Journal of Motor Behavior, 2007, Vol. 39, No. 2, 103–114 Copyright © 2007 Heldref Publications
Planning and Control of Sequential Rapid Aiming in Adults With Parkinson’s Disease A. L. Smiley-Oyen K. A. Lowry
J. P. Kerr Voxel Research Ames, Iowa
Motor Control and Learning Research Laboratory Department of Health and Human Performance Iowa State University, Ames
ABSTRACT. Eight people with Parkinson’s disease (PD), 8 agematched older adults, and 8 young adults executed 3-dimensional rapid aiming movements to 1, 3, 5, and 7 targets. Reaction time, flight time, and time after peak velocity to the 1st target indicated that both neurologically healthy groups implemented a plan on the basis of anticipation of upcoming targets, whereas the PD group did not. One suggested reason for the PD group’s deficiency in anticipatory control is the greater variability in their initial force impulse. Although the PD group scaled peak velocity and time to peak velocity similarly to the other groups, their coefficients of variation were greater, making consistent prediction of the movement outcome difficult and thus making it less advantageous to plan too far in advance. A 2nd finding was that the PD group exhibited increased slowing in time after peak velocity in the final segments of the longest sequence, whereas the other 2 groups did not. The increased slowing could be the result of a different movement strategy, increased difficulty modulating the agonist and antagonist muscle groups later in the sequence, or both. The authors conclude that people with PD use more segmented planning and control strategies than do neurologically healthy older and young adults when executing movement sequences and that the locus of increased bradykinesia in longer sequences is in the deceleration phase of movement.
performing a movement sequence than when performing discrete movements, and they exhibit longer pause times between elements of the sequence (Agostino, Berardelli, Formica, Accornero, & Manfredi, 1992; Benecke, Rothwell, Dick, Day, & Marsden, 1987; Harrington & Haaland, 1991; Weiss, Stelmach, & Hefter, 1997). They also exhibit an increase in bradykinesia, that is, movement slowness, in the later segments of a sequence (Agostino et al., 1992; Agostino et al., 1994; Harrington & Haaland). One reason that execution of movement sequences may be different in PD is that they are planned differently. One can measure planning on the basis of changes in reaction time (RT, i.e., the time required to preprogram the movement); as the number of elements in a rapidly executed series of discrete movements increases, so does RT. Investigators have concluded that as this type of sequence increases in number of elements, the movement is more complex; and therefore more neurological and psychological processing time is required before the movement is initiated (Christina, 1992; Fischman, 1984; Garcia-Colera & Semjen, 1987; Gordon & Meyer, 1987; Henry & Rogers, 1960; Lajoie & Franks, 1997; Sidaway, 1991; Sidaway, Christina, & Shea, 1988; Smiley-Oyen & Worringham, 2001; Sternberg, Monsell, Knoll, & Wright, 1978). A longer RT for longer sequences indicates that young adults plan a movement sequence in an integrative fashion, taking into account many segments, not just the first one or two segments. People with PD exhibit increased preprogramming time comparable with that of age-matched controls for two- and
Key words: anticipation, basal ganglia, motor planning, sequence, variability
I
t is well established that people with Parkinson’s disease (PD) exhibit difficulties with movement sequencing in such tasks as handwriting (Phillips, Chiu, Bradshaw, & Iansek, 1995; Teulings, Contreras-Vidal, Stelmach, & Adler, 1997), buttoning (Rocca, Maraganore, McDonnell, & Schaid, 1998; Soliveri, Brown, Jahanshahi, & Marsden, 1992), keyboarding (Freund, 1989; Longstreth, Nelson, Linde, & Munoz, 1992), and even finger tapping (Camicioli, Grossmann, Spencer, Hudnell, & Anger, 2001; Harrington, Haaland, & Hermanowicz, 1998). People with PD move more slowly when
Correspondence address: A. L. Smiley-Oyen, 244 Forker Building, Department of Health and Human Performance, Iowa State University, Ames, IA 50011, USA. E-mail address: asmiley @iastate.edu 103
A. L. Smiley-Oyen, K. A. Lowry, & J. P. Kerr
three-segment sequences (Behrman, Cauraugh, & Light, 2000; Harrington & Haaland, 1991; Rafal, Inhoff, Friedman, & Berstein, 1987; Reed & Franks, 1998). But, unlike age-matched controls, RT for the PD group does not increase when five elements are added to a sequence (Harrington & Haaland; Stelmach, Worringham, & Strand, 1987). Thus, they do not use preprogramming to the same degree for longer movement sequences, which indicates that they may approach the control of a movement sequence in a more segmented, less integrated, fashion. Another way to examine how a movement sequence is planned is to observe how the movement itself is influenced by changes in the sequence. For example, a unique element in a sequence (such as a smaller target) affects not only the movement to the small target but also movement to the surrounding targets (Rand, Alberts, Stelmach, & Bloedel, 1997; Rand & Stelmach, 2000; Short, Fischman, & Wang, 1996; Sidaway, Sekiya, & Fairweather, 1995; Smiley-Oyen & Worringham, 1996). Another example is that movement time to the first target increases as targets are added to the sequence (Chamberlin & Magill, 1989; Fischman & Reeve, 1992; Smiley-Oyen & Worringham, 2001). The results of the cited research indicated that neurologically healthy people plan for more than just the immediate segment being executed; rather, they incorporate integrative planning across segments. People with PD exhibit that type of integrative planning in short sequences. The first movement of a two-segment sequence was influenced by the size of the second target (Weiss et al., 1997) and by the distance to the second object (Gentilucci & Negrotti, 1999) similarly in PD patients and in age-matched controls. But another study’s results indicated that a small target (high accuracy demand) in the first of two targets causes the PD group to execute the two segments more independently (Rand, Van Gemmert, & Stelmach, 2002). More work is needed, especially in longer sequences, to enable us to better understand the degree to which people with PD use integrative planning and to examine the nature of the increased slowing later in a movement sequence. Our purpose in this study was to examine how people with PD plan and control a movement sequence as we increased the complexity of the movement by adding segments to the sequence. We compared their performance with that of agematched and young neurologically healthy adults to better understand the impact of both healthy aging and aging influenced by PD. To provide a better understanding of how the sequence is controlled, we parsed each segment into acceleration and deceleration phases. Parsing each segment in that manner reflects changes in the initial force impulse separate from the process of honing-in on the target. In addition, we used a velocity symmetry ratio (acceleration phase divided by deceleration phase) to assess the relative contribution of each phase within a segment. Young adults exhibit a symmetrical curve for discrete aiming, with approximately 50% 104
of the movement in each of the two phases, whereas older adults require a longer deceleration phase (Brown, 1996). In addition to examining performance within each condition, we used RT and kinematics of Segment 1 across conditions to examine the degree to which planning is approached in an integrative manner. An increase in the initial RT as targets are added would strongly suggest that during preprogramming, the motor system takes into account the overall movement. Likewise, if time to the first target increases and kinematics are adjusted with the addition of targets, then that would show clearly that the initial movement is being influenced by the anticipation of the upcoming targets. Thus, scaling in those measurements with the addition of targets supports the position that participants who employ those strategies are integrating aspects of the later movements into the planning and execution of the movement to the first target. Method Participants There were three groups of 8 participants (5 men and 3 women) in this study. One consisted of people with PD (mean age = 66 years, SD = 4.9 years). The second was an age-matched neurologically healthy older adult (OA) group (mean age = 65.1 years, SD = 3.3 years). The mean age of the third group, which consisted of young adults (YA), was 23.1 years (SD = 3 years). Patients had no greater than moderate bradykinesia, rigidity, or tremor; mild or no manifestations of cognitive impairment; no history of drug or alcohol abuse; and no neuropathology other than PD. Clinical evaluations were obtained for all PD patients (see Table 1). We tested patients while they were on medication and during the time when the medication was most effective. The university’s Institutional Review Board approved the study. Participants’ consent was obtained before their inclusion in the study and according to the Declaration of Helsinki. A comparison of the groups is shown in Table 2. Their Mini-Mental State Examination scores (MMSE; Folstein, Folstein, & McHugh, 1975) were not significantly different, although one person with PD had a score of 24, indicative of mild cognitive impairment. All other MMSE scores were within the normal range. On the basis of the Beck Depression Inventory-II (Beck & Steer, 1987), we found no significant differences in depression, although one person with PD was categorized as having moderate depression (score of 22), and one healthy older adult was categorized as having mild to moderate depression (score of 14). As expected, the PD group performed significantly slower than the other two groups on the Purdue Pegboard and the Reciprocal Tap tasks. Apparatus Targets were positioned in a straight line and were elevated 2 cm above a base. That positioning allowed us to visually and auditorily determine misses (see Figure 1). Holding a stylus, participants moved from right to left if the Journal of Motor Behavior
Online Control in Sequential Rapid Aiming
TABLE 1. Characteristic Symptoms of the Most Affected Upper Limb for People in the Parkinson’s Disease Group
ID
Age (years)
Duration (years)
Hoehn and Yahr
Limb bradykinesia
Limb resting tremor
Limb rigidity
1
75
11.0
2
carbidopa/ levadopa ropinirole venlafaxine clonazepam
moderate
mild
moderate
mild dyskinesia, bilateral involvement moderate on/off fluctuations
2
68
3.5
2
pramipexole selegeline
minimal
none
minimal
bilateral involvement
3
66
2.5
2
none
moderate
moderate
minimal
moderate depression, bilateral involvement
4
63
2.0
1
none
mild
mild
mild
unilateral involvement
5
60
3.5
1
carbidopa/ levadopa
mild
none
mild
unilateral involvement
6
61
3.0
1
ropinirole
moderate
minimal
mild
unilateral involvement
7
67
18.0
2
carbidopa/ levadopa pramipexole amantadine
mild
none
mild
moderate dyskinesia, bilateral involvement
8
66
20.0
3
carbidopa/ levadopa ropinirole paroxetine clozapine
moderate
none
moderate
mild dyskinesia, bilateral involvement, mild cognitive impairment
Medications
Other comments
Note. The Hoehn and Yahr (1987) scale ranges from 1 (mildly impaired) to 5 (severely impaired).
TABLE 2. Average Performance on Emotional, Cognitive, and Motor Function Tests PD patients Measure Education (years) Beck Depression Scalea Mini-Mentalb Purdue pegboardc Reciprocal tappingd
Older adults
Young adults
M
SD
M
SD
M
SD
16.4 7.3 28.0 10.7* 30.1*
4.2 7.3 2.1 2.3 6.2
16.9 2.4 29.5 14.5 38.0
2.60 4.80 0.54 1.30 7.70
16.10 4.38 30.00 15.20 38.50
2.2 3.3 0.0 1.1 5.1
Note. Participants in the Parkinson’s disease (PD) group used their most affected hand for Purdue pegboard and reciprocal tap, and people in the control group’s hand use was matched; that is if the nondominant hand was the most affected, then the matched person in the control group used the nondominant hand. a The range of values for the Beck Depression Scale is 0–63; 5–9 is considered normal. bThe range for the Mini-Mental State Examination is 0–30; 27–30 is considered normal. cThe unit of measurement for the Purdue pegboard test is number of pegs, and dnumber of taps is the unit of measurement for the reciprocal tapping test. * p < .01.
March 2007, Vol. 39, No. 2
105
A. L. Smiley-Oyen, K. A. Lowry, & J. P. Kerr
FIGURE 1. Experimental apparatus. Pictured is the seventarget sequence. When using the right hand, participants moved the stylus from right to left. A mirror image of the pattern was provided for those who used their left hand. The x, y, and z coordinates of stylus tip were recorded.
right hand was used and from left to right if the left hand was used. Patients used their most affected hand; we matched OA and YA participants so that the same number of people used either their dominant or nondominant hand. We used an infrared light-emitting diode (IRED), placed 2.5 cm above the tip of the stylus, to record x, y, and z coordinates of the movement with the Optotrak camera system (Model 3020, Northern Digital, Inc., Waterloo, Ontario, Canada) at a spatial resolution of 1 mm. We collected data at 200 Hz. Conditions (Cond) varied in number of targets in the movement sequence, as follows: Cond1 (one-target discrete aim), Cond3 (three-target sequence), Cond5 (five-target sequence), and Cond7 (seven-target sequence). The targets were physically removed when not being used. All targets were 3 cm in diameter and 5 cm apart from center to center, including the distance from the start position to Target 1 (index of difficulty from one target to the next = 1.83, based on log2 [2 amplitude/width]); Fitts, 1954). Participants were to move the stylus as quickly as possible from target to target while attempting to maintain a 90% accuracy rate. Procedure While holding the stylus (keeping their hand and arm off the surface), participants contacted the surface of each target with the stylus tip. After a warning tone and go signal (we used durations between tones that randomly varied between 800 and 1,200 ms to reduce anticipation), they were to execute the sequence as quickly and as accurately as possible and to remain in contact with the last target until a tone sounded indicating the end of the trial. (We required contact with the last target to control the end of the movement.) All trials were completed in less than 3 s. After completion of the trial, we verbally reported to participants their response time 106
(RT + movement time), and, for purposes of motivation, we told them whether that trial was faster than any of their previous trials. If a participant had more than one spatial error after completing 10 trials, then we instructed them to make fewer errors, even if it meant moving more slowly. If there were no errors, then we told participants to move more quickly. When RT was less than 100 ms (indicative of anticipation) or greater than 750 ms (indicative of being distracted) or if they made a spatial error, we deleted and replaced that trial. Training occurred before testing. Participants repeated each sequence approximately 10 times. During those practice trials, we encouraged participants to move more quickly until they missed a target, and we then asked them to slow down slightly until they could consistently contact the targets. We followed that procedure to help all participants identify the speed–accuracy tradeoff that resulted in their fastest time but also to assure that they maintained a 90% accuracy rate. After practice, they executed 20 error-free trials of that condition. There were three different orders in which the conditions were presented: (a) Cond3, Cond7, Cond1, and Cond5; (b) Cond5, Cond3, Cond1, and Cond7; and (c) Cond7, Cond5, Cond3, and Cond1. We randomly assigned people in the PD group to each of the conditions, and we assigned the controls on the basis of the placement of their PD counterparts. All participants completed all conditions. Data Reduction and Analysis We filtered the data with a second-order dual-pass Butterworth filter by using a 21-Hz low-pass cutoff frequency (Smith, 1989). Because abrupt movements tend to consist of higher frequency components, which are important for accurate analysis, we selected that cutoff frequency to optimally remove noise without eliminating important signal information (Winter, 1990). The onset of movement was defined as the first point of five consecutive samples in which the magnitude of tangential velocity increased. We determined the magnitude of tangential velocity (referred to from this point on as velocity) on the basis of the absolute change in the position data over time. We separated the movements to each target into flight times (from the onset of movement or from release of contact with a target until contact with the next target) and contact times.1 The end of movement was at contact with the last target. We stored the trajectory of the IRED for each trial and replayed it in a three-dimensional visual environment by using customdesigned software to check for correct identification of the start and end of the movement and to separate the movement into segments. If any of the algorithms incorrectly identified the beginning, the end, or the parsing of the movement into segments, then the experimenter corrected the incorrect event on the basis of visual replay of the trial and examination of the accompanying position, velocity, acceleration, or jerk profiles, alone or in combination. Dependent Variables We used RT (time from the go signal to the onset of Journal of Motor Behavior
Online Control in Sequential Rapid Aiming
movement) to measure time in preprogramming. We separated movement time (from onset of movement to contact on the last target) into the flight time (FT) for each segment. We identified peak velocity for each segment, and we separated the FT into the acceleratory phase, measured by time to peak velocity (TTPV), and the deceleratory phase, measured by time after peak velocity (TAPV). We calculated a symmetry ratio (TTPV/TAPV) to provide the relative contribution of each phase to the FT. To assess adjustment of the movement before target contact, we counted the number of error-corrective movements, which we defined as the number of times after peak deceleration that the acceleration curve crossed the zero baseline and remained above the baseline for at least 20 ms (four consecutive samples). In addition we assessed coefficients of variation of peak velocity, TTPV, and TAPV. We used the coefficient of variation (calculated as standard deviation/mean * 100) rather than standard deviation to account for mean differences between groups, thus controlling for the inherently greater variability in higher means. We reduced data by calculating a mean across trials for each participant. The goal accuracy rate was 90%, and the PD, OA, and YA accuracy rates, respectively, for Cond1, Cond3, Cond5, and Cond7 were 94%, 98%, and 98%; 94%, 94%, and 94%; 93%, 91%, and 91%; and 92%, 88%, and 87%. Analysis of the error rates indicated a significant increase in errors with number of targets in the sequence (p < .002) but no group difference and no interaction. Other errors such as anticipation or not attending to the start of the movement were approximately 3% for all the groups and conditions (PD = 6%; OA < 1%; YA = 2%). We conducted Group × Segment analyses of variance (ANOVAs) with repeated measures on the second factorto examine changes in performance within a condition across segments. We used Group × Condition ANOVAs with repeated measures on the second factor to analyze RT and kinematics in Segment 1 so that we could assess changes in planning early in the sequence. Note that for the comparison of the kinematics of Segment 1 across conditions, the overt movement requirements were similar. Thus, biomechanical factors and serial order effects were similar across conditions. (The only exception was that the participants did not move off Target 1 in Cond1.) Thus, changes in kinematics in Segment 1 reflected the degree to which the upcoming targets influenced movement to the first target. For some variables, the distribution was skewed, and we therefore used a log transform to adjust them into ranges below 1.0 to establish a more normal distribution before analysis. In all ANOVAs, if the test for sphericity was not met, then we used the Huynh–Feldt adjustment to determine significance. We conducted Tukey honestly significant difference post hoc tests to determine differences between groups. We express variability around the mean as standard error unless otherwise specified. Alpha level was set at .05. March 2007, Vol. 39, No. 2
Results General Description of Execution As expected, the PD group’s overall movement time increased more than those of the other two groups as targets were added to the sequence. The Group × Condition interaction for movement time was significant, F(3.3, 34.2) = 3.28, p = .029, although the post hoc tests indicated that the PD group was significantly different from only the YA group. The average time (in ms) for PD, OA, and YA groups per condition were, respectively, 213 (21), 159 (8), and 141 (7) for Cond1; 739 (43), 631 (40), and 567 (21) for Cond3; 1,225 (40), 1,221 (72), and 1,077 (35) for Cond5; and 1,875 (64), 1,744 (101), and 1,543 (60) for Cond7. We divided FT into TTPV and TAPV, and, as can be seen in Figure 2, the pattern of times within a condition was similar between groups, with the exception of Cond7. Statistical analyses within each condition revealed that the PD group exhibited longer times in FT, TTPV, and TAPV than did the other two groups, but there were no significant Group × Condition interactions except for a trend in TAPV for Cond7, p < .08. Because we were particularly interested in how the final segments of a long sequence were executed, we analyzed follow-up ANOVAs on TTPV and TAPV of the final three segments of Cond7. There were no significant results for TTPV, but, because TTPV was trending downward, we examined peak velocity to determine if it, too, was trending downward. Analysis of peak velocity revealed a group main effect, F(2, 21) = 4.749, p = .02. Post hoc analysis showed that the PD group exhibited a lower peak velocity than those of the other two groups but no session main effect and no interaction (peak velocity [cm/s] for Segments 5, 6, and 7, respectively, were 48.1 [3], 48.2 [3], and 47.9 [5] for the PD group; 68.6 [3], 65.1 [3], and 65.5 [4] for the OAs; and 63.9 [6], 61.2 [6], and 60.7 [6] for the YAs). Analysis of TAPV revealed a significant interaction, F(4, 42) = 2.716, p = .042. The PD group exhibited an increase, whereas the other two groups exhibited a decrease. RT To determine whether previous results with young adults were replicated in the present data set, we conducted a oneway ANOVA across the conditions on the YA RT data. The ANOVA revealed a significant difference, F(3, 21) = 4.05, p = .02; Cond1 produced the shortest RT and Cond7 the longest, thus replicating previous results. The Group × Condition ANOVA revealed a significant interaction, F(6, 63) = 2.97, p = .013, in which the PD and OA groups exhibited an increase only from Cond1 to Cond3 (Figure 3A). FT 1 Again, to determine whether previous results with young adults were replicated in the present data set, we conducted a one-way ANOVA across the conditions on the YA FT data. A condition effect, F(1.851, 12.957) = 1.579, p < .002, was 107
A. L. Smiley-Oyen, K. A. Lowry, & J. P. Kerr
PD 260
YA
OA
(A) FT
Time (ms)
240 220 200 180 160 140 1 120
Time (ms)
110 100
–
1
2
3
–
1
2
3
4
5
–
1
2
3
4
5
6
7
2
3
–
1
2
3
4
5
–
1
2
3
4
5
6
7
2
3
–
1
2
3
4
5
–
1
2
3
4
5
6
7
(B) TTPV
90 80 70 60 50 40 1 180
–
1
(C) TAPV
Time (ms)
160 140 120 100 80 60 1 Cond1
–
1
Cond3
Cond5
Cond7
Target Position FIGURE 2. (A) Flight time (FT), (B) time to peak velocity (TTPV), and (C) time after peak velocity (TAPV) for each group and segment within each condition. Error bars are standard errors. Note the increase in TAPV for the Parkinson’s disease group at the end of Condition 7 (Cond7 = seven-target sequence).
revealed in the analysis. FT1 for Cond1 was the shortest and FT1 for Cond7 was the longest, thus replicating previous findings. Analysis of the three groups revealed a Group × Condition interaction, F(6, 63) = 3.3, p = .007. FT1 increased in a similar manner in the OA and the YA groups, but the PD group exhibited an increase only from Cond1 to Cond3 (see Figure 3B). 108
Peak Velocity 1 (PV1) Peak velocity to the first target for each group is plotted in Figure 3C. The ANOVA revealed main effects only for group, F(2, 21) = 7.71, p = .003, and condition, F(3, 63) = 9.34, p < .001. Post hoc tests indicated that the PD group’s PV1 (M = 70 cm/s [7 cm/s]) was significantly lower than that of the OA group (106 cm/s [6 cm/s]). The YA group (83 Journal of Motor Behavior
Online Control in Sequential Rapid Aiming
PD
OA
(A) Reaction Time
260
YA
(B) Flight Time
250 240
240 230
220 Time (ms)
Time (ms)
220 210 200
200 180
190 160
180 170
140
160 Cond1 120
Cond3
Cond5
Cond7
Cond1 110
(C) Peak Velocity 1
Cond3
Cond5
Cond7
(D) Time to Peak Velocity 1
105
110
100 95 Time (ms)
Velocity (cm/s)
100 90 80
90 85 80
70
75
60
70 65
50 Cond1
Cond3
Cond5
Cond7
Cond1
Cond3
Cond5
Cond7
FIGURE 3. (A) Reaction time, (B) Flight Time 1, (C) Peak Velocity 1, and (D) time to peak velocity 1 for each group and condition (Cond). Error bars are standard errors. (A) and (B) portray significant interactions; interactions in (C) and (D) were not significant. The significant main effects for peak velocity and time to peak velocity are described in the text.
cm/s [7 cm/s]) did not significantly differ from either group. The ANOVA also revealed that PV1 was higher for Cond1 than for the other three conditions (PV1s for Cond1, Cond3, Cond5, and Cond7 were, respectively, 86 cm/s [7 cm/s], 71 cm/s [5 cm/s], 72 cm/s [7 cm/s], and 69 cm/s [5 cm/s]). Those results indicate that scaling of PV1 occurred only from the single-target condition to the multitarget conditions and that the pattern was similar across groups. March 2007, Vol. 39, No. 2
TTPV1 TTPV to the first target for each group is plotted in Figure 3D. The ANOVA revealed only a condition main effect, F(3, 63) = 11.8, p < .001, with Cond1 shorter than all others, and no other differences (TTPV1s for Cond1, Cond3, Cond5, Cond7 were, respectively, 73 ms [4 ms], 85 ms [6 ms], 83 ms [5 ms], and 86 ms [7 ms]). Thus, participants scaled both PV1 and TTPV1 from a discrete aiming task 109
A. L. Smiley-Oyen, K. A. Lowry, & J. P. Kerr
(Cond1) to sequential aiming, but no further scaling was observed as the sequence increased in number of targets. TAPV1 In contrast to TTPV1, there was a different pattern of responses between groups for TAPV1. The YA and OA groups exhibited an increase as targets were added to the sequence, whereas the PD group did not. The pattern was statistically significant, as evidenced by a Group × Condition interaction, F(6, 63) = 5.19, p < .001 (Figure 4A). To control for overall duration, we normalized TAPV1 by dividing by FT1; that interaction was also significant, F(6, 63) = 3.76, p = .003. Symmetry Ratio 1 To simultaneously examine relative contributions of TTPV1 and TAPV1 to the duration of FT1, we calculated a symmetry ratio by dividing TTPV1 by TAPV1. When the ratio is equal to 1.0, then the velocity profile is a bell curve, with both components contributing equally to FT. If the ratio is greater than 1.0, then the acceleration phase is longer; and if the ratio is less than 1.0, then the deceleration phase predominates. For discrete aiming (Cond1), the YA group exhibited a ratio of nearly 1.0, whereas, as expected, both of the older groups exhibited a ratio less than 1.0. With the addition of targets, the YA group’s ratio shifted to less than 0.8 (Figure 4B). A 3 (group) × 4 (condition) ANOVA with repeated measures on the second variable resulted in a significant interaction, F(6, 63) = 3.314, p = .007. Number of Submovements in Segment 1 The ANOVA revealed a trend for a Group × Condition interaction (p = .092): The YA and OA groups increased the number of submovements as the number of targets increased, whereas the PD group exhibited no increase (see Figure 4C). Because there were many trials in which there were no submovements, we conducted a second analysis to compare the number of trials in which there were submovements. That analysis resulted in a significant Group × Condition interaction, F(6, 63) = 2.51, p = .03. Most interesting, only the YA group exhibited an increase with number of targets in that variable (see Figure 4D). The differential outcome of the OA group between number of submovements and percentage of trials with submovements indicates that the OA group incorporated at least one submovement in many of the Cond1 and Cond3 trials (approximately 75%), whereas the YA group executed more of those trials without any submovements. Variability of PV1 and TTPV1 We analyzed those variables to determine the variability in the initial force impulse. There was a group main effect for the coefficient of variation of PV1, F(2, 21) = 8.44, p = .002—for the PD, OA, and YA, respectively, Ms = 22.66 (3.7), 19.65 (3.2), and 15.69 (1.4). The Tukey post hoc test indicated that the PD group exhibited more variability than 110
the YA group, and there was a trend for the OA group to exhibit more variability than the YA group (p < .07). There was also a significant group main effect for the coefficient of variation of TTPV1, F(2, 21) = 4.74, p = .02—for PD, OA, and YA, respectively, Ms = 25.13 (4), 14.04 (3), and 14.07 (3). The post hoc tests indicated that the PD group was more variable than both of the other groups and that the OA and YA groups were statistically similar. Variability of TAPV1 The ANOVA revealed a trend for a condition main effect, F(3, 63) = 2.4, p = .076—for Cond1, Cond3, Cond5, and Cond7, respectively, Ms = 20.3 (2.2), 12.8 (2.1), 29.3 (4.1), and 19.1 (2.3)—but no group main effect and no interaction. Thus, group differences in variability were evident only in the initial force impulse, not in the honing phase of the movement. Discussion The data supported the position that the PD group used less integrative planning at the initial stages of the movement sequence than did either the age-matched controls or the YA group. However, the PD group did not differ from the other groups in all aspects of control. In fact, in some ways they were similar to even the YAs. As indicated by changes in RT and FT1, the PD and the OA groups similarly increased preprogramming from the discrete aiming condition to the shortest sequence (three targets). In contrast, the YAs increased preprogramming for the five- and seven-target sequences. Previous work indicated that, for young adults, the upper limit for preprogramming is nine targets (Smiley-Oyen & Worringham, 2001). Our results concur with previous findings indicating that people with PD use preplanning for shorter sequences but not for longer sequences (Behrman et al., 2000; Harrington & Haaland, 1991; Rafal et al., 1987; Reed & Franks, 1998; Stelmach et al., 1987). To our knowledge, we are the first researchers to separate the acceleration phase data from the deceleration phase data to examine how each is adjusted as the sequence is increased in number of segments beyond a two-segment sequence. Most interesting, all groups scaled the first force impulse in a similar manner. They increased TTPV and decreased peak velocity in the first segment only from the one-target to the three-target condition. In other words, the first force impulse was adjusted between discrete aiming and sequential aiming, but additional targets in the sequence did not cause further scaling. Our results concur with previous findings that people with PD and controls scale peak velocity and TTPV from discrete aiming to a sequential aiming task of two segments in a similar manner (Reed & Franks, 1998). In contrast, the deceleration phase of the movement to the first target differentiated the PD group from the other two groups. The YA and OA groups increased time decelerating to Target 1 in anticipation of up to four targets, but the PD Journal of Motor Behavior
Online Control in Sequential Rapid Aiming
PD
OA
(A) TAPV1
1.2
YA
(B) Symmetry Ratio 1
160 1.0 TTPV/TAPV
Time (ms)
140
120
100
0.8
0.6 80 60
0.4 Cond1
Cond3 Cond5 Cond7
Cond1 Cond3 Cond5 Cond7 100
(C) Submovements 1
(D) % with submovements
1.2
80
1.0 Percentage
No. of Submovements
90
0.8
0.6
70 60 50 40
0.4 Cond1
Cond3 Cond5 Cond7
Cond1 Cond3 Cond5 Cond7
FIGURE 4. (A) Time After Peak Velocity 1, (B) Symmetry Ratio 1 (time to peak velocity [TTPV]/time after peak velocity [TAPV], (C) number of Submovements 1 (averaged across trials; thus, fractions are plotted), and (D) percentage of trials with submovements to the first target for each group and condition (Cond). Error bars are standard errors. The horizontal line in the symmetry ratio represents a movement in which 50% involved acceleration and 50% involved deceleration.
group did not. The increased deceleration time with the addition of targets reflects the implementation of a plan in which multiple segments are taken into account. The symmetry ratio of the velocity profile to the first target also indicated that the YA and OA groups increased their reliance on the deceleration phase as targets were added to the sequence, but the PD group changed very little. We interpret March 2007, Vol. 39, No. 2
those findings as evidence that the PD group’s plan did not reflect anticipation for the upcoming targets and thus was planned and controlled in a more segmented fashion. The results of previous work indicated that additional targets in a sequence constrain movement early in the sequence, thus making FT1 longer (Fischman & Reeve, 1992; Lajoie & Franks, 1997). Evidence supporting that 111
A. L. Smiley-Oyen, K. A. Lowry, & J. P. Kerr
position is that the flight path to the first target is less variable and that there is less scatter on the first target when there are more targets in the sequence (Short et al., 1996; Sidaway et al., 1995). There is also more central processing with more targets, which may involve increased monitoring of the trajectory (Smiley-Oyen & Worringham, 2001). The results of those studies, together with our findings, indicate that when a movement that incorporates multiple targets is planned, the accuracy demands of the movement to the first target are greater; therefore, based on the speed–accuracy tradeoff, a slower movement is required (Fischman & Reeve; Short et al.; Sidaway, 1991; Sidaway et al., 1988; Sidaway et al., 1995; Smiley-Oyen & Worringham; van Donkelaar & Franks, 1991). However, the constraining of the movement is based only on the degree to which the plan incorporates multiple segments of the movement, thus anticipating the upcoming targets. The PD group did not exhibit that anticipation and thus was executing the sequence in a less integrated and more segmented fashion. An alternative explanation for the apparent absence of anticipatory control is a problem of implementation—that is, the PD group could not execute TAPV any faster—thus masking the time used for the planning and anticipation of the movement sequence. It has long been known that force generation is affected by PD (Hallett & Khoshbin, 1980), with more recent evidence indicating that modulation of the agonist–antagonist muscle groups is also problematic even with medication (Kelly & Bastian, 2005; Robichaud, Pfann, Comella, & Corcos, 2002). Although we did observe an overall slowness in movement in people with PD—thus, slowness of movement cannot be completely ruled out as an explanation—the critical question is whether movement slowness masked the scaling of times. For two reasons, we think that it did not. First, the PD group scaled TTPV1 in a manner similar to that of the neurologically healthy groups, showing that movement slowness did not dominate the first force impulse, which is where it might be expected given the just-noted findings. Second, the PD group did exhibit a faster TAPV in other segments, indicating that they could execute that portion of the movement faster than the speed they exhibited to the first target. Thus, adjustment of times was apparent in other variables and segments, indicating that bradykinesia did not mask all variation. Another possible explanation for group differences is that people with PD may learn this movement task more slowly and therefore may not plan the movement as age-matched controls do. However, previous studies demonstrated that people with PD learn a sequential rapid aiming movement in a manner similar to age-matched controls, in terms of RT and movement time (Behrman et al., 2000; Smiley-Oyen, Worringham, & Cross, 2002). In addition, Klapp (1995) found that simple RT decreased with practice. Therefore, if the control groups were learning the sequence more quickly, then they would have exhibited decreased scaling of RT, not greater scaling. Thus, we do not think differences in 112
learning affected the observed differences in how the sequences were planned. One reason why people with PD may not use a plan that takes into account the overall movement is that the disease creates more inherent variability in force generation. If there is greater variability in force generation, then it is more difficult to accurately implement a plan that involves multiple segments. That explanation is consistent with the tenets of the optimized submovement model. According to that model, the overall speed of a discrete aiming movement is optimized on the basis of the variability of the force impulse (Meyer, Kornblum, Abrams, Wright, & Smith, 1988). In the current study, the PD group did exhibit greater variability than the other two groups in both peak velocity and TTPV. Other studies also indicated that force generation is more variable in people with PD (Reed & Franks, 1998; Vaillancourt, Slifkin, & Newell, 2002; Wierzbicka, Wiegner, Logigian, & Young, 1991) and that individuals may use different movement strategies to reduce that variability (Longstaff et al., 2003). If a movement has greater inherent variability, then it would be less advantageous to create a plan that takes into account many targets. It would be more advantageous to approach a movement sequence in a more segmented fashion, anticipating only one or possibly two targets in advance of the immediate movement. Even in a two-element sequence, there is evidence that the second element exhibits more adjustments, suggesting that the plan has to be updated (Gentilucci & Negrotti, 1999; Reed & Franks). Thus, using a plan that incorporates multiple segments would be less efficient. A final observation is that the PD group showed increased slowness in the later segments of the longest sequence specifically in the deceleration phase. In contrast, the neurologically healthy groups moved faster in the deceleration phase at the end of the seven-target sequence. That result is expected because they have a decreased need for anticipatory control at the end of the sequence. Although many studies have shown that people with PD slow segment times later in a sequence (Agostino et al., 1992; Agostino et al., 1994; Harrington & Haaland, 1991), our data indicated that the slowing is specific to deceleration and not to the initial force impulse or to a combination of both phases. That finding indicates that difficulty with modulation of the agonist–antagonist bursts may increase as they execute a movement sequence. In summary, the results of this study show that both young adults and neurologically healthy older adults adjust movement to the first target in anticipation of up to four targets removed from the immediate movement. In contrast, people with PD do not exhibit that anticipation. They organize a movement in a more segmented fashion, anticipating only up to one or two targets in advance of the immediate movement. We suggest that one factor contributing to the deficiency in integrative planning is greater inherent variability in their force impulse, which makes it more difficult to consistently predict the movement outcome, thus making it less advantaJournal of Motor Behavior
Online Control in Sequential Rapid Aiming
geous to plan too far in advance. Their more segmented planning and control may be the best solution for executing a movement quickly and accurately. However, more segmented planning may contribute to movement slowness and to greater demands for attention. Further work is needed to better understand the degree to which people with PD use anticipatory control in a variety of movement situations. ACKNOWLEDGMENTS National Institutes of Health Grants NS 043733 and NS 36752 supported this research. This research was presented in part at the 2005 conference for the North American Society for the Psychology of Sport and Physical Activity (NASPSPA). We thank Q. Emerson for recruiting and organizing the participants and B. Boden and A. Ibrahim for their help with data processing. We also thank the following physicians for their help with recruitment and clinical assessment: T. Ajax, M. Kitchell, S. Spencer, and L. Struck. We appreciate the helpful comments provided on earlier drafts by J. Bloedel, M. Fischman, J. Gallagher, J. Thomas, and the anonymous reviewers. Finally, we thank all participants in this study. NOTE 1. Pairs of acceleration spikes identified contact with a target. The first acceleration spike was identified as contact. From the second acceleration spike, there was a backward search to the first jerk spike (second derivative of velocity), which was identified as release from contact. REFERENCES Agostino, R., Berardelli, A., Formica, A., Accornero, N., & Manfredi, M. (1992). Sequential arm movements in patients with Parkinson’s disease, Huntington’s disease and dystonia. Brain, 115, 1481–1495. Agostino, R., Berardelli, A., Formica, A., Stocchi, F., Accornero, N., & Manfredi, M. (1994). Analysis of repetitive and nonrepetitive sequential arm movements in patients with Parkinson’s disease. Movement Disorders, 9, 311–314. Beck, A. T., & Steer, R. A. (1987). Manual for the revised Beck Depression Inventory. San Antonio, TX: Psychological Corporation. Behrman, A. L., Cauraugh, J. H., & Light, K. E. (2000). Practice as an intervention to improve speeded motor performance and motor learning in Parkinson’s disease. Journal of the Neurological Sciences, 174, 127–136. Benecke, R., Rothwell, J. C., Dick, J. P. R., Day, B. L., & Marsden, C. D. (1987). Disturbance of sequential movements in patients with Parkinson’s disease. Brain, 110, 361–379. Brown, S. H. (1996). Control of simple arm movements in the elderly. In A.-M. Ferrandez & N. Teasdale (Eds.), Changes in sensory motor behavior in aging (pp. 27–52). Amsterdam: Elsevier. Camicioli, R., Grossmann, S. J., Spencer, P. S., Hudnell, K., & Anger, W. K. (2001). Discriminating mild parkinsonism: Methods for epidemiological research. Movement Disorders, 16, 33–40. Chamberlin, C. J., & Magill, R. A. (1989). Preparation and control of rapid, multisegmented responses in simple and choice environments. Research Quarterly for Exercise and Sport, 60, 256–267. Christina, R. W. (1992). The 1991 C. H. McCloy Research Lecture: Unraveling the mystery of the response complexity effect in skilled movements. Research Quarterly for Exercise and Sport, 63, 218–230. Fischman, M. G. (1984). Programming time as a function of numMarch 2007, Vol. 39, No. 2
ber of movement parts and changes in movement direction. Journal of Motor Behavior, 16, 405–423. Fischman, M. G., & Reeve, T. G. (1992). Slower movement times may not necessarily imply online programming. Journal of Human Movement Studies, 22, 131–144. Fitts, P. M. (1954). The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47, 381–391. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). MiniMental State: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. Freund, H. J. (1989). Motor dysfunctions in Parkinson’s disease and premotor lesions. European Neurology, 29, 133–137. Garcia-Colera, A., & Semjen, A. (1987). The organization of rapid movement sequences as a function of sequence length. Acta Psychologica, 66, 237–250. Gentilucci, M., & Negrotti, A. (1999). The control of an action in Parkinson’s disease. Experimental Brain Research, 129, 269–277. Gordon, P. C., & Meyer, D. E. (1987). Control of serial order in rapidly spoken syllable sequences. Journal of Memory and Language, 26, 300–321. Hallett, M., & Khoshbin, S. (1980). A physiological mechanism of bradykinesia. Brain, 103, 301–314. Harrington, D. L., & Haaland, K. Y. (1991). Sequencing in Parkinson’s disease. Abnormalities in programming and controlling movement. Brain, 114(Pt. 1A), 99–115. Harrington, D. L., Haaland, K. Y., & Hermanowicz, N. (1998). Temporal processing in the basal ganglia. Neuropsychology, 12, 3–12. Henry, F. M., & Rogers, D. E. (1960). Increased response latency for complicated movements and a “memory drum” theory of neuromotor reaction. Research Quarterly, 31, 448–458. Hoehn, M., & Yahr, M. (1987). Parkinsonism: Onset, progression and mortality. Neurology, 17, 427–442. Kelly, V. E., & Bastian, A. J. (2005). Antiparkinson medications improve agonist activation but not antagonist inhibition during sequential reaching movements. Movement Disorders, 20, 694–704. Klapp, S. T. (1995). Motor response programming during simple and choice-reaction time: The role of practice. Journal of Experimental Psychology: Human Perception and Performance, 21, 1015–1027. Lajoie, J. M., & Franks, I. M. (1997). Response programming as a function of accuracy and complexity: Evidence from latency and kinematic measures. Human Movement Science, 16, 485–505. Longstaff, M. G., Mahant, P. R., Stacy, M. A., Van Gemmert, A. W. A., Leis, B. C., & Stelmach, G. E. (2003). Discrete and dynamic scaling of the size of continuous graphic movements of parkinsonian patients and elderly controls. Journal of Neurology, Neurosurgery & Psychiatry, 74, 299–304. Longstreth, W. T., Nelson, L., Linde, M., & Munoz, D. (1992). Utility of the sickness impact profile in Parkinson’s disease. Journal of Geriatric Psychiatry and Neurology, 5, 142–148. Meyer, D. E., Kornblum, S., Abrams, R. A., Wright, C. E., & Smith, J. E. K. (1988). Optimality in human motor-performance: Ideal control of rapid aimed movements. Psychological Review, 95, 340–370. Phillips, J. G., Chiu, E., Bradshaw, J. L., & Iansek, R. (1995). Impaired movement sequencing in patients with Huntington’s disease: A kinematic analysis. Neuropsychologia, 33, 365–369. Rafal, R. D., Inhoff, A. W., Friedman, J. H., & Berstein, E. (1987). Programming and execution of sequential movements in Parkinson’s disease. Journal of Neurology, Neurosurgery & Psychiatry, 50, 1267–1273. 113
A. L. Smiley-Oyen, K. A. Lowry, & J. P. Kerr Rand, M. K., Alberts, J. L., Stelmach, G. E., & Bloedel, J. R. (1997). The influence of movement segment difficulty on movements with two-stroke sequence. Experimental Brain Research, 115, 137–146. Rand, M. K., & Stelmach, G. E. (2000). Segment interdependency and difficulty in two-stroke sequences. Experimental Brain Research, 134, 228–236. Rand, M. K., Van Gemmert, A. W., & Stelmach, G. E. (2002). Segment difficulty in two-stroke movements in patients with Parkinson’s disease. Experimental Brain Research, 143, 383–393. Reed, C. L., & Franks, I. M. (1998). Evidence for movement preprogramming and on-line control in differentially impaired patients with Parkinson’s disease. Cognitive Neuropsychology, 15, 723–745. Robichaud, J. A., Pfann, K. D., Comella, C. L., & Corcos, D. M. (2002). Effect of medication on EMG patterns in individuals with Parkinson’s disease. Movement Disorders, 17, 950–960. Rocca, W. A., Maraganore, D. M., McDonnell, S. K., & Schaid, D. J. (1998). Validation of a telephone questionnaire for Parkinson’s disease. Journal of Clinical Epidemiology, 51, 517–523. Short, M. W., Fischman, M. G., & Wang, Y. T. (1996). Cinematographical analysis of movement pathway constraints in rapid target-striking tasks. Journal of Motor Behavior, 28, 157–163. Sidaway, B. (1991). Motor programming as a function of constraints on movement initiation. Journal of Motor Behavior, 23, 120–130. Sidaway, B., Christina, R. W., & Shea, J. B. (1988). A movement constraint interpretation of the response complexity effect on programming time. In A. Colley & J. Beech (Eds.), Cognition and action in skilled behaviour (pp. 87–102). Amsterdam: North-Holland. Sidaway, B., Sekiya, H., & Fairweather, M. (1995). Movement variability as a function of accuracy demand in programmed serial aiming responses. Journal of Motor Behavior, 27, 67–76. Smiley-Oyen, A. L., & Worringham, C. J. (1996). Distribution of programming in a rapid aimed sequential movement. Quarterly Journal of Experimental Psychology, 49A, 379–397. Smiley-Oyen, A. L., & Worringham, C. J. (2001). Peripheral constraint versus on-line programming in rapid aimed sequential movements. Acta Psychologica, 108, 219–245. Smiley-Oyen, A. L., Worringham, C. J., & Cross, C. L. (2002).
114
Practice effects in three-dimensional sequential rapid aiming in Parkinson’s disease. Movement Disorders, 17, 1196–1204. Smith, G. (1989). Padding point extrapolation techniques for Butterworth digital filter. Journal of Biomechanics, 22, 967–971. Soliveri, P., Brown, R. G., Jahanshahi, M., & Marsden, C. D. (1992). Effect of practice on performance of a skilled motor task in patients with Parkinson’s disease. Journal of Neurology Neurosurgery, & Psychiatry, 55, 454–460. Stelmach, G. E., Worringham, C. J., & Strand, E. A. (1987). The programming and execution of movement sequences in Parkinson’s disease. International Journal of Neuroscience, 36, 55–65. Sternberg, S., Monsell, S., Knoll, R. L., & Wright, C. E. (1978). The latency and duration of rapid movement sequences: Comparison of speech and typewriting. In G. E.Stelmach (Ed.), Information processing in motor control and learning (pp. 117–152). New York: Academic Press. Teulings, H. L., Contreras-Vidal, J. L., Stelmach, G. E., & Adler, C. H. (1997). Parkinsonism reduces coordination of fingers, wrist, and arm in fine motor control. Experimental Neurology, 146, 159–170. Vaillancourt, D. E., Slifkin, A. B., & Newell, K. M. (2002). Interdigit individuation and force variability in the precision grip of young, elderly, and Parkinson’s disease participants. Motor Control, 6, 113–128. van Donkelaar, P., & Franks, I. M. (1991). Preprogramming vs online control in simple movement sequences. Acta Psychologica, 77, 1–19. Weiss, P., Stelmach, G. E., & Hefter, H. (1997). Programming of a movement sequence in Parkinson’s disease. Brain, 120, 91–102. Wierzbicka, M. M., Wiegner, A. W., Logigian, E. L., & Young, R. R. (1991). Abnormal most-rapid isometric contractions in patients with Parkinson’s disease. Journal of Neurology, Neurosurgery & Psychiatry, 54, 210–216. Winter, D. (1990). Biomechanics and motor control of human movement. New York: Wiley Interscience. Submitted April 5, 2006 Revised June 8, 2006
Journal of Motor Behavior
