalamus
Length £ width £ height of lesion (cm)
Time passed
Side of lesion
Hand preference before lesion
Affected hand
Hand preference changed
Degree of paresis
2 £ 1 £ 1.8 not known 1 £ 1 £ 1.6 4 £ 2 £ 3.0
1y8 3y6 5y8 6y2
left right right left
right right left right
right left left right
no – no no
4– 5 3– 4 4 4– 5
1 £ 1 £ 1.6 2 £ 1 £ 2.4 2 £ 2 £ 2.4 2.5 £ 4 £ 6.4 1.5 £ 1 £ 3.2 1.5 £ 1 £ 2.4 2 £ 2 £ 3.0 1.5 £ 1 £ 1.6 5 £ 2 £ 4.0
5y1m 4y8m 4y2m 4y0m 3y8m 4y5m 6y1m 8 y 10 m 4y2m
left left left left left left left left right
right right right right right right right right right
right right right right right right right right left
no yes yes no no no yes no –
4 4 4– 5 4– 5 4– 5 4– 5 4– 5 4– 5 4– 5
2.2 £ 1.6 £ 2.8
4y9m (1 y 8 m – 8 y 10 m)
10 left, 3 right
12 right, 1 left
10 right, 3 left
3 changes
4– 5 (3–4 to 4– 5)
m m m m
M, male; F, female; cr, corona radiata; ci, capsula interna; y, years; m, months. MRI (computed tomography (CT)) scans were made at the acute stage. Length and width of lesion were measured in the MRI (CT) slice with the largest extent of lesion; height is given by the number of slices in which the lesion was visible (Patient 2’s MRI scan got lost before this precise measurement could be performed. Details about his lesion sites were taken from his clinical files). Careful neurological examination was performed immediately before EEG recording to evaluate the degree of paresis and any signs of impairment not due to the known infarction. ‘Hand preference before lesion’ was asked in retrospect by applying the items of the Edinburgh Handedness Inventory. No entry was made in the column ‘preference changed’ if the affected hand was not the preferred one. Summary: means and ranges are indicated for ‘age’ and ‘time passed’, median and range for ‘degree of paresis’ (scale ranging from 0, plegia, to 5, no impairment), mean of each dimension separately (length, width, height) for length £ width £ height of lesion.
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
Code
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
The session consisted of 5 repetitions of two blocks, spontaneous force and graded force. Working through the entire session of 5 repetitions £ two blocks took about 75 min, plus about 10 min for instructions and practice at the beginning. For this report, data were pooled across 150 trials for either hand, from the unimanual spontaneous force and the 30 and 60% graded force trials, where participants had to press the key with the right or left hand, applying 100, 60, or 30% of their maximum convenient force, respectively. Subjects looked at the 17 inches screen from about 1.2 m. In each trial, two events occurred on the screen, cue and IS. The cue was the appearance of two blue bars, with the left bar assigned to the left hand and the right bar to the right hand. IS was the turn of the colour of the bars from blue to yellow, 2 s after cue onset. As soon as participants exerted force on the keys, the momentarily applied force was displayed by a green bar for either hand, of same size as the blue/yellow bars. Four seconds after IS, the next trial started with a new cue. In spontaneous force, the two blue bars appeared side by side. Simultaneously, an upward-pointing light-blue arrow appeared below the left, right, or both bars, forming the cue. Participants had to press the key(s) denoted by the arrow(s) briskly and forcefully when the blue bars had turned yellow (IS). The 3 tasks (right, left, both hands, 10 trials in each of the 5 repetitions across the session) alternated in random order. In graded force, two vertical scales appeared simultaneously with the blue bars, at the outer side of either bar. Either scale consisted of a vertical line separated in 4 equal parts by 5 horizontal tick marks at its outer side. Zero percent was marked at the lowest mark, 50% at the middle one, and 100% at the upper mark. The scale of 100%was the maximum force calculated from the preceding spontaneous force condition. The left and the right blue bar appeared at heights of 0, 30, or 60% of their corresponding scale, varying over trials, this way denoting which force had to be exerted after IS by which hand. Half the trials were unimanual tasks. Participants had to try and cover the yellow bar(s) by the corresponding green bar(s) after IS, by applying graded force. 2.3. Recording, data processing, and analysis For data analysis, all patients were considered to have a left-sided lesion and, therefore, right hand paresis. To this end, hand-force recordings and lateral EEG recording sites were reversed for the 3 patients who had the lesion in their right hemispheres. 2.4. EEG potentials EEG was recorded from 26 Ag/AgCl scalp electrodes (Fp1, Fp2, F3, Fz, F4, FC5, FC1, FCz, FC2, FC6, C3, C1, Cz, C2, C4, CP5, CP1, CP2, CP6, P7, P3, Pz, P4,
1471
P8, O1, O2) referred to an electrode at the nose tip. Vertical and horizontal electrooculogram (EOG) were recorded bipolarly from above vs. below the left eye and from the outer canthi of the eyes, for recognizing ocular artifacts. EEG and EOG were amplified (band pass 0.03– 35 Hz), digitized with 100 Hz per channel and digitally stored. Data were edited in several steps. First, trials with amplifier blocking (zero lines) were rejected from further analysis. Then, the impact of vertical EOG deflections on each EEG channel was estimated (by linear regression in epochs of large EOG activity) and these artifacts were removed from the EEG by subtraction (Verleger et al., 1982), this way keeping the affected epochs for further analysis. Third, in the same way, the impact of horizontal EOG deflections was estimated and removed (Anderer et al., 1992). Fourth, trials with other artifacts were rejected from further analysis (large amplitudes, fast shifts, and slow drifts). Finally, trials with premature, missing (. 1500 ms), or wrong side responses were rejected. The mean number (range) of remaining artifact-free and correctly responded unimanual trials was 85 (29 – 109) and 85 (40 – 112) for healthy (left) hand and affected (right) hand trials in the patients, and 114 (94 –128) and 119 (79 – 132) for left and right hand trials in the control group. The group difference was significant (F1;19 ¼ 13:8, P ¼ 0:001) reflecting both more erroneous responses and more EEG artifacts in the patients’ rejected data. These data were averaged separately for either hands and each participant in two ways, time-locked to the stimuli and response. Response-locked averages covered the time from 400 ms before response onset to 400 ms afterwards. Parameters quantified were mean amplitudes before response onset (2 220 to 2 180 ms), at response onset (2 20 to þ 20 ms), and after response onset (þ 160 to þ 200 ms). Stimuluslocked averages covered the entire 4 s recording of the trials from 100 ms before cue onset to 1.9 s after IS onset. Parameters quantified were contingent negative variation (CNV; mean amplitude 90 –0 ms before IS onset) and movement-accompanying negativity (MAN; most negative peak 350 – 1500 ms after IS). Time-courses of both response-locked and stimulus-locked potentials were analyzed mainly from C3 and C4 (the scalp sites overlying the motor cortex). Topographic maps were done of the scalp distribution of potentials for the mentioned parameters, using spherical spline interpolation. Analyses of variance (ANOVA) had the factors Group, Hand, and Recordings (two when analyzing C3 and C4, otherwise 22). For comparing contra- and ipsilateral topography, ANOVA had the factors Hemisphere (contralateral vs. ipsilateral) and recordings (10: F3, FC5, FC1, C3, C1, CP5, CP1, P7, P3, O1 on the left; F4, FC6, FC2, C4, C2, CP6, CP2, P8, P4, O2 on the right). When repeated-measurement factors had more than two levels, degrees of freedom were corrected by Huynh –Feldt’s 1 In this case, the corrected P values will be reported together with the original degrees of freedom.
1472
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
2.5. Force The voltage output of both force-sensitive keys was sampled with 100 Hz from 0.1 s before to 3.9 s after the cue (1.9 s after IS) and stored on hard disk. Data were converted to Newton by reference to calibration with defined weights and further analyzed with self-developed software. (Newton is the unit of force, 1 N is the force exerted by a mass of 102 g on its basis). First, 3 types of errors were separately counted, each calculated as number of error trials divided by the total number of trials in which the involved hand could commit the error: premature responses (i.e. before the imperative signal), missing responses, and wrong side. ‘Wrong side’ was counted for any pressure exceeding 1 N on the wrong key even if the stronger response was on the correct side. Four parameters of the force curves were determined separately for either hand in each errorless single trial and were averaged across trials. (1) Response time (RT): the time elapsed from IS onset until force increased above 1 N in spontaneous force and above 9% of maximum force in graded force. (2) Movement time (MT): the time from RT until force maximum in spontaneous force and until the moment of best fit in graded force. (3) Slope: the force difference between RT þ 200 ms and RT, to measure the initial impulse exerted on the keys. (4) Force: in spontaneous force, the amplitude of the force maximum and in graded force, the best fit, defined as the smallest deviation from target force, are expressed as percentage of maximum force. These parameters were analyzed by ANOVA with 3 factors, Group (patients vs. control group) and, as repeatedmeasurement factor, Hand (left vs. right ¼ healthy vs. affected in the patients).
3. Results 3.1. Behavior The mild degree of palsy of patients’ affected hands (range 3.5 –4.5) was well visible in the force measures. As Fig. 1 shows, (upper panels) maximum force was weaker when patients pressed the keys with the affected hand than with the unaffected hand (Hand £ Group: F1;19 ¼ 6:77, P ¼ 0:02; effect of Hand in the patients: F1;12 ¼ 4:23, P ¼ 0:06), the initial slope of force was less steep with this hand (Hand £ Group: F1;19 ¼ 10:69, P ¼ 0:004; effect of Hand in the patients: F1;12 ¼ 11:80, P ¼ 0:005), and (not shown by Fig. 1) deviation from target force was larger with this hand (Hand £ Group: F1;19 ¼ 7:33, P ¼ 0:01; effect of Hand in the patients: F1;12 ¼ 4:39, P ¼ 0:06). In contrast, temporal parameters of the response did not differentiate between groups, i.e. patients were neither reliably slower in initiating their response (RT) nor in executing their response (MT) (main effect of group:
Fig. 1. Response-locked EEG potentials. Grand means across all participants, time-locked to response onset. Depicted are response forces and EEG potentials, averaged separately for left and right hand trials. X-axis is in milliseconds relative to response onset. Upper panels: response force (gray denoting the affected right hand, black the unaffected left hand). Yaxis is in Newton. Second and third panels: EEG potentials recorded from sites overlying the left and right motor cortex, C3 (gray, affected left side) and C4 (black, unaffected right side), cf. the schematic head on the right. Yaxis is in microvolts (EEG), with negative values plotted upwards. Baseline for these data is the 100 ms interval before the cue, like in stimulus-locked potentials (Fig. 2). Lower panels: differences contralateral 2 ipsilateral. Black lines denote the C4 –C3 difference from key-presses with the unaffected hand (i.e. the differences between the C4 and C3 waveshapes in the third panels from above), gray lines denote the C3 –C4 difference from key-presses with the affected hand (i.e. the differences between the C3 and C4 waveshapes in the second panels from above). The arrows mark the time point when the contra-ipsilateral difference returned to zero, after movement onset in the patients’ recordings.
F1;19 ¼ 2:28, P ¼ 0:15 for RT; F1;19 ¼ 0:22, P ¼ 0:64 for MT). Nor was there a differential effect for the patients’ affected hand (Hand £ Group: F1;19 ¼ 0:05, P ¼ 0:83 for RT; F1;19 ¼ 1:43, P ¼ 0:25 for MT). Finally, there were some differences between groups in errors: patients had more responses missing ðRT . 1500 msÞ than the control group (9 vs. 4% on average), in particular with their affected hand (10% in the affected, 8% in the unaffected hand; Hand £ Group: F1;19 ¼ 6:76, P ¼ 0:02). Patients also pressed the key on the wrong side more often (6.5 vs. 2%; Group: F1;19 ¼ 4:97, P ¼ 0:04) without reliable differences between hands. There was no group difference in the number of premature responses (about 2% of all trials). 3.2. Response-locked potentials The analysis of main interest was on the data time-locked to response onset, following the IS. Response-locked potentials recorded from scalp sites C3 and C4, situated above the left (affected) and right (healthy) motor cortex, are displayed in the middle panels of Fig. 1, as grand means of either group. Being referred to the epoch before the cue as their baseline, these response-locked waveshapes started on negative levels, more so in the control group than in the patients, due to the preceding larger CNV of the control group (Fig. 2 and analysis below). Of importance, for patients and control group alike, EEG became more
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
negative at the site overlying the contralateral than the ipsilateral motor cortex about 200 ms before movement onset, with this contra-ipsilateral difference increasing until movement onset. This is a standard result (e.g. Kutas and Donchin, 1980; Ball et al., 1999; Kunieda et al., 2000). These differences between contra- and ipsilateral waveshapes are more clearly seen in the difference waveshapes in the lower panels of Fig. 1 (similar to the ‘lateralized readiness potential’, Coles, 1989, where additionally data from both hands are pooled). These differences appeared to be larger in patients than in the control group, but this was not significant. After movement onset, however, a specific feature emerged when patients were using their affected hand: the contra-ipsilateral difference disappeared in this case. This difference in time-course, measured from onset of the contra – ipsilateral difference via response onset to 180 ms after response onset, was significant in the patients (time-course £ Hand £ Group F2;38 ¼ 3:4, P ¼ 0:06; timecourse £ Hand in the control group: F2;14 ¼ 0:2, n.s.; timecourse £ Hand in the patients F2;24 ¼ 5:8, P ¼ 0:02). This offset of the contra – ipsilateral difference might be due either to loss of contralateral or to increase of ipsilateral activation. Inspection of the time-courses of the contra- and ipsilateral recordings (middle panels of Fig. 1) does not provide a clear answer to this question. Yet, the latter alternative is strongly suggested by the maps of topographical distribution (Fig. 3a). In the control group, there was an exclusive contralateral focus both at movement onset and 180 ms afterwards, and the same was true when patients used their unaffected left hand. In contrast, when patients used their affected hand, there was likewise an exclusive contralateral focus at movement onset, but 180 ms after-
Fig. 2. Stimulus-locked EEG potentials. Grand means across all participants, time-locked to presentation of the stimuli. Depicted are EEG potentials, averaged separately for left and right hand trials, as time-course from electrodes overlying the left and right motor cortex, C3 (gray, affected left side) and C4 (black, unaffected right side), cf. the inserted schematic head. X-axis is in milliseconds from cue onset, thus cue is presented at 0 ms, IS at 2000 ms. Y-axis is in microvolts, with negative values plotted upwards. Note that response onset, the temporal reference for the data shown in Fig. 1, occurred in each trial within 2200–3500 ms on the time scale of the present figure, varying between trials.
1473
Fig. 3. Maps of topographic distributions of amplitudes. Scalp sites are viewed from the top. Actual recording sites are marked as white dots and indicated in the symbolized head. To optimize resolution maps were individually scaled from zero (blue) to 22 times the vector norm (square root of the sum of squared amplitudes; red). (a) Response-locked activity: values are displayed for the two time points at response onset and 180 ms afterwards (^20 ms), referred to the first 100 ms of the response-locked averages. Note contralateral foci in all maps and additional ipsilateral focus in patients’ right hand press at 180 ms. Negative maxima (first right hand, then left hand), control group: 23.4, 25.0, 24.8, 26.3 mV; patients: 27.8, 210.1, 25.7, 28.5 mV. (b) Stimulus-locked activity: values are displayed for CNV (mean of 90 – 0 ms before S2) and MAN (mean of 600 –800 ms after S2). Note contralateral foci of MAN and general lack of such asymmetry in CNV. Topography of CNV differed between groups, with relatively less activity in the patients at parietal sites and at the lateral central sites overlying both motor cortices. An additional topographic map has been inserted (patients’ mean amplitudes 600–800 ms after the cue indicating press with the affected hand) to clarify the nature of the tonic contralateral negativity seen in the upper-right waveshapes of Fig. 2 (probably artifactual, see text). Negative maxima (first right hand, then left hand), control group: 216.3, 23.5, 216.0, 24.8 mV; patients: 24.3, 29.6, 27.4, 29.0, 27.0 mV.
wards, this focus was accompanied by a second ipsilateral focus. Indeed, topography differed between left and right hemispheres within 3 of the patients’ 4 maps shown in Fig. 3a (topography £ hemisphere: F9;108 . 4:2, P , 0:001) reflecting the unilateral focus, but not within the þ 180 ms map for presses of the affected hand (F9;108 ¼ 1:2, P ¼ 0:32) because in this map, ipsi- and contralateral recordings had the same topography, with foci both at C3 and C4 overlying both the affected and unaffected sensorimotor cortex. Thus, these data suggest that there is indeed ipsilateral activity when patients move their affected hands but that this activity does not start before movement onset. 3.3. Stimulus-locked potentials To show the context in which the response-locked
1474
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
potentials were embedded, stimulus-locked potentials are presented in Fig. 2, spanning the entire trial. Following the visual potential, evoked by cue onset, and a phasic negative wave peaking at about 350 ms, negativity gradually increased until onset of the IS (CNV; cf. Walter et al., 1964; Verleger et al., 2000). Following the visual evoked potential evoked by IS, a broad, large, phasic negativity (MAN) spanned the time during which participants exerted force on the keys (Slobounov et al., 1998; Verleger et al., 1999). MAN was markedly larger at sites contralateral than ipsilateral to the key-pressing hand (F1;19 ¼ 18:8, P , 0:001). However, since responses occurred with varying latencies after IS, this stimulus-locked analysis cannot distinguish between parts of MAN that occurred before movement, thus probably reflect movement preparation, and parts of MAN that occurred during responding, thus might reflect both movement control and reafferent processing. It is due to this reason that the response-locked analysis had been performed. One might suspect that ipsilateral activation had already started before the time-window covered by the responselocked data. However, there was no hint in the stimuluslocked data for any increased ipsilateral activation when patients pressed with their affected hand, neither in CNV nor in MAN. The only feature specific to patients’ pressing with the affected hand occurred about 500– 1500 ms after the cue (1500 –500 ms before the IS) with increased contralateral rather than ipsilateral activity (difference of gray and black waveshapes in upper-right graph of Fig. 2). However, as the dichotomously configurated scalp map (Fig. 3b) suggests and inspection of the horizontal EOG recording confirmed (not shown), this feature was most probably an artifact due to incomplete removal of transmission of ocular potentials, because patients tended to look towards their affected hand. In any case, this was an increase of contralateral rather than ipsilateral activity. Patients generally had smaller CNV amplitudes than the control group (main effect of Group: F1;19 ¼ 4:5, P ¼ 0:048) without any differences between responses of affected and unaffected hands. Nor were there any differences in this reduced activation between the recordings from affected and unaffected hemispheres (Fig. 2), although other topographical differences occurred (group £ recording: F21;399 ¼ 3:0, P ¼ 0:005), with the patients’ amplitude reductions being most marked at posterior sites (e.g. main effect of Group at Pz: F1;19 ¼ 5:8, P ¼ 0:03) and at the two sites overlying the left and right motor cortex (C3: F1;19 ¼ 5:4, P ¼ 0:03; C4: F1;19 ¼ 5:0, P ¼ 0:04) whereas amplitudes did not differ significantly between groups at central mesial sites (e.g. Cz: F1;19 ¼ 3:0, P ¼ 0:10; FCz: F1;19 ¼ 0:7, P ¼ n.s.) (see maps in Fig. 3b).
4. Discussion We found evidence for activation of the hemisphere
ipsilateral to the affected hand, at recording sites overlying the motor areas, when patients were responding with this hand. This ipsilateral activation is likely to be the electrophysiological counterpart of the ipsilateral activation found in preceding studies on cerebral blood flow (Weiller et al., 1993; Cramer et al., 1997, 1999; Cao et al., 1998; Seitz et al., 1999; Nelles et al., 1999; Marshall et al., 2000; Carey et al., 2002). Of most importance, ipsilateral activation was found to prevail only after movement onset, starting approximately at movement onset, i.e. about 200 ms after regular activation of the contralateral side, and reaching its maximum another 200 ms later, whereby topographical foci became apparent at sites above the motor cortex of either hemispheres. This timing of ipsilateral activation provides valuable information about its function. Ipsilateral activation obviously occurred too late for initiating the response. Conversely, its timing rather can be taken to suggest that it was initiated by the response. Immediately after patients had initiated responses with their affected hand, the opposite motor system became activated. There are at least two alternative accounts for why this might have happened. Either this activation helped the contralateral, affected cortex in maintaining, modifying, or continuing the response. To do so, this activation must have access to the ipsilateral hand, be it directly, via some ipsilateral connection, or indirectly, via transcallosal connections to the opposite, contralateral motor cortex (cf. Ziemann et al., 1999, for a thorough discussion of possible ipsilateral pathways). However, this account of ipsilateral activation does not appear to be very probable, because the contralateral motor cortex was intact in our subcortically lesioned patients, thus did not necessarily need support, and because of the lacking role of ipsilateral activation in response initiation. If no support was needed for response initiation, why would it be needed for further control of the response? The alternative account is that this activation of the opposite hemisphere indeed means activation of the other, unaffected hand: this activation is rather contralateral than ipsilateral, referring not to the affected hand but to its unaffected counterpart. Such activation might be prophylactic and might make much sense in everyday life, where patients would put the unaffected hand in readiness as soon as they start using their affected hand, in order that the unaffected hand may come into play in case the affected hand fails, just like parents are in constant readiness when allowing their little child to handle some breakable object. One argument in favor of this alternative is that the focus of this ‘ipsilateral’ activation has been found to be anterior and lateral to the motor cortex, thus probably in the premotor cortex (Cramer et al., 1999; Nelles et al., 1999; Seitz et al., 1999), as it were one step before actually using the motor system. Another argument is that the amount of ipsilateral activation was found to be a predictor for poor recovery of the affected hand (Turton et al., 1996; Netz et al., 1997), which is plausible because, rational as such prophylactic activation of the unaffected hand might be, it might prevent patients from exhausting and
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
boosting the full potential of their affected hand. Thus, concurring with the notion of ‘learned non-use’ of the affected hand (Taub et al., 1998; Liepert et al., 2000) we interpret this ipsilateral activation to consist of learned reflectory preactivation of the motor system opposite to the affected hand. In this interpretation, the present data provide another rationale why treatments like ‘constraint-induced therapy’ might be effective (Taub et al., 1998; Liepert et al., 2000). The reflexlike habit of learned ‘overprotective’ activation of the healthy hand must be unbroken in order that the affected motor system can become more self-reliant and can make further progress. 4.1. Relation to previous EEG-potential studies Ipsilateral activation was also found in previous EEGpotential studies in small groups of patients or in single cases (Kitamura et al., 1996; Kopp et al., 1999; Green et al., 1999; cf. Verleger, 2002, for a comprehensive review), and very recently in a group of patients very similar to the one studied here (Gerloff et al., 2002). Evidence provided by these studies is extended by the present data in important ways because none of these studies made use of the fine-grained temporal resolution as given by the present approach. 4.2. Measurement of motor output Ipsilateral activation was here found under conditions where overt ‘mirror movements’, i.e. force exertion by the unaffected hand, did not occur. Any pressure by the uninvolved hand that exceeded 1 N led to rejection of the trial from analysis (6.5% of trials in the patients). This is certainly a more stringent control than the mere observation of patients’ behavior used by the previous studies on cerebral blood flow that found ipsilateral activation. As recently confirmed in a study on neurological patients, force-sensitive keys are an extremely precise and sensitive means to measure force output, in particular being sensitive to any accessory force exerted by the hand that is not cued to press (Franz and Miller, 2002). Confirming the sensitivity of force measurement, our patients differed in several parameters of their force output from the healthy control group. Since our instructions to participants focused on overt motor output, we did not control EMG activity of their finger-moving muscles. This might be done in future studies to clarify the question whether activation of the unaffected motor areas remains entirely prophylactic, without any peripheral excitation, or whether there is already some ‘leakage’ of this activation to peripheral muscle activity. Either outcome would be compatible with the present account of ipsilateral cortical activation. 4.3. Stimulus-locked EEG potentials The stimulus-locked data provided important additional information. In the absence of these data, it could have been argued that the patients’ late ipsilateral activation was only compensating for some earlier imbalance between contra- and
1475
ipsilateral sides, occurring already during response preparation. This, however, was not the case. We did find a difference in the slow preparatory negativity (CNV) between patients and control group, with CNV amplitudes being smaller in the patients particularly at parietal midline (replicating recent Bereitschaftspotential data, Platz et al., 2000) and above both motor cortices. This difference, however, was surprisingly unspecific, being present when preparing movements of both the affected and unaffected hand and being present at both C3 and C4, i.e. at sites overlying both the affected and unaffected motor cortex. Thus, there was no difference between contra- and ipsilateral side and between affected and unaffected side during response preparation. The paresis of the affected hand might have general impairing effects on patients’ manual mobility (Debaere et al., 2001), therefore, the generally lowered CNV above both motor cortices might be due to this general decrease of skillful use of the hands. Alternatively, keeping in mind that all amplitudes were measured relative to the baseline before the cue, the patients’ smaller increase might be smaller only relative to baseline. They might have maintained some tonic activation of their motor areas throughout the entire session such that they needed a lower increase of activation between cue and IS. Such tonic activation can only be measured with carefully prepared (and artifact-prone) DC recordings (e.g. Marshall et al., 1998) or with PET. Indeed, Weiller et al. (1993) reported relatively large activation of both motor systems in their patients, measured by PET. These latter considerations have an important implication for the main finding of the present study, the time-course of ipsilateral activity. It must be noted that this measure of ipsilateral activation is relative to the level at trial beginning. No statement can be made about the absolute amount of contra- and ipsilateral activation. Therefore, we do not know whether there is some tonic activation of the motor cortices throughout the entire session and whether this hypothetical tonic activation is larger in patients than in controls or vice versa, and whether it is larger contralaterally than ipsilaterally or vice versa. What can be said from the present data are statements about changes of contra- and ipsilateral activity in response to the requirement to use one or the other hand (stimulus-locked, evoked by the cue) and when actually preparing and executing the movement. What we found is that these relative changes when actually executing movements were larger on the ipsilateral side in patients when the movement is with the affected hand.
5. Conclusion In summary, extending the results of previous PET and fMRI studies, the present results for the first time show the time-course of contra- and ipsilateral activation. Starting after onset of contralateral activation and even after response onset, ipsilateral activation probably has no function in initiating movements of the affected hand.
1476
R. Verleger et al. / Clinical Neurophysiology 114 (2003) 1468–1476
Acknowledgements Hardware and software for force measurements were installed and developed by Piotr Jas´kowski, software for experimental control, recording, and data analysis by Edmund Wascher, software for mapping topographical distributions by Peter Trillenberg. Manfred Kaps gave an important impulse for planning this study, and Andreas Moser and Matthias Nitschke provided helpful advice. M.R. was supported by DFG grant Ve 110/10-3 to R.V.
References Anderer P, Semlitsch HV, Saletu B, Barbanoj MJ. Artifact processing in topographic mapping of electroencephalographic activity in neuropsychopharmacology. Psychiatr Res Neuroimage 1992;45:79–93. Ball T, Schreiber A, Feige B, Wagner M, Lu¨cking CH, Kristeva-Feige R. The role of higher-order motor areas in voluntary movement as revealed by high-resolution EEG and fMRI. Neuroimage 1999;10:682–94. Cao Y, D’Olhaberriague L, Vikingstad EM, Levine SR, Welch KMA. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke 1998;29:112–22. Carey JR, Kimberley TJ, Lewis SM, Auerbach EJ, Dorsey L, Rundquist P, Ugurbil K. Analysis of fMRI and finger tracking training in subjects with chronic stroke. Brain 2002;125:773–88. Chen R, Cohen LG, Hallett M. Nervous system reorganization following injury. Neuroscience 2002;111:761–73. Coles MGH. Modern mind-brain reading: psychophysiology, physiology, and cognition. Psychophysiology 1989;26:251–69. Cramer SC, Nelles G, Benson RR, Kaplan JD, Parker RA, Kwong KK, Kennedy DN, Finklestein SP, Rosen BR. A functional MRI study of subjects recovered from hemiparetic stroke. Stroke 1997;28:2518–27. Cramer SC, Finklestein SP, Schaechter JD, Bush G, Rosen BR. Activation of distinct motor cortex regions during ipsilateral and contralateral finger movements. J Neurophysiol 1999;81:383 –7. Debaere F, Van Assche D, Kiekens C, Verschueren SMP, Swinnen SP. Coordination of upper and lower limb segments: deficits on the ipsilesional side after unilateral stroke. Exp Brain Res 2001;141:519–29. Franz EA, Miller J. Effects of response readiness on reaction time and force output in people with Parkinson’s disease. Brain 2002;125:1733–50. Gerloff C, Sailer A, Braun C, Schmitt C, Bushara K, Wittenberg GF, Chen R, Ishii K, Waldvogel D, Wassermann EM, Cohen LG, Hallett M. Plasticity after stroke: homologous areas with non-homologous function. Paper presented at the Fourth Workshop on Cortical Plasticity, March 15–17, 2002, Castle of Schwetzingen, Germany. Gilbert PFC. An outline of brain function. Cogn Brain Res 2001;12:61– 74. Green JB, Bialy Y, Sora E, Ricamato A. High-resolution EEG in poststroke hemiparesis can identify ipsilateral generators during motor tasks. Stroke 1999;30:2659 –65. Kitamura J-I, Shibasaki H, Takeuchi T. Cortical potentials preceding voluntary elbow movement in recovered hemiparesis. Electroenceph clin Neurophysiol 1996;98:149 –56. Kopp B, Kunkel A, Mu¨hlnickel W, Villringer K, Taub E, Flor H. Plasticity in the motor system related to therapy-induced improvement of movement after stroke. NeuroReport 1999;10:807– 10. Kunieda T, Ikeda A, Ohara S, Yazawa S, Nagamine T, Taki W, Hashimoto N, Shibasaki H. Different activation of presupplementary motor area, supplementary motor area proper, and primary sensorimotor area, depending on the movement repetition rate in humans. Exp Brain Res 2000;135:163–72. Kutas M, Donchin E. Preparation to respond as manifested by movementrelated potentials. Brain Res 1980;202:95–115.
Liepert J, Weiller C. Mapping plastic brain changes after acute lesions. Curr Opin Neurol 1999;12:709 –13. Liepert J, Bauder H, Miltner WHR, Taub E, Weiller C. Treatment-induced cortical reorganization after stroke in humans. Stroke 2000;31:1210 –6. Marshall L, Mo¨lle M, Fehm HL, Born J. Scalp recorded direct current brain potentials during human sleep. Eur J Neurosci 1998;10:1167–78. 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. Nelles G, Spiekermann G, Jueptner M, Leonhardt G, Mu¨ller S, Gerhard H, Diener HC. Evolution of functional reorganization in hemiplegic stroke: a serial positron emission tomographic activation study. Ann Neurol 1999;46:901–9. Netz J, Lammers T, Ho¨mberg V. Reorganization of motor output in the non-affected hemisphere after stroke. Brain 1997;120:1579– 86. Platz T, Kim IH, Pintschovius H, Winter T, Kieselbach A, Villringer K, Kurth R, Mauritz K-H. Multimodal EEG analysis in man suggests impairment-specific changes in movement-related electric brain activity after stroke. Brain 2000;123:2475–90. Rizzolatti G, Luppino G, Matelli M. The organization of the cortical motor system: new concepts. Electroenceph clin Neurophysiol 1998;106:283–96. Seitz RJ, Azari NP. Cerebral reorganization in man after acquired lesions. In: Stefan H, Andermann F, Chauvel P, Shorvon S, editors. Advances in neurology, vol. 81. Philadelphia, PA: Lippincott Williams and Wilkins; 1999. p. 37– 47. Seitz RJ, Knorr U, Azari NP, Herzog H, Freund H-J. Visual network activation in recovery from sensorimotor stroke. Restorat Neurol Neurosci 1999;14:25–33. Slobounov SM, Ray WJ, Simon RF. Movement-related potentials accompanying unilateral finger movements with special reference to rate of force development. Psychophysiology 1998;35:537 –48. Small SL, Hlustik P, Noll DC, Genovese C, Solodkin A. Cerebellar hemispheric activation ipsilateral to the paretic hand correlates with functional recovery after stroke. Brain 2002;125:1544–57. Taub E, Crago JE, Uswatte G. Constraint-induced movement therapy, a new approach to treatment in physical rehabilitation. Rehabil Psychol 1998;43:152–70. Turton A, Wroe S, Trepte N, Fraser C, Lemon RN. Contralateral and ipsilateral EMG responses to transcranial magnetic stimulation during recovery of arm and hand function after stroke. Electroenceph clin Neurophysiol 1996;101:316–28. Verleger R. Event-related EEG potential research in neurological patients. In: Zani A, Proverbio AM, editors. The cognitive electrophysiology of mind and brain. San Diego, CA: Academic Press; 2002. p. 309 –41. Verleger R, Gasser T, Mo¨cks J. Correction of EOG artifacts in event related potentials of the EEG: aspects of reliability and validity. Psychophysiology 1982;19:472– 80. Verleger R, Wascher E, Wauschkuhn B, Jas´kowski P, Allouni B, Trillenberg P, Wessel K. Consequences of altered cerebellar input for the cortical regulation of motor coordination, as reflected in EEG potentials. Exp Brain Res 1999;127:409 –22. Verleger R, Wauschkuhn B, Van der Lubbe RHJ, Jas´kowski P, Trillenberg P. Posterior and anterior contributions of hand-movement preparation to late CNV. J Psychophysiol 2000;14:69 –86. Walter WG, Cooper R, Aldridge VJ, McCallum WC, Winter AL. Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature 1964;203:380–4. Weiller C, Ramsay SC, Wise RJS, Friston KJ, Frackowiak RSJ. Individual patterns of functional reorganization in the human cerebral cortex after capsular infarction. Ann Neurol 1993;33:181–9. Ziemann U, Ishii K, Borgheresi A, Yaseen Z, Battaglia F, Hallett M, Cincotta M, Wassermann EM. Dissociation of the pathways mediating ipsilateral and contralateral motor-evoked potentials in human hand and arm muscles. J Physiol 1999;518:895–906.