- Email: [email protected]
elbow motor task in stroke survivors
Journal of the Neurological Sciences 256 (2007) 21 – 29 www.elsevier.com/locate/jns
Abnormal cognitive planning and movement smoothness control for a complex shoulder/elbow motor task in stroke survivors Yin Fang a , Guang H. Yue a,b,c , Kenneth Hrovat d , Vinod Sahgal c , Janis J. Daly d,e,⁎ a
Department of Biomedical Engineering, The Cleveland Clinic Foundation, Cleveland, OH 44195, United States Orthopaedic Research Center, The Lerner Research Institute, The Cleveland Clinic Foundation, Cleveland, OH 44195, United States c Department of Physical Medicine and Rehabilitation, The Cleveland Clinic Foundation, Cleveland, OH 44195, United States d Louis Stokes Cleveland Department of Veterans Affairs Medical Center, Cleveland, OH 44106, United States Department of Neurology, Case Western Reserve School of Medicine, LSC VA Medical Center, Research Service, Cleveland, OH 44106, United States b
e
Received 25 August 2006; received in revised form 5 January 2007; accepted 23 January 2007 Available online 27 March 2007
Abstract Purpose: Cortical function is not well understood in stroke survivors with persistent dyscoordination. The study purpose was two-fold: 1) characterize cognitive planning time and cognitive effort level for a circle-drawing motor task in stroke survivors using shoulder/elbow muscles and 2) identify the relationship between cognitive effort level and movement smoothness. Methods: Twelve stroke survivors with shoulder/elbow coordination deficits (N12 mo) and eight controls were enrolled. The motor task was to draw a circle on a horizontal surface using only shoulder/elbow muscles. Outcome measures were: EEG-derived cognitive planning time, cognitive effort level, and movement smoothness. Comparisons between stroke and controls were made using t-tests. The Pearson's correlation model was analyzed to determine the relationship between movement smoothness and cognitive effort level. Results: Stroke subjects showed a statistically significant prolonged motor planning time versus controls for both lesion and non-lesion sides ( p = 0.013 and 0.049, respectively). They also showed a statistically significant elevated effort level versus controls for both sides (p = 0.016 and 0.013). The patients exhibited statistically significant poor movement smoothness in the medial/lateral and forward/backward movement directions versus controls (p = 0.035 and 0.037, respectively). For stroke, there was a significant correlation between cognitive effort level on the non-lesion side and smoothness of movement in the medial/lateral and forward/backward directions (r = 0.54, p = 0.036 and r = 0.76, p = 0.002, respectively). On the lesion side, results were mixed (r = 0.268, p = 0.2 r = 0.59, p = 0.023, respectively). Conclusions: Stroke survivors with upper limb motor deficits exhibit a longer cognitive planning time and elevated cognitive effort for performance of a complex shoulder/elbow motor coordination task. The elevated cognitive effort level was associated with poor (jerky) motor performance, suggesting a potential role of the CNS in controlling movement smoothness of the arm. © 2007 Elsevier B.V. All rights reserved. Keywords: Stroke; Cognition; Motor control; Electroencephalography; EEG; Motor planning; Motor-related cortical potential; MRCP; MCP; Coordination
1. Introduction In order to develop successful stroke rehabilitation therapies, it is essential to develop an understanding of how stroke can interfere with the cognitive processes that control motor function. Some stroke survivors exhibit poor control of movement ⁎ Corresponding author. Department of Neurology, Case Western Reserve School of Medicine, LSC VA Medical Center, Research Service, 151-W, 10701 E. Blvd., Cleveland, OH 44106, United States. Tel.: +1 216 791 3800x4129. E-mail address: [email protected] (J.J. Daly). 0022-510X/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jns.2007.01.078
smoothness [1], and some improve as spontaneous recovery proceeds [2]. Smoothness is a critical characteristic of coordinated human movements. Precise timing, accurate path maintenance, and smooth pursuit of the path are fundamental for an efficient and successful movement. Proper central nervous system (CNS) planning helps achieve movement smoothness during normal reaching movements [3– 5]. Normal motor planning optimizes the smoothness of different movement features, (e.g., minimum jerk profile [3], muscle force [4], and joint torque [5]). Although the details of each of these models [3–5] are different, the common findings
22
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
support the concept that CNS motor planning is responsible for programming and optimizing critical parameters to achieve an accurate and smooth movement. For circle and object drawing, similar evidence exists. For example, in normal subjects, there was a significant relationship between activation intensity of the motor cortex during visual presentation of geometrical shape and the speed of the upcoming movement in geometrical shape copying [6]. This finding suggests that the motor cortex plans the movement before the movement onset. Circle drawing is a complicated movement in that multiple muscles at multiple joints are sequentially activated. The smoothness of performance depends on the precise coordination of these activation patterns. In contrast, Rymer and Krylow [7] challenged the role of CNS planning regarding movement smoothness. They recorded movement trajectories produced at joints with loaded inertia while keeping the excitatory input constant. They concluded that the intrinsic mechanical properties of muscle may suffice to account for much of the smoothing of voluntary motion, obviating the need for an optimizing neural strategy. Although many stroke survivors do exhibit altered muscle tone characteristics, it is reasonable to hypothesize that poor movement smoothness in stroke is related at least to a large degree to impaired motor planning function at the cortical level, where the CNS injury occurred. Currently, few data are available which directly quantify the relationship between movement smoothness and the CNS motor planning in persons who suffered stroke. The study purpose was two-fold: 1) characterize cognitive planning time and cognitive effort level in stroke survivors for a complex shoulder/elbow motor coordination task (circle drawing); 2) identify the relationship between cognitive effort level and movement smoothness for a complex shoulder/elbow motor coordination task. 2. Methods 2.1. Subjects Twelve stroke patients with upper limb motor deficits and eight healthy controls were enrolled. This study was performed under the oversight of the Institutional Review Board of the Louis Stokes Cleveland VA Medical Center, and conducted according to the Declaration of Helsinki. All subjects gave informed consent prior to their participation. 2.2. Data recording 2.2.1. Apparatus The InMotion2 shoulder/elbow robot apparatus (Interactive Motion Technologies, Inc., Cambridge, MA) was used to standardize the task and acquire motor function data. The apparatus cradled, supported, and stabilized the patient′s forearm and hand in a position of neutral forearm, 30 degree wrist extension, and with fingers flexed around a cone. The apparatus allowed only shoulder/elbow movement in the horizontal plane. The apparatus measured position and time during the motor task (200 Hz).
2.2.2. Motor task and procedure For each trial, a “go” signal was provided for movement initiation. Subjects then traced a circle template (28 cm diameter) as smoothly as possible using shoulder and elbow muscles (left-limb affected, clockwise; right-limb affected, counter-clockwise). After completing each circle-tracing, the subject's hand rested for 10 s on the robot apparatus arm cradle at the beginning point of the movement and waited for the next “go” signal. Five sets of 10 repetitions were performed with a 2-min rest between sets. 2.2.3. EMG recording Surface EMG signals were recorded from the triceps and anterior deltoid muscles (1000 Hz) using bipolar electrodes. The reference electrode was applied at the lateral epicondyle. EMG signals were amplified (X1,000) and filtered (10–1000 Hz) using a Grass Neurodata system (Astro-Med, Inc., West Warwick, RI). The EMG was digitized using Spike2 (Cambridge Electronics Design, Cambridge, UK). 2.2.4. EEG recording A 64-channel NeuroScan EEG system (NeuroScan Labs, El Paso, TX) was used to acquire EEG signals from the surface of the scalp. The configuration of the electrode arrangement in the cap was based on the International 10–20 System [8]. The EEG electrodes were referenced to the common linked electrodes at the mastoid processes. An impedance map was displayed on a computer monitor. Impedance for all electrodes was monitored at a level below 10,000 Ω prior the initiation of data collection. All channels of the EEG signals were amplified (headbox X150, main amplifier X500), low-pass filtered (0–45 Hz), and digitized (1000 sample/s) using the NeuroScan Labs software. A customized goniometer was attached to the robot arm. The angle signal generated by a potentiometer in the goniometer was recorded in both the Spike2 and NeuroScan systems to synchronize the EMG and EEG signals. A trigger signal was generated each time an angle threshold was reached for the subsequent triggered-averaging of the EEG and peripheral signals (EMG and position). 2.3. Data processing and analysis 2.3.1. Movement smoothness (jerk cost) Movement smoothness was quantified in each of two orthogonal directions, medial/lateral and forward/backward on the work surface. Jerk can be calculated as the third derivative of displacement or the second derivative of velocity with respect to time [9], so for the medial/lateral component of movement, the jerk is given by: d3 xðtÞ dt 3 d2 Vx ðtÞ ¼ Vx WðtÞ ¼ dt 2
jerk ¼ xj ðtÞ ¼
where x(t) is the medial/lateral displacement, Vx(t) is the medial/lateral velocity and t is time. In order to quantify
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
23
Fig. 1. The example shows EEG data for a single control (A) and stroke (B) subject for motor related cortical potential (MRCP) from the C3 electrode (top graph in both A and B). The vertical arrow illustrates MRCP amplitude. The intersection of horizontal and diagonal shaded bands illustrates MRCP onset time, and the beginning of cognitive planning time. The end of cognitive planning time was defined as the onset of EMG. The EMG data (bottom graph in both A and B) is shown for single control subject A and stroke subject B from triceps brachii muscle. The EMG onset is given a time “0” and defines not only movement onset, but also the end of cognitive planning time. The duration of cognitive planning time is represented by the space between the two horizontal arrows facing each other.
smoothness using our velocity measurements, we calculated the jerk cost for the medial/lateral component of movement as the integral of the squared jerk as follows: Z t2 jerk cost ¼ ðVx WðtÞÞ2 dt t1
Likewise, we calculated jerk cost for the forward/backward component of movement as follows: Z
t2
jerk cost ¼ t1
ðVy WðtÞÞ2 dt
24
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
where Vy(t) is the forward/backward velocity t1 = 0, and t2 = time to complete circle. Smoother movement results in smaller jerk cost, while movements that are less smooth result in higher jerk cost [9]. 2.3.2. EEG processing The EEG recording was epoched for each trial. All epoched EEG data were inspected visually. Trials with eye blinks or other signal artifacts were excluded. The EEG and EMG data were then trigger-averaged across the remaining trials (28–40 trials per stroke subject), which yielded movement-related cortical potential (MRCP) and average EMG at each channel for a given subject. Electrodes over the sensorimotor region at lesion side and non-lesion side (C3 or C4 depending on the affected side) were selected for MRCP analysis because they show strong MRCP [10,11], and may reveal the differences between affected and unaffected sides. Generally, the MRCP is thought to represent cortical activity associated with planning and execution of the motor action [12]. It is usually difficult to identify a single point indicating the MRCP onset, despite a prominent MRCP waveform. We used a curve-fitting method: a straight line was drawn along the approximate shape of the baseline potential and the MRCP rising slope [10]; the intersection of the two lines was defined as the onset time of the MRCP (horizontal solid line and diagonal solid line; Fig. 1A and B). In some cases while the readiness potential component could be distinguished from later negative slope component, we used the slope of readiness potential as the MRCP rising slope. The amplitude of the MRCP was defined as the difference between peak and baseline (Fig. 1A, B), and represented cognitive effort level associated with planning and execution of the motor action [12]. The latency from the onset of MRCP to the onset of EMG (the earliest burst from among three muscles) was calculated, indicating the motor planning time (MPT) from beginning of the cortical activity to beginning of the muscle activity. 2.3.3. Two-dimensional (2-D) mapping of MRCP We constructed a 2-D map (NeuroScan software) of electrical potential of the brain, using mean MRCP amplitude across trials for stroke and control group. This “topographical” map was used to show spatial and temporal distributions of cortical electrical activities during the planning phase of the task. The analysis included a time period of 4900 ms, beginning from 4000 ms before the movement initiation to 900 ms after that. (On average, EMG began at 463 ms (for stroke) before the movement initiation.) A map was created at each 100 ms interval, resulting in a time series of maps. 2.4. Statistical analysis Descriptive statistics were generated for group mean and standard deviations (SD). Comparisons between the stroke and control groups were made using t-tests (alpha = .05). The
Table 1 Subject information Group
Number of subjects Stroke type a
Stroke Control
Time after stroke (years)
Age (years)
LI
LH
RI
RH
1–2
N2
48–59
60–72
6
2
3
1
5
7
5 5
7 3
a LI = Left ischemia, LH = left hemorrhagic, RI = right ischemia, RH = right hemorrhagic.
data were reported as mean ± SD unless otherwise specified. Pearson correlation was used to determine the level and significance of the relationship between the MRCP amplitude and movement smoothness. 3. Results Twelve stroke survivors were enrolled with mean age, 56.52 ± 7.73 years, range 48–72 years. Eight healthy control subjects were enrolled with age, 60.62 ± 6.25 years, range 53–69 years). The stroke survivors were 12–111 mo after stroke (Table 1). Stroke survivors exhibited statistically significant prolonged motor planning time (MPT) compared with healthy controls (Fig. 2A, B). The average MPT on the lesion side (from onset of the MRCP to onset of the EMG) for stroke was 2601 ± 499 ms and for controls was 1866 ± 481 ms ( p = 0.013). On the non-lesion side, the average MPT for stroke was 2277 ± 448 ms and for controls was 1759 ± 583 ms ( p = 0.049). Stroke survivors exhibited abnormally elevated cortical effort level (MRCP amplitude) compared with healthy controls (Fig. 2C, D). On the lesion side, the MRCP amplitude was 8.34 ± 3.61 μV for the stroke survivors and 5.24 ± 2.56 μV for the controls (p = 0.016). On the non-lesion side, the amplitude was 7.69 ± 2.99 μV and 4.69 ± 1.81 μV for the stroke survivors and control groups, respectively (p = 0.013). The 2-D mapping of MRCP (Fig. 3) suggested that the MRCP (negative potential, blue color) onset was prolonged for stroke (top figure) versus controls (bottom figure), and a greater brain area (covered by blue color) was involved in movement planning in stroke versus controls. Stroke survivors exhibited poor motor performance compared with healthy controls, according to the measurement of movement smoothness (jerk cost). Fig. 4 shows a comparison of a control subject example (Fig. 4A, C, E) and a stroke subject example (Fig. 4B, D, F). There was a statistically significant difference in the medial/lateral component of jerk cost for stroke versus controls: 727044 ± 857211 m 2 /s 5 and 161381 ± 337863 m2/s5, respectively ( p = 0.035; Fig. 2F). There was a statistically significant difference in the forward/backward component of jerk cost for stroke versus controls: 660,955 ± 941,364 m2 /s5 and 88,172 ± 64,647 m2/s5 , respectively ( p = 0.037; Fig. 2F).
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
25
Fig. 2. The statistically significant difference ( p b .05) is illustrated (asterisk between the columns) between stroke (cross-hatched column) and control (diagonal line column) subjects for the following measures: motor planning time for the lesion side (A) and for the non-lesion side (B); cognitive effort level (MRCP amplitude) on the lesion side (C) and the non-lesion side (D); and smoothness of movement (jerk cost) in the medial/lateral movement directions (E) and the forward/backward movement directions (F). The relationship is shown between the following measures: on the lesion side, MRCP amplitude and movement smoothness in the medial/lateral (G, r = 0.268, p = 0.2); on the lesion side, MRCP amplitude and forward/backward movement smoothness (H, r = 0.585; p = 0.023); on the non-lesion side, MRCP amplitude and med/lat movement smoothness (I, r = 0.536, p = 0.036); and on the non-lesion side, MRCP amplitude and forward backward movement smoothness (J, r = 0.76, p = 0.002).
26
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
Fig. 3. Group averaged 2-D MRCP amplitude maps. Each map represents MRCP amplitude data from 64 electrodes, beginning with the first map that occurred 4000 ms before the movement onset and the last map 900 ms after the movement onset. An MRCP map was created and is presented for every 100 ms. The red arrow (second row) indicates MRCP onset. The blue arrow on the fourth row indicates EMG onset, and the black arrow on the bottom row indicates the movement onset. The data suggest that duration of MRCP time (between MRCP onset and EMG onset was longer for stroke versus controls (longer duration of blue brain maps). Also the region of activated brain (blue area within a given map) was larger for stroke versus control during the planning phase of the movement.
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
27
Fig. 4. Single subject examples of movement trajectory in healthy control (A) and stroke (B), velocity profile in medial/lateral movement direction in control (C) and stroke (D), and velocity profile in forward/backward in control (E) and stroke (F).
Our data indicated that there were moderate to good and significant correlations between smoothness of movement (jerk cost) and cognitive effort level (MRCP amplitude) on the lesion side and non-lesion side for stroke survivors (Fig. 2H, I, J). On the non-lesion side, the correlation was 0.54 (p = 0.036) between the MRCP amplitude and smoothness of medial/lateral movement (Fig. 2I), and the correlation was 0.76 (p = 0.002) between MRCP amplitude and the smoothness of forward/backward movement components (Fig. 2J). On the lesion side, the correlation between cognitive effort level and movement smoothness in forward/ backward components was 0.59 (p = 0.02; Fig. 2H). For the lesion side, there was no significant correlation between cognitive effort level and smoothness of medial/lateral movement (r = 0.268; p = 0.2 (Fig. 2G). 4. Discussion Our results extended the literature by providing evidence that stroke survivors exhibited prolonged cognitive planning time for a complex series of movements involving con-
current and sequential shoulder and elbow movements. Our results of cortical signal measures are consistent with the work of others for stroke survivors that focused simply on the physical manifestation of movement delay, that is, delay in motor response time from the command to movement onset [13–15]. The prolongation of the planning time could arise from at least two sources: 1) time to schedule and plan; and 2) time for signal transmission from the CNS to the muscle. Our MPT measure was the duration of time from the MRCP onset to the EMG onset, including both sources. It is difficult to pinpoint the mechanisms that contribute to differences in MPT between healthy versus stroke subjects. In healthy controls, an easy, well-learned motor task is planned and controlled by the primary motor region; whereas, only more difficult or novel motor tasks are controlled by the secondary (e.g., pre-motor and supplementary motor) and association (e.g., cingulate and prefrontal) regions. For healthy controls, more time is required to plan a more difficult or novel motor task versus an easier task [16]. In contrast to healthy controls, stroke survivors not only exhibit activation of the primary motor region for seemingly simple
28
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
tasks, but they also activate the secondary and association regions [17] that are normally reserved for control of only more difficult or novel tasks in healthy controls. This distribution of the activation pattern suggests that for stroke survivors, a seemingly simple task may actually be perceived to be more difficult than would be the case for healthy controls. If a given task is perceived to be more difficult, stroke survivors could require a more prolonged motor planning time for that task. A prolonged planning time would then be consistent with the process used by healthy controls during planning for a more difficult task. Stroke survivors with motor impairment exhibited motor difficulty during the circle task, compared with the ease of performance exhibited by healthy control subjects. Therefore, it could have been perceived by stroke survivors as a more difficult motor task, possibly explaining the prolonged cognitive planning time. The study provided evidence that significantly elevated cognitive effort was expended in stroke survivors versus controls, during the planning of complex shoulder/elbow movements. Others showed that in healthy controls and other patient populations, more difficult movements were associated with higher MRCP amplitude and more errors [10,11,19]. Cognitive measures (timing, amplitude) before movement onset were related to the complexity, difficulty, speed, and amplitude of the upcoming movement [16,18,19]. In the current study of stroke survivors, in the presence of weakness and dyscoordination, the circle-drawing task was a difficult task, explaining the elevated cognitive effort level. This study showed a significant correlation between cognitive effort level and smoothness of movement: the jerkier the movement, the higher the cognitive effort level that was involved in planning the movement. Clearly, correlation does not support a cause and effect relationship, and so multiple explanations should be considered. Therefore, we will consider the possibility of covariation of the two variables, as well as possible cause and effect relationships. First, the cause of elevated effort level could be the perceived greater difficulty of the task. The difficulty of the task could arise from a number of sources, one of which could be the inability to perform the task in an accurate manner, which required smooth movement control as an inherent aspect of the task, execution of the circular movement pathway. The greater the impairment in movement smoothness control, the more difficult the task was perceived. That is, the stroke survivor with dyscoordination and/or weakness is very much aware, prior to the attempted movement, that it is going to be quite difficult to perform the circle-drawing movement pattern. The greater the dyscoordination and/or weakness, the more difficult it will be to activate and deactivate the combined shoulder/elbow muscles quickly enough and in the proper sequence to continually maintain the desired movement pathway. The greater these difficulties, the jerkier the movement, as motor control errors are made and sudden stops, starts, and abnormal movement reversals occur. Because the subject is aware of the looming difficulties in
maintaining the desired circular movement pathway, the motor control system may categorize the task as more difficult than would be the case for healthy controls. And consistent with healthy controls, the stroke survivor would then employ greater effort level for a perceived more difficult task. This explanation is consistent with the findings of others for healthy control performance of easy versus difficult motor tasks [10,11,16,18,19]. A second explanation for the correlation could be that elevated effort level and movement smoothness covary as a result of their relationship to another variable(s). For example, in a functional motor task, the actual function (e.g., reach for an object; bring food to the mouth, etc.) is the primary motor behavior goal. In contrast, the motor control system may assign movement smoothness as a secondary goal, especially if limitation of cognitive resources is an issue. The weakness or dyscoordination after stroke may render the cortical control of the task ineffective, interfering with both the primary goal of accurate muscle selection for the joint movements, as well as the secondary goal of movement smoothness, which entails finely graded activation/deactivation of shoulder and elbow muscles, as well as the accurate timing of sequential activations of muscle combinations required for the circle-drawing task. In compensating for the impairments, greater cognitive effort may be required for the primary task of accurate muscle selection for the joint movements. Though movement smoothness is also disrupted, movement smoothness may be uncompensated, if it is a secondary goal. In this case, both cognitive effort level and movement smoothness would covary with the magnitude of weakness and/or dyscoordination and would be correlated. Though this explanation could apply here, it would be strongest for a linear reaching movement to grasp an object or hit a target. In the current study, the task was to move in a prescribed circular pattern. This task requires a continuously curved movement pattern, which demands a continuous, smooth transition of graded muscle activations/deactivations of muscle combinations of shoulder and elbow joint muscles in the proper sequence and the proper instant in time. This particular task requires smooth movement in order to execute critical aspects of the task. A third possible explanation for the correlation between elevated effort level and movement smoothness is that the motor control system may categorize the circle-drawing task as a “novel” or not a well-learned task, since the stroke survivors were not able to perform the task well and had no reason to perform the required movements for this task for many months (Their poor control of the movements gave them no functional motivation). Many studies [e.g., [20,21]] have shown that the level of brain activation is higher during the early stage of learning a motor skill than the later stage. Stroke survivors also need greater effort to re-learn a relatively “new” motor skill compared to performing a learned one. Assignment of the motor task as novel or not welllearned may partially explain the greater effort level for those with more severe dyscoordination.
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
On the lesion side, the correlation of smoothness and effort level was relatively higher for the forward/backward movement component, versus the medial/lateral component (Fig. 2G compared to 2H). Since functional reaching movements utilize the forward/backward component to a greater degree than the medial/lateral component, there could be not only a greater level of skill control for the forward/ backward component, but also more accurate perception of the cognitive effort level that will be needed to control the movement. That is, for forward/backward movement, there was an accurately perceived need and greater measured cognitive effort for one who exhibited a more jerky, less smooth movement. And there was an accurately perceived need for less cognitive effort and lower measured cognitive effort for one who performed a smoother movement. Other factors may also have influenced the difference between medial/lateral and forward/backward jerk cost correlation with the cognitive effort. First, simultaneous two-joint (elbow and shoulder) movements occur during the forward/backward component, but the elbow joint contributes little (relative to shoulder) to the medial/lateral (shoulder abduction/adduction) component. The control of two joints is more complex than control of a single joint and that may have contributed to the stronger relationship in the forward/ backward direction of the movement. Second, the two-joint movements needed for the forward/backward component were “out-of-synergy” activities that involved flexion at one joint and extension at the other (shoulder/elbow). Stroke survivors with dyscoordination have greater difficulty with out-of-synergy movements [22] at two contiguous joints than with single joint movements. Third, during movement in the medial/lateral direction, both the forearm and upper arm were moved, causing greater inertia versus movement of mostly the forearm during forward/backward movement directions. If poor smoothness was caused by erratic or poorly controlled muscle force, then the larger inertia of the arm during medial/ lateral movements would tend to mask the jerkiness or reduce the movement error in the medial/lateral direction. The lack of association between cognitive effort level (medial/lateral component) and smoothness of movement, especially on the lesion side (Fig. 2G) could reflect inaccuracies in the perceived cognitive effort level needed. For example, some produced a higher measured cognitive effort level, but performed very smooth movement, while some produced a lower measured cognitive effort level, but performed very jerky, less smooth movement (Fig. 2G). The presence of the lesion in that hemisphere, could explain the lack of association between measured effort level and smoothness performance. 5. Conclusions Stroke survivors with upper limb motor deficits exhibit a longer cognitive planning time and elevated cognitive effort for performance of a complex shoulder/elbow motor coordination task. The elevated cognitive effort level was
29
associated with poor (jerky) motor performance, suggesting a potential role of the CNS in controlling movement smoothness of the arm. References [1] Krebs HI, Hogan N, Aisen ML, Volpe BT. Robot-aided neurorehabilitation. IEEE Trans Rehabil Eng 1998;6:75–87. [2] Rohrer B, Fasoli S, Krebs HI, Hughes R, Volpe B, Frontera WR, et al. Movement smoothness changes during stroke recovery. J Neurosci 2002;22(18):8297–304. [3] Hogan N. An organizing principal for a class of voluntary movements. J Neurosci 1984;4:2745–54. [4] Soechting JF, Flanders M. Movement planning: kinematics, dynamics, both or neither? In: Harris L, Jenkin M, editors. Vision and action. New York: Cambridge Univ. Press; 1998. p. 352–71. [5] Uno Y, Kawato M, Suzuki R. Formation and control of optimal trajectory in human multijoint arm movement: minimum torquechange model. Biol Cybern 1989;61:89–101. [6] Lewis SM, Jerde TA, Tzagarakis C, Georgopoulos MA, Tsekos N, Amirikian B, et al. Cerebellar activation during copying geometrical shapes. J Neurophysiol 2003;90(6):3874–87. [7] Rymer AM, Krylow WZ. Role of intrinsic muscle properties in producing smooth movements. IEEE Trans Biomed Eng 1997;44:165–76. [8] Jasper HH. Report of the committee on methods of clinical examination in electroencephalography. Appendix: the ten twenty electrode system of the international federation. Electroencephalogr Clin Neurophysiol 1958;10:370–5. [9] Hogan N. Adaptive control of mechanical impedance by coactivation of antagonist muscles. IEEE Trans Automat Contr 1984;29:681–90. [10] Fang Y, Siemionow V, Sahgal V, Xiong F, Yue GH. Greater movementrelated cortical potential during human eccentric versus concentric muscle contractions. J Neurophysiol 2001;86(4):1764–72. [11] Fang Y, Siemionow V, Sahgal V, Xiong F, Yue GH. Distinct brain activation patterns for human maximal voluntary eccentric and concentric muscle actions. Brain Res 2004;1023(2):200–12. [12] Hallett M. Movement-related cortical potentials. Electromyogr Clin Neurophysiol 1994;34:5–13. [13] Kaizer F, Korner-Bitensky N, Mayo N, Becker R, Coopersmith H. Response time of stroke patients to a visual stimulus. Stroke 1988;19:335–9. [14] Gerritsen MJ, Berg IJ, Deelman BG, Visser-Keizer AC, Meyboom-de Jong B. Speed of information processing after unilateral stroke. J Clin Exp Neuropsychol 2003;25(1):1–13. [15] Spikeman J, van Zomeren A, Deelman B. Deficits of attention after closedhead injury: slowness only? J Clin Exp Neuropsychol 1996;18:755–67. [16] Schreiber H, Lang M, Lang W, Kornhuber A, Heise B, Keidel M, et al. Frontal hemispheric differences in the Bereitschafts potential associated with writing and drawing. Hum Neurobiol 1983;2(3):197–202. [17] Calautti C, Baron JC. Functional neuroimaging studies of motor recovery after stroke in adults: a review. Stroke 2003;34(6):1553–66. [18] Yue GH, Liu JZ, Siemionow V, Ranganathan VK, Ng TC, Sahgal V. Brain activation during human finger extension and flexion movements. Brain Res 2000;856(1–2):291–300. [19] Yue GH, Ranganathan VK, Siemionow V, Liu JZ, Sahgal V. Evidence of inability to fully activate human limb muscle. Muscle Nerve 2000;23(3):376–84. [20] Jenkins IH, Brooks DJ, Nixon PD, Frackowiak RS, Passingham RE. Motor sequence learning: a study with positron emission tomography. J Neurosci 1994;14:3775–90. [21] Seitz RJ, Roland E, Bohm C, Greitz T, Stone-Elander S. Motor learning in man: a positron emission tomographic study. NeuroReport 1990;1:57–60. [22] Dewald JP, Pope PS, Given JD, Buchanan TS, Rymer WZ. Abnormal muscle coactivation patterns during isometric torque generation at the elbow and shoulder in hemiparetic subjects. Brain 1995;118:495–510.
Abnormal cognitive planning and movement smoothness control for a complex shoulder/elbow motor task in stroke survivors Yin Fang a , Guang H. Yue a,b,c , Kenneth Hrovat d , Vinod Sahgal c , Janis J. Daly d,e,⁎ a
Department of Biomedical Engineering, The Cleveland Clinic Foundation, Cleveland, OH 44195, United States Orthopaedic Research Center, The Lerner Research Institute, The Cleveland Clinic Foundation, Cleveland, OH 44195, United States c Department of Physical Medicine and Rehabilitation, The Cleveland Clinic Foundation, Cleveland, OH 44195, United States d Louis Stokes Cleveland Department of Veterans Affairs Medical Center, Cleveland, OH 44106, United States Department of Neurology, Case Western Reserve School of Medicine, LSC VA Medical Center, Research Service, Cleveland, OH 44106, United States b
e
Received 25 August 2006; received in revised form 5 January 2007; accepted 23 January 2007 Available online 27 March 2007
Abstract Purpose: Cortical function is not well understood in stroke survivors with persistent dyscoordination. The study purpose was two-fold: 1) characterize cognitive planning time and cognitive effort level for a circle-drawing motor task in stroke survivors using shoulder/elbow muscles and 2) identify the relationship between cognitive effort level and movement smoothness. Methods: Twelve stroke survivors with shoulder/elbow coordination deficits (N12 mo) and eight controls were enrolled. The motor task was to draw a circle on a horizontal surface using only shoulder/elbow muscles. Outcome measures were: EEG-derived cognitive planning time, cognitive effort level, and movement smoothness. Comparisons between stroke and controls were made using t-tests. The Pearson's correlation model was analyzed to determine the relationship between movement smoothness and cognitive effort level. Results: Stroke subjects showed a statistically significant prolonged motor planning time versus controls for both lesion and non-lesion sides ( p = 0.013 and 0.049, respectively). They also showed a statistically significant elevated effort level versus controls for both sides (p = 0.016 and 0.013). The patients exhibited statistically significant poor movement smoothness in the medial/lateral and forward/backward movement directions versus controls (p = 0.035 and 0.037, respectively). For stroke, there was a significant correlation between cognitive effort level on the non-lesion side and smoothness of movement in the medial/lateral and forward/backward directions (r = 0.54, p = 0.036 and r = 0.76, p = 0.002, respectively). On the lesion side, results were mixed (r = 0.268, p = 0.2 r = 0.59, p = 0.023, respectively). Conclusions: Stroke survivors with upper limb motor deficits exhibit a longer cognitive planning time and elevated cognitive effort for performance of a complex shoulder/elbow motor coordination task. The elevated cognitive effort level was associated with poor (jerky) motor performance, suggesting a potential role of the CNS in controlling movement smoothness of the arm. © 2007 Elsevier B.V. All rights reserved. Keywords: Stroke; Cognition; Motor control; Electroencephalography; EEG; Motor planning; Motor-related cortical potential; MRCP; MCP; Coordination
1. Introduction In order to develop successful stroke rehabilitation therapies, it is essential to develop an understanding of how stroke can interfere with the cognitive processes that control motor function. Some stroke survivors exhibit poor control of movement ⁎ Corresponding author. Department of Neurology, Case Western Reserve School of Medicine, LSC VA Medical Center, Research Service, 151-W, 10701 E. Blvd., Cleveland, OH 44106, United States. Tel.: +1 216 791 3800x4129. E-mail address: [email protected] (J.J. Daly). 0022-510X/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jns.2007.01.078
smoothness [1], and some improve as spontaneous recovery proceeds [2]. Smoothness is a critical characteristic of coordinated human movements. Precise timing, accurate path maintenance, and smooth pursuit of the path are fundamental for an efficient and successful movement. Proper central nervous system (CNS) planning helps achieve movement smoothness during normal reaching movements [3– 5]. Normal motor planning optimizes the smoothness of different movement features, (e.g., minimum jerk profile [3], muscle force [4], and joint torque [5]). Although the details of each of these models [3–5] are different, the common findings
22
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
support the concept that CNS motor planning is responsible for programming and optimizing critical parameters to achieve an accurate and smooth movement. For circle and object drawing, similar evidence exists. For example, in normal subjects, there was a significant relationship between activation intensity of the motor cortex during visual presentation of geometrical shape and the speed of the upcoming movement in geometrical shape copying [6]. This finding suggests that the motor cortex plans the movement before the movement onset. Circle drawing is a complicated movement in that multiple muscles at multiple joints are sequentially activated. The smoothness of performance depends on the precise coordination of these activation patterns. In contrast, Rymer and Krylow [7] challenged the role of CNS planning regarding movement smoothness. They recorded movement trajectories produced at joints with loaded inertia while keeping the excitatory input constant. They concluded that the intrinsic mechanical properties of muscle may suffice to account for much of the smoothing of voluntary motion, obviating the need for an optimizing neural strategy. Although many stroke survivors do exhibit altered muscle tone characteristics, it is reasonable to hypothesize that poor movement smoothness in stroke is related at least to a large degree to impaired motor planning function at the cortical level, where the CNS injury occurred. Currently, few data are available which directly quantify the relationship between movement smoothness and the CNS motor planning in persons who suffered stroke. The study purpose was two-fold: 1) characterize cognitive planning time and cognitive effort level in stroke survivors for a complex shoulder/elbow motor coordination task (circle drawing); 2) identify the relationship between cognitive effort level and movement smoothness for a complex shoulder/elbow motor coordination task. 2. Methods 2.1. Subjects Twelve stroke patients with upper limb motor deficits and eight healthy controls were enrolled. This study was performed under the oversight of the Institutional Review Board of the Louis Stokes Cleveland VA Medical Center, and conducted according to the Declaration of Helsinki. All subjects gave informed consent prior to their participation. 2.2. Data recording 2.2.1. Apparatus The InMotion2 shoulder/elbow robot apparatus (Interactive Motion Technologies, Inc., Cambridge, MA) was used to standardize the task and acquire motor function data. The apparatus cradled, supported, and stabilized the patient′s forearm and hand in a position of neutral forearm, 30 degree wrist extension, and with fingers flexed around a cone. The apparatus allowed only shoulder/elbow movement in the horizontal plane. The apparatus measured position and time during the motor task (200 Hz).
2.2.2. Motor task and procedure For each trial, a “go” signal was provided for movement initiation. Subjects then traced a circle template (28 cm diameter) as smoothly as possible using shoulder and elbow muscles (left-limb affected, clockwise; right-limb affected, counter-clockwise). After completing each circle-tracing, the subject's hand rested for 10 s on the robot apparatus arm cradle at the beginning point of the movement and waited for the next “go” signal. Five sets of 10 repetitions were performed with a 2-min rest between sets. 2.2.3. EMG recording Surface EMG signals were recorded from the triceps and anterior deltoid muscles (1000 Hz) using bipolar electrodes. The reference electrode was applied at the lateral epicondyle. EMG signals were amplified (X1,000) and filtered (10–1000 Hz) using a Grass Neurodata system (Astro-Med, Inc., West Warwick, RI). The EMG was digitized using Spike2 (Cambridge Electronics Design, Cambridge, UK). 2.2.4. EEG recording A 64-channel NeuroScan EEG system (NeuroScan Labs, El Paso, TX) was used to acquire EEG signals from the surface of the scalp. The configuration of the electrode arrangement in the cap was based on the International 10–20 System [8]. The EEG electrodes were referenced to the common linked electrodes at the mastoid processes. An impedance map was displayed on a computer monitor. Impedance for all electrodes was monitored at a level below 10,000 Ω prior the initiation of data collection. All channels of the EEG signals were amplified (headbox X150, main amplifier X500), low-pass filtered (0–45 Hz), and digitized (1000 sample/s) using the NeuroScan Labs software. A customized goniometer was attached to the robot arm. The angle signal generated by a potentiometer in the goniometer was recorded in both the Spike2 and NeuroScan systems to synchronize the EMG and EEG signals. A trigger signal was generated each time an angle threshold was reached for the subsequent triggered-averaging of the EEG and peripheral signals (EMG and position). 2.3. Data processing and analysis 2.3.1. Movement smoothness (jerk cost) Movement smoothness was quantified in each of two orthogonal directions, medial/lateral and forward/backward on the work surface. Jerk can be calculated as the third derivative of displacement or the second derivative of velocity with respect to time [9], so for the medial/lateral component of movement, the jerk is given by: d3 xðtÞ dt 3 d2 Vx ðtÞ ¼ Vx WðtÞ ¼ dt 2
jerk ¼ xj ðtÞ ¼
where x(t) is the medial/lateral displacement, Vx(t) is the medial/lateral velocity and t is time. In order to quantify
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
23
Fig. 1. The example shows EEG data for a single control (A) and stroke (B) subject for motor related cortical potential (MRCP) from the C3 electrode (top graph in both A and B). The vertical arrow illustrates MRCP amplitude. The intersection of horizontal and diagonal shaded bands illustrates MRCP onset time, and the beginning of cognitive planning time. The end of cognitive planning time was defined as the onset of EMG. The EMG data (bottom graph in both A and B) is shown for single control subject A and stroke subject B from triceps brachii muscle. The EMG onset is given a time “0” and defines not only movement onset, but also the end of cognitive planning time. The duration of cognitive planning time is represented by the space between the two horizontal arrows facing each other.
smoothness using our velocity measurements, we calculated the jerk cost for the medial/lateral component of movement as the integral of the squared jerk as follows: Z t2 jerk cost ¼ ðVx WðtÞÞ2 dt t1
Likewise, we calculated jerk cost for the forward/backward component of movement as follows: Z
t2
jerk cost ¼ t1
ðVy WðtÞÞ2 dt
24
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
where Vy(t) is the forward/backward velocity t1 = 0, and t2 = time to complete circle. Smoother movement results in smaller jerk cost, while movements that are less smooth result in higher jerk cost [9]. 2.3.2. EEG processing The EEG recording was epoched for each trial. All epoched EEG data were inspected visually. Trials with eye blinks or other signal artifacts were excluded. The EEG and EMG data were then trigger-averaged across the remaining trials (28–40 trials per stroke subject), which yielded movement-related cortical potential (MRCP) and average EMG at each channel for a given subject. Electrodes over the sensorimotor region at lesion side and non-lesion side (C3 or C4 depending on the affected side) were selected for MRCP analysis because they show strong MRCP [10,11], and may reveal the differences between affected and unaffected sides. Generally, the MRCP is thought to represent cortical activity associated with planning and execution of the motor action [12]. It is usually difficult to identify a single point indicating the MRCP onset, despite a prominent MRCP waveform. We used a curve-fitting method: a straight line was drawn along the approximate shape of the baseline potential and the MRCP rising slope [10]; the intersection of the two lines was defined as the onset time of the MRCP (horizontal solid line and diagonal solid line; Fig. 1A and B). In some cases while the readiness potential component could be distinguished from later negative slope component, we used the slope of readiness potential as the MRCP rising slope. The amplitude of the MRCP was defined as the difference between peak and baseline (Fig. 1A, B), and represented cognitive effort level associated with planning and execution of the motor action [12]. The latency from the onset of MRCP to the onset of EMG (the earliest burst from among three muscles) was calculated, indicating the motor planning time (MPT) from beginning of the cortical activity to beginning of the muscle activity. 2.3.3. Two-dimensional (2-D) mapping of MRCP We constructed a 2-D map (NeuroScan software) of electrical potential of the brain, using mean MRCP amplitude across trials for stroke and control group. This “topographical” map was used to show spatial and temporal distributions of cortical electrical activities during the planning phase of the task. The analysis included a time period of 4900 ms, beginning from 4000 ms before the movement initiation to 900 ms after that. (On average, EMG began at 463 ms (for stroke) before the movement initiation.) A map was created at each 100 ms interval, resulting in a time series of maps. 2.4. Statistical analysis Descriptive statistics were generated for group mean and standard deviations (SD). Comparisons between the stroke and control groups were made using t-tests (alpha = .05). The
Table 1 Subject information Group
Number of subjects Stroke type a
Stroke Control
Time after stroke (years)
Age (years)
LI
LH
RI
RH
1–2
N2
48–59
60–72
6
2
3
1
5
7
5 5
7 3
a LI = Left ischemia, LH = left hemorrhagic, RI = right ischemia, RH = right hemorrhagic.
data were reported as mean ± SD unless otherwise specified. Pearson correlation was used to determine the level and significance of the relationship between the MRCP amplitude and movement smoothness. 3. Results Twelve stroke survivors were enrolled with mean age, 56.52 ± 7.73 years, range 48–72 years. Eight healthy control subjects were enrolled with age, 60.62 ± 6.25 years, range 53–69 years). The stroke survivors were 12–111 mo after stroke (Table 1). Stroke survivors exhibited statistically significant prolonged motor planning time (MPT) compared with healthy controls (Fig. 2A, B). The average MPT on the lesion side (from onset of the MRCP to onset of the EMG) for stroke was 2601 ± 499 ms and for controls was 1866 ± 481 ms ( p = 0.013). On the non-lesion side, the average MPT for stroke was 2277 ± 448 ms and for controls was 1759 ± 583 ms ( p = 0.049). Stroke survivors exhibited abnormally elevated cortical effort level (MRCP amplitude) compared with healthy controls (Fig. 2C, D). On the lesion side, the MRCP amplitude was 8.34 ± 3.61 μV for the stroke survivors and 5.24 ± 2.56 μV for the controls (p = 0.016). On the non-lesion side, the amplitude was 7.69 ± 2.99 μV and 4.69 ± 1.81 μV for the stroke survivors and control groups, respectively (p = 0.013). The 2-D mapping of MRCP (Fig. 3) suggested that the MRCP (negative potential, blue color) onset was prolonged for stroke (top figure) versus controls (bottom figure), and a greater brain area (covered by blue color) was involved in movement planning in stroke versus controls. Stroke survivors exhibited poor motor performance compared with healthy controls, according to the measurement of movement smoothness (jerk cost). Fig. 4 shows a comparison of a control subject example (Fig. 4A, C, E) and a stroke subject example (Fig. 4B, D, F). There was a statistically significant difference in the medial/lateral component of jerk cost for stroke versus controls: 727044 ± 857211 m 2 /s 5 and 161381 ± 337863 m2/s5, respectively ( p = 0.035; Fig. 2F). There was a statistically significant difference in the forward/backward component of jerk cost for stroke versus controls: 660,955 ± 941,364 m2 /s5 and 88,172 ± 64,647 m2/s5 , respectively ( p = 0.037; Fig. 2F).
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
25
Fig. 2. The statistically significant difference ( p b .05) is illustrated (asterisk between the columns) between stroke (cross-hatched column) and control (diagonal line column) subjects for the following measures: motor planning time for the lesion side (A) and for the non-lesion side (B); cognitive effort level (MRCP amplitude) on the lesion side (C) and the non-lesion side (D); and smoothness of movement (jerk cost) in the medial/lateral movement directions (E) and the forward/backward movement directions (F). The relationship is shown between the following measures: on the lesion side, MRCP amplitude and movement smoothness in the medial/lateral (G, r = 0.268, p = 0.2); on the lesion side, MRCP amplitude and forward/backward movement smoothness (H, r = 0.585; p = 0.023); on the non-lesion side, MRCP amplitude and med/lat movement smoothness (I, r = 0.536, p = 0.036); and on the non-lesion side, MRCP amplitude and forward backward movement smoothness (J, r = 0.76, p = 0.002).
26
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
Fig. 3. Group averaged 2-D MRCP amplitude maps. Each map represents MRCP amplitude data from 64 electrodes, beginning with the first map that occurred 4000 ms before the movement onset and the last map 900 ms after the movement onset. An MRCP map was created and is presented for every 100 ms. The red arrow (second row) indicates MRCP onset. The blue arrow on the fourth row indicates EMG onset, and the black arrow on the bottom row indicates the movement onset. The data suggest that duration of MRCP time (between MRCP onset and EMG onset was longer for stroke versus controls (longer duration of blue brain maps). Also the region of activated brain (blue area within a given map) was larger for stroke versus control during the planning phase of the movement.
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
27
Fig. 4. Single subject examples of movement trajectory in healthy control (A) and stroke (B), velocity profile in medial/lateral movement direction in control (C) and stroke (D), and velocity profile in forward/backward in control (E) and stroke (F).
Our data indicated that there were moderate to good and significant correlations between smoothness of movement (jerk cost) and cognitive effort level (MRCP amplitude) on the lesion side and non-lesion side for stroke survivors (Fig. 2H, I, J). On the non-lesion side, the correlation was 0.54 (p = 0.036) between the MRCP amplitude and smoothness of medial/lateral movement (Fig. 2I), and the correlation was 0.76 (p = 0.002) between MRCP amplitude and the smoothness of forward/backward movement components (Fig. 2J). On the lesion side, the correlation between cognitive effort level and movement smoothness in forward/ backward components was 0.59 (p = 0.02; Fig. 2H). For the lesion side, there was no significant correlation between cognitive effort level and smoothness of medial/lateral movement (r = 0.268; p = 0.2 (Fig. 2G). 4. Discussion Our results extended the literature by providing evidence that stroke survivors exhibited prolonged cognitive planning time for a complex series of movements involving con-
current and sequential shoulder and elbow movements. Our results of cortical signal measures are consistent with the work of others for stroke survivors that focused simply on the physical manifestation of movement delay, that is, delay in motor response time from the command to movement onset [13–15]. The prolongation of the planning time could arise from at least two sources: 1) time to schedule and plan; and 2) time for signal transmission from the CNS to the muscle. Our MPT measure was the duration of time from the MRCP onset to the EMG onset, including both sources. It is difficult to pinpoint the mechanisms that contribute to differences in MPT between healthy versus stroke subjects. In healthy controls, an easy, well-learned motor task is planned and controlled by the primary motor region; whereas, only more difficult or novel motor tasks are controlled by the secondary (e.g., pre-motor and supplementary motor) and association (e.g., cingulate and prefrontal) regions. For healthy controls, more time is required to plan a more difficult or novel motor task versus an easier task [16]. In contrast to healthy controls, stroke survivors not only exhibit activation of the primary motor region for seemingly simple
28
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
tasks, but they also activate the secondary and association regions [17] that are normally reserved for control of only more difficult or novel tasks in healthy controls. This distribution of the activation pattern suggests that for stroke survivors, a seemingly simple task may actually be perceived to be more difficult than would be the case for healthy controls. If a given task is perceived to be more difficult, stroke survivors could require a more prolonged motor planning time for that task. A prolonged planning time would then be consistent with the process used by healthy controls during planning for a more difficult task. Stroke survivors with motor impairment exhibited motor difficulty during the circle task, compared with the ease of performance exhibited by healthy control subjects. Therefore, it could have been perceived by stroke survivors as a more difficult motor task, possibly explaining the prolonged cognitive planning time. The study provided evidence that significantly elevated cognitive effort was expended in stroke survivors versus controls, during the planning of complex shoulder/elbow movements. Others showed that in healthy controls and other patient populations, more difficult movements were associated with higher MRCP amplitude and more errors [10,11,19]. Cognitive measures (timing, amplitude) before movement onset were related to the complexity, difficulty, speed, and amplitude of the upcoming movement [16,18,19]. In the current study of stroke survivors, in the presence of weakness and dyscoordination, the circle-drawing task was a difficult task, explaining the elevated cognitive effort level. This study showed a significant correlation between cognitive effort level and smoothness of movement: the jerkier the movement, the higher the cognitive effort level that was involved in planning the movement. Clearly, correlation does not support a cause and effect relationship, and so multiple explanations should be considered. Therefore, we will consider the possibility of covariation of the two variables, as well as possible cause and effect relationships. First, the cause of elevated effort level could be the perceived greater difficulty of the task. The difficulty of the task could arise from a number of sources, one of which could be the inability to perform the task in an accurate manner, which required smooth movement control as an inherent aspect of the task, execution of the circular movement pathway. The greater the impairment in movement smoothness control, the more difficult the task was perceived. That is, the stroke survivor with dyscoordination and/or weakness is very much aware, prior to the attempted movement, that it is going to be quite difficult to perform the circle-drawing movement pattern. The greater the dyscoordination and/or weakness, the more difficult it will be to activate and deactivate the combined shoulder/elbow muscles quickly enough and in the proper sequence to continually maintain the desired movement pathway. The greater these difficulties, the jerkier the movement, as motor control errors are made and sudden stops, starts, and abnormal movement reversals occur. Because the subject is aware of the looming difficulties in
maintaining the desired circular movement pathway, the motor control system may categorize the task as more difficult than would be the case for healthy controls. And consistent with healthy controls, the stroke survivor would then employ greater effort level for a perceived more difficult task. This explanation is consistent with the findings of others for healthy control performance of easy versus difficult motor tasks [10,11,16,18,19]. A second explanation for the correlation could be that elevated effort level and movement smoothness covary as a result of their relationship to another variable(s). For example, in a functional motor task, the actual function (e.g., reach for an object; bring food to the mouth, etc.) is the primary motor behavior goal. In contrast, the motor control system may assign movement smoothness as a secondary goal, especially if limitation of cognitive resources is an issue. The weakness or dyscoordination after stroke may render the cortical control of the task ineffective, interfering with both the primary goal of accurate muscle selection for the joint movements, as well as the secondary goal of movement smoothness, which entails finely graded activation/deactivation of shoulder and elbow muscles, as well as the accurate timing of sequential activations of muscle combinations required for the circle-drawing task. In compensating for the impairments, greater cognitive effort may be required for the primary task of accurate muscle selection for the joint movements. Though movement smoothness is also disrupted, movement smoothness may be uncompensated, if it is a secondary goal. In this case, both cognitive effort level and movement smoothness would covary with the magnitude of weakness and/or dyscoordination and would be correlated. Though this explanation could apply here, it would be strongest for a linear reaching movement to grasp an object or hit a target. In the current study, the task was to move in a prescribed circular pattern. This task requires a continuously curved movement pattern, which demands a continuous, smooth transition of graded muscle activations/deactivations of muscle combinations of shoulder and elbow joint muscles in the proper sequence and the proper instant in time. This particular task requires smooth movement in order to execute critical aspects of the task. A third possible explanation for the correlation between elevated effort level and movement smoothness is that the motor control system may categorize the circle-drawing task as a “novel” or not a well-learned task, since the stroke survivors were not able to perform the task well and had no reason to perform the required movements for this task for many months (Their poor control of the movements gave them no functional motivation). Many studies [e.g., [20,21]] have shown that the level of brain activation is higher during the early stage of learning a motor skill than the later stage. Stroke survivors also need greater effort to re-learn a relatively “new” motor skill compared to performing a learned one. Assignment of the motor task as novel or not welllearned may partially explain the greater effort level for those with more severe dyscoordination.
Y. Fang et al. / Journal of the Neurological Sciences 256 (2007) 21–29
On the lesion side, the correlation of smoothness and effort level was relatively higher for the forward/backward movement component, versus the medial/lateral component (Fig. 2G compared to 2H). Since functional reaching movements utilize the forward/backward component to a greater degree than the medial/lateral component, there could be not only a greater level of skill control for the forward/ backward component, but also more accurate perception of the cognitive effort level that will be needed to control the movement. That is, for forward/backward movement, there was an accurately perceived need and greater measured cognitive effort for one who exhibited a more jerky, less smooth movement. And there was an accurately perceived need for less cognitive effort and lower measured cognitive effort for one who performed a smoother movement. Other factors may also have influenced the difference between medial/lateral and forward/backward jerk cost correlation with the cognitive effort. First, simultaneous two-joint (elbow and shoulder) movements occur during the forward/backward component, but the elbow joint contributes little (relative to shoulder) to the medial/lateral (shoulder abduction/adduction) component. The control of two joints is more complex than control of a single joint and that may have contributed to the stronger relationship in the forward/ backward direction of the movement. Second, the two-joint movements needed for the forward/backward component were “out-of-synergy” activities that involved flexion at one joint and extension at the other (shoulder/elbow). Stroke survivors with dyscoordination have greater difficulty with out-of-synergy movements [22] at two contiguous joints than with single joint movements. Third, during movement in the medial/lateral direction, both the forearm and upper arm were moved, causing greater inertia versus movement of mostly the forearm during forward/backward movement directions. If poor smoothness was caused by erratic or poorly controlled muscle force, then the larger inertia of the arm during medial/ lateral movements would tend to mask the jerkiness or reduce the movement error in the medial/lateral direction. The lack of association between cognitive effort level (medial/lateral component) and smoothness of movement, especially on the lesion side (Fig. 2G) could reflect inaccuracies in the perceived cognitive effort level needed. For example, some produced a higher measured cognitive effort level, but performed very smooth movement, while some produced a lower measured cognitive effort level, but performed very jerky, less smooth movement (Fig. 2G). The presence of the lesion in that hemisphere, could explain the lack of association between measured effort level and smoothness performance. 5. Conclusions Stroke survivors with upper limb motor deficits exhibit a longer cognitive planning time and elevated cognitive effort for performance of a complex shoulder/elbow motor coordination task. The elevated cognitive effort level was
29
associated with poor (jerky) motor performance, suggesting a potential role of the CNS in controlling movement smoothness of the arm. References [1] Krebs HI, Hogan N, Aisen ML, Volpe BT. Robot-aided neurorehabilitation. IEEE Trans Rehabil Eng 1998;6:75–87. [2] Rohrer B, Fasoli S, Krebs HI, Hughes R, Volpe B, Frontera WR, et al. Movement smoothness changes during stroke recovery. J Neurosci 2002;22(18):8297–304. [3] Hogan N. An organizing principal for a class of voluntary movements. J Neurosci 1984;4:2745–54. [4] Soechting JF, Flanders M. Movement planning: kinematics, dynamics, both or neither? In: Harris L, Jenkin M, editors. Vision and action. New York: Cambridge Univ. Press; 1998. p. 352–71. [5] Uno Y, Kawato M, Suzuki R. Formation and control of optimal trajectory in human multijoint arm movement: minimum torquechange model. Biol Cybern 1989;61:89–101. [6] Lewis SM, Jerde TA, Tzagarakis C, Georgopoulos MA, Tsekos N, Amirikian B, et al. Cerebellar activation during copying geometrical shapes. J Neurophysiol 2003;90(6):3874–87. [7] Rymer AM, Krylow WZ. Role of intrinsic muscle properties in producing smooth movements. IEEE Trans Biomed Eng 1997;44:165–76. [8] Jasper HH. Report of the committee on methods of clinical examination in electroencephalography. Appendix: the ten twenty electrode system of the international federation. Electroencephalogr Clin Neurophysiol 1958;10:370–5. [9] Hogan N. Adaptive control of mechanical impedance by coactivation of antagonist muscles. IEEE Trans Automat Contr 1984;29:681–90. [10] Fang Y, Siemionow V, Sahgal V, Xiong F, Yue GH. Greater movementrelated cortical potential during human eccentric versus concentric muscle contractions. J Neurophysiol 2001;86(4):1764–72. [11] Fang Y, Siemionow V, Sahgal V, Xiong F, Yue GH. Distinct brain activation patterns for human maximal voluntary eccentric and concentric muscle actions. Brain Res 2004;1023(2):200–12. [12] Hallett M. Movement-related cortical potentials. Electromyogr Clin Neurophysiol 1994;34:5–13. [13] Kaizer F, Korner-Bitensky N, Mayo N, Becker R, Coopersmith H. Response time of stroke patients to a visual stimulus. Stroke 1988;19:335–9. [14] Gerritsen MJ, Berg IJ, Deelman BG, Visser-Keizer AC, Meyboom-de Jong B. Speed of information processing after unilateral stroke. J Clin Exp Neuropsychol 2003;25(1):1–13. [15] Spikeman J, van Zomeren A, Deelman B. Deficits of attention after closedhead injury: slowness only? J Clin Exp Neuropsychol 1996;18:755–67. [16] Schreiber H, Lang M, Lang W, Kornhuber A, Heise B, Keidel M, et al. Frontal hemispheric differences in the Bereitschafts potential associated with writing and drawing. Hum Neurobiol 1983;2(3):197–202. [17] Calautti C, Baron JC. Functional neuroimaging studies of motor recovery after stroke in adults: a review. Stroke 2003;34(6):1553–66. [18] Yue GH, Liu JZ, Siemionow V, Ranganathan VK, Ng TC, Sahgal V. Brain activation during human finger extension and flexion movements. Brain Res 2000;856(1–2):291–300. [19] Yue GH, Ranganathan VK, Siemionow V, Liu JZ, Sahgal V. Evidence of inability to fully activate human limb muscle. Muscle Nerve 2000;23(3):376–84. [20] Jenkins IH, Brooks DJ, Nixon PD, Frackowiak RS, Passingham RE. Motor sequence learning: a study with positron emission tomography. J Neurosci 1994;14:3775–90. [21] Seitz RJ, Roland E, Bohm C, Greitz T, Stone-Elander S. Motor learning in man: a positron emission tomographic study. NeuroReport 1990;1:57–60. [22] Dewald JP, Pope PS, Given JD, Buchanan TS, Rymer WZ. Abnormal muscle coactivation patterns during isometric torque generation at the elbow and shoulder in hemiparetic subjects. Brain 1995;118:495–510.