Human Movement Science 20 (2001) 643±674
www.elsevier.com/locate/humov
The one-target advantage: A test of the movement integration hypothesis Werner F. Helsen a,*, Jos. J. Adam b, Digby Elliott c, Martinus J. Buekers a a
c
Department of Kinesiology, Katholieke Universiteit Leuven, Tervuursevest 101, 3001 Heverlee, Leuven, Belgium b Department of Movement Sciences, Maastricht University, P.O. Box 616, 6200 MD Maastricht, Netherlands Department of Kinesiology, McMaster University, Hamilton, Ont. L8S 4K1, Canada
Abstract Two experiments were conducted to compare the temporal structure of single aiming movements to two-component movements involving either a reversal in direction or an extension. For reversal movements, there was no cost associated with the movement time for the ®rst segment of the movement. However, regardless of movement direction, initiation instructions, handedness or eector, two-component extension movements were always associated with a longer movement time for the ®rst movement segment. This disadvantage for extension movements, but not reversal movements, is consistent with the notion that there is interference between the execution of the ®rst movement and implementation of the second movement. By contrast, because the muscular force used to break the ®rst movement is also used to propel the second movement, reversal movements are organised as an integrated unit. Ó 2001 Elsevier Science B.V. All rights reserved. PsycINFO classi®cation: 2300; 2330; 2500; 2520 Keywords: Tapping movements; One-target advantage; Manual asymmetries
*
Corresponding author. Tel.: +32-16-329068; fax: +32-16-329197. E-mail address: werner.helsen@¯ok.kuleuven.ac.be (W.F. Helsen).
0167-9457/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 9 4 5 7 ( 0 1 ) 0 0 0 7 1 - 9
644
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
1. Introduction Discrete aiming and reaching movements are skills pervasive in everyday life: from turning on the light, to opening a door, to picking up a glass of wine. In addition, we often have to make a series of movements that vary in spatial±temporal diculty, such as dialing a telephone, using a calculator or entering a security code in a bank terminal. In this study, we examine the phenomenon that a rapid aimed hand movement is executed faster when it is performed as a single, discrete movement than when it is followed by a second movement. Traditionally, this observation was a key topic in the research on the response-programming stage of rapid movement sequences (e.g., Henry & Rogers, 1960; Klapp & Irwin, 1976; Sternberg, Monsell, Knoll, & Wright, 1978). Speci®cally, these studies examined the eect of movement complexity on reaction time (RT). In a typical study, Henry and Rogers (1960) measured the RT for a lowcomplexity movement (®nger lift), a medium-complexity movement (singleball grasp), and a high-complexity movement (double-ball strike). The eect of movement complexity on RT was straightforward: 159 milliseconds for the ®nger lift, 195 milliseconds for the single-ball grasp, and 208 milliseconds for the double-ball strike. Because the initiation of all three movements was signalled by the same auditory stimulus, the authors attributed the increases in RT to the time required to access and assemble more complex neuromotor coordination patterns from memory. Although the terminology has changed over the years, the idea that simple RT increases as a function of the number of units in the response has received a great deal of empirical support (e.g., Christina, 1992; Klapp, 1995; Sidaway, 1991; Smiley-Oyen & Worringham, 1996). Whereas these studies employed the reaction time technique to examine the planning of movements, other publications have focused on the execution of the actual movements and have used movement time as the main dependent variable. In the typical paradigm ®rst presented by Glencross (1980), participants were instructed either to move as fast as possible to a ®rst target and stop (i.e., the 1-tap condition), or to contact the ®rst target and move on to the second target (i.e., the 2-tap condition). The ``one-target advantage'' (OTA) then refers to the dierence in movement time (to the ®rst target) between 1-tap and 2-tap conditions. According to Adam et al. (2000), this phenomenon is a robust one resulting in, on average, a 19millisecond advantage across studies. Alternatively, this OTA can also be understood as a disadvantage for the 2-tap:extension movement. Typically,
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
645
all these studies (Chamberlin & Magill, 1989; Christina, Fischman, Lambert, & Moore, 1985; Fischman, 1984; Fischman & Lim, 1991; Fischman & Reeve, 1992; Sidaway, 1991) have focussed on fast, ballistic, adduction movements (from right to left) of the preferred hand, characterised by a short movement time (generally 400 milliseconds) and higher indices of diculty (>4 bits). These movements also had dwell times between the two movements in excess of the time associated with a simple reaction (i.e., >220 milliseconds). According to Adam et al.'s account of same-direction movements, this interval is long enough for independent preparation of the second movement. Unfortunately, for the purpose of comparison, these movements were made from ipsilateral to contralateral hemispace and vice versa, whereas in all the other ``one-target'' studies, movements were completed within one and the same hemispace. In single aiming movements, hemispace has been shown to be a powerful determinant of both RT and MT (e.g., Elliott et al., 1993) and thus it 1 The distinction the MIH makes between programming and implementation closely parallels the distinction of Ivry (1986) between program construction, in which the separate response elements are loaded into a response or motor buer, and program implementation, which includes retrieval, readout, and execution of response elements.
648
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
Table 1 Mean performance measures (in milliseconds) from the Adam et al. (2000) study examining the movement integration hypothesis of the one-target advantage
Exp. Exp. Exp. Exp. Exp.
a
1 2 3a 4a 5
1-tap
2-tap
1-target advantage
DT
MT2
ID
139 146 116 131 188
157 162 140 149 202 195 492 157 149 146 137 136 126
18 16 24 18 14 7 )4 14 6 3 15 14 4
114 110 96 79 89 93 179 73 63 63 67 72 65
132 144 125 124 178 183 509 143 126 140 125 426 121
2.4 2.4 2.4 2.4 3.7
Exp. 6 Exp. 7
496 143
Exp. 8
122
ext rev ext ext rev±direct rev±new ext rev±loop rev±direct
5.1 2.7 2.4
Values for Experiments 3 and 4 are based on the short movement distance condition.
complicates interpretation of the results accordingly. Consequently, in this study, we adopted the tapping protocol typically used to examine the interdependence between movement elements (e.g., the OTA). To test the predictions of the MIH with respect to the direction of the second movement, in Experiment 1 we used a 2-tap:reversal task and a 2tap:extension task in addition to a 1-tap task. A generic version of the on-line programming hypothesis predicts an OTA for both reversal and extension movements, while the MIH predicts no signi®cant one-target advantage in the reversal condition. First and second movements were symmetric in terms of the distance to be covered, and, by consequence, also in terms of the amount of agonist and antagonist force needed to achieve the target positions. According to the MIH, this should result in optimal integration of both movements. In addition, we manipulated two other variables that have not yet been studied in the literature. To examine the in¯uence of movement direction, we compared abduction to adduction movements. Until now, most studies have used adduction movements (from R to L) in one hemispace, resulting in a typical sequence of ¯exion and extension movements. Recently, Helsen et al. (1998a) have found that the type of the ®rst movement (¯exion or extension) may play a role in aiming performance. Thus, the present study examined whether the one-target advantage depends on ¯exion/extension order. To study the eects of speci®c task instructions as well as the time available for preparation prior to movement onset on the actual movement
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
649
times, we compared an externally (or computer-) paced condition to a condition in which participants initiated their movements when ready. Presumably, in the latter condition, participants could optimally prepare the single movement or the movement sequence prior to movement initiation. From a methodological point of view, we believe this is the ®rst experiment to directly compare externally paced and self-paced initiation procedures. By examining pacing procedure, we hoped to identify both dierences and commonalities between the response planning (e.g., Sidaway, 1991) and response execution (e.g., Adam et al., 2000) protocols used to examine the interdependence of movement elements. Thus, in terms of dependent variables, we examined how the requirements imposed by a second movement in a two-target movement aected not only the movement time, but also the time to initiate the sequence. 2. Experiment 1 2.1. Method 2.1.1. Participants Fifteen male undergraduate students, ranging in age from 18 to 24 years (mean age 19.3 years), volunteered to participate. According to an adaptation of Bryden's handedness questionnaire, they were strongly right handed on all six items of the ®ve-point scale (M 29:0, where 30 is maximal right hand preference) (cf. Bryden, 1977). Based on self-report, all participants had normal or corrected-to-normal vision. Participants were naive with regard to the hypotheses being tested. Informed consent was obtained prior to participation. Because of data collection problems, one participant had to be eliminated from the analysis, leaving 14 participants for the ®nal analysis. Both experiments have been conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki, as approved by the Committee for Ethical Considerations in Human Experimentation of the Faculty of Physical Education and Physiotherapy from Katholieke Universiteit Leuven. 2.1.2. Apparatus The tapping apparatus consisted of a 1 m aluminium cylinder (4 cm diameter) fastened 31 cm above the table surface. Both the home and target
650
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
positions were 1.2 cm circular push buttons mounted on the cylinder surface facing the participant. The distance between adjacent targets was 20 cm (all centre-to-centre). This resulted in an index of diculty of 4.94 bits (Fitts, 1954). All buttons were white and connected to micro-switches requiring a vertical travel distance of 2.2 mm and 390 g to be activated. Using an electronic circuit, the test apparatus was connected to a standard microcomputer that recorded initiation times, movement times, contact times, and errors (i.e., target hits versus misses). Sampling frequency was 1000 Hz. Throughout the experiment, participants wore an Applied Sciences Laboratory headmounted eye tracker fastened on a hockey helmet (750 g) to measure eye and head movements. In this paper, however, we restrict ourselves to describing only the ®nger data. 2.1.3. Task The tapping task was an extension of an aiming task used previously (Helsen, Starkes, & Buekers, 1997; Helsen et al., 1998a; Helsen, Starkes, Elliott, & Ricker, 1998b, ). Participants sat at a table with the right index ®nger on a push button in front of them at eye level (Fig. 1). A red light was
Fig. 1. The test set-up and the experimental conditions (3 Task by 2 Instruction by 2 Movement Direction).
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
651
initiated behind this home button. When the light went o, the participant had either to move to the ®rst target and stop (i.e., the 1-tap condition), to push the ®rst target and move on to the second target (i.e., the 2-tap:extension condition), or to push the ®rst target and move back to the home target (i.e., the 2-tap:reversal condition). Each of the three movements was to be done as fast as possible. 2.1.4. Procedure Participants were seated on an adjustable chair in front of the table on which the aiming apparatus was mounted. They were positioned so that the body midline was aligned with the left target. Each participant ®rst received standardised instructions. The participants were then ®tted with the eyemovement recording apparatus. Participants were instructed to place the index ®nger of the right hand on the home button and to move as fast as possible to press the corresponding target button(s). After each trial, participants were provided feedback about movement time-1 in milliseconds. In the externally paced condition, the red light at the home position was extinguished, following a 2±4-second random foreperiod (i.e., to prevent anticipation). This was the participant's signal to initiate the tapping movement, and to move, as fast as possible, to one of the three target buttons. In the case of the internally paced condition, participants had a 10-second interval following stimulus oset to execute the movement. Participants were instructed to return their right index ®nger to the home position immediately after each trial. Throughout trials, participants were allowed to move the eyes, head and trunk freely. All tapping conditions (1-tap, 2-tap:reversal, and 2-tap:extension) were blocked and conducted during one single test session, with task, instruction, and direction counterbalanced across participants. For each condition, ®ve practice trials were given separated by a 1-minute rest with 20 test trials. Two experimenters were required to conduct this study. One experimenter monitored the actual movement times and provided the participants with feedback regarding movement 1. The second experimenter monitored another set of computer screens for the control of data acquisition and also ensured that the ®nger was aligned on the home position before stimulus oset. 2.1.5. Dependent variables In line with previous work (Adam et al., 2000), the following performance measures were calculated: (1) movement time-1 (MT1) corresponded
652
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
to the duration between the moment of ®nger lift and the depression of the ®rst target button; (2) dwell time or down time (DT) was de®ned as the amount of time the ®nger was in contact with the ®rst target; (3) movement time-2 (MT2) was de®ned as the time between the moment of hand release from the ®rst target button and the depression of the second target button; (4) percentage of errors on the ®rst target. Typically, an error occurred when the participant failed to contact the corresponding target button and, by consequence, failed to activate the corresponding micro-switch. In addition to previous work (Adam et al., 2000), hand initiation time (IT) was also taken into account (5) for the externally paced movements. This parameter was de®ned as the time interval between the stimulus oset and the moment the ®nger was lifted from the home button. All dependent measures were formed from the average of each participant's test trials. Outliers (i.e., those trials on which raw MT1, MT2, DT, or IT scores were 3 SD from their respective block mean) were discarded from analysis. 2.1.6. Design Two separate analyses (initiation time and movement time-1) were conducted using a 3 Task (1-tap, 2-tap:reversal, 2-tap:extension) by 2 Instruction (externally paced, internally paced) by 2 Movement Direction (abduction, adduction) analysis of variance (ANOVA) design with repeated measures on all factors. To further investigate the temporal structure of the 2-tap movements, a down time and movement time-2 analysis were conducted using a 2 Task (2tap:reversal, 2-tap:extension) by 2 Instruction (externally paced, internally paced) by 2 Direction (abduction, adduction) repeated-measures analysis of variance. When necessary, eects involving more than one degree of freedom were adjusted using the Greenhouse±Geisser corrected critical F -values. Alpha levels were set at 0.05 and all signi®cant results involving more than two means were examined using Tukey's HSD procedures. These statistical procedures were also used in Experiment 2. Across conditions, the outliers procedure removed 2.6±12.9% of the data (M 6:9%). Only for the 2-tap:extension task in the externally paced abduction condition (M 12:91%) did the outliers procedure remove more than 10% of the data.
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
653
2.2. Results 2.2.1. Initiation time As might be expected, the hand initiation data analysis showed a signi®cant main eect of Instruction (F
1; 12 82:72; p < 0:0001). According to the instructions, participants took signi®cantly more time to initiate movement in the self-paced condition
M 754 227 milliseconds than in the externally paced condition (M 249 25 milliseconds). Therefore, we removed the main eect of Instruction from the analysis and conducted an initiation time analysis on the externally paced condition using a 3 Task (1-tap, 2-tap:reversal, 2-tap:extension) by 2 Movement Direction (abduction, adduction) ANOVA design with repeated measures on all factors. Unlike previous ®ndings (Henry & Rogers, 1960; Klapp, 1995; Sternberg et al., 1978), this analysis showed no signi®cant eect. Thus, there is no one-target initiation time advantage for any of the 2-tap movements when subjects have to react to an external signal. 2.2.2. Movement time-1 The analysis of MT1 data showed signi®cant main eects of Task (F
2; 24 4:72; p < 0:0186) and Direction (F
1; 13 19:67; p < 0:0008). Speci®cally, movement times-1 were signi®cantly shorter for both 1-tap
M 295 63 milliseconds and 2-tap:reversal
M 297 50 milliseconds movements than for 2-tap:extension movements
M 310 59 milliseconds. As a result, a disadvantage of 15 milliseconds materialised for the 2-tap:extension movement. In addition, movement times-1 were shorter for adduction
M 285 49 milliseconds than for abduction movements
M 316 62 milliseconds. 2.2.3. Down time The analysis of down time data showed signi®cant main eects of Instruction (F
1; 13 7:15; p < 0:0191) and Direction (F
1; 13 6:94; p < 0:0206). Speci®cally, down times were shorter in the self-paced condition
M 124 43 milliseconds than in the externally paced condition
M 140 38 milliseconds. Perhaps more complete preparation, aorded by the self-paced condition, resulted in a decreased necessity for programming speci®c to the second movement during the interval between the two movements. Alternatively, it may be that self-initiation simply allows for movement execution to occur at a time when the system is optimally alert. In terms of direction, the down times for adduction movements
654
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
Fig. 2. Mean movement time-2 (milliseconds) in Experiment 1 as a function of task and movement direction.
M 125 40 milliseconds were shorter than for abduction movements
M 138 42 milliseconds. Given that this eect is independent of task, it may be that movements toward the midline are less dicult than movements away from the midline. 2.2.4. Movement time-2 The analysis of hand MT2 data showed signi®cant main eects of Task (F
1; 13 28:48; p < 0:0001) and Direction (F
1; 13 16:28; p < 0:0014). For the 2-tap:extension movement
M 292 53 milliseconds, MT2 was shorter than for the 2-tap:reversal movement
M 326 78 milliseconds. For abduction
M 313 66 milliseconds, MT2 was shorter than for adduction
M 305 71 milliseconds. The Task Direction interaction, however, was also signi®cant (F
1; 13 45:44; p < 0:0001). As is evident in Fig. 2, the situation involving a second movement away from the midline was the most time consuming. As mentioned earlier, perhaps the midline provides an important anchor for goal-directed aiming. 2.2.5. Percentage errors The analysis of errors on target 1 showed signi®cant main eects of Task (F
2; 28 14:65; p < 0:0001). Speci®cally, the percentage of errors was signi®cantly higher in the 2-tap:reversal condition
M 9:7 5:4% than in
W.F. Helsen et al. / Human Movement Science 20 (2001) 643±674
655
Fig. 3. Error percentage in Experiment 1 as a function of task and movement direction.
both the 1-tap condition
M 4:5 4:3% and 2-tap:extension condition
M 6:5 4:9%. In addition, the Task Direction interaction was signi®cant (F
2; 28 5:51; p < 0:0096). As can be seen from Fig. 3, dierences between the three task conditions were more pronounced for abduction than adduction movements. 2.3. Discussion To date, studies examining the one-target advantage in rapid aimed hand movements have focussed on fast, ballistic, and right-to-left adduction movements (Adam et al., 2000; Chamberlin & Magill, 1989; Christina et al., 1985; Fischman, 1984; Fischman & Lim, 1991; Fischman & Reeve, 1992; Sidaway, 1991). The tasks used have been characterised by a short movement time (generally
