| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Jacobs Center for Lifelong Learning and Institutional Development, Jacobs University Bremen, Germany.
2 Department of Biomedical Engineering and Center for Neurological Restoration, The Cleveland Clinic
3 Center for Functional Electrical Stimulation, Louis Stokes VAMC, Cleveland, Ohio.
Address correspondence to Jay L. Alberts, Department of Biomedical Engineering, Center for Neurological Restoration, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail: albertj{at}ccf.org
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
MANY daily activities such as opening containers or dressing require the precise control and coordination of grasping forces while simultaneously processing external information. The force control impairments in older adults are well documented (e.g., Cole, 1991
; Cole, Rotella, & Harper, 1999; Enoka & Fuglevand, 2001; Galganski, Fuglevand, & Enoka, 1993), as are declines in information processing (Hedden & Gabrieli, 2004). In general, the motor performance and cognitive performance of older adults have largely been studied independent of one another. New insights into potential mechanisms underlying diminished force control in older adults may be gained through the use of dual-task paradigms.
In a dual-task paradigm, participants perform two tasks simultaneously. If a person is able to perform the two tasks simultaneously, with no reduction in performance of either task, then attention is assumed to have been divided successfully (McDowd & Shaw, 2000). Woollacott and Shumway-Cook (2002) defined attention as "the information processing capacity of an individual" (p. 1) and asserted that all tasks require attention. Limits of attention are considered in light of several theoretical positions: For example, the central bottleneck theory states that, as a result of an information-processing bottleneck, only one task can be processed at a time; processing of a second task could not commence until the first is complete. This bottleneck usually results in a longer response time for one of the two tasks within a dual-task paradigm (Pashler, 1994; Pashler, Johnston, & Ruthruff, 2001; Welford, 1967). Attentional resource theory suggests that declines in performance under dual-task conditions result from interference caused by competing demands for attentional resources, which results in less attention available for each task (Kahnemann, 1973). The four-dimensional multiple-resource model (Wickens, 2002) proposes that there will be greater interference between two tasks to the extent that they share stages, sensory modalities, processing codes, and channels of visual information.
Age is thought to be associated with reduced processing efficiency (e.g. decrease in nerve conduction speed; see Hedden & Gabrieli, 2004; Kramer, Hahn, & Gopher, 1999). With advances in age, it is proposed that sensory and motor aspects of performance are increasingly in need of cognitive control and supervision. Proposed mechanisms that necessitate the need for greater cognitive control include sensory losses, impaired sensorimotor performance, and declines in the efficiency of cognitive control processes (Baltes & Lindenberger, 1997; Kramer, Larish, & Strayer, 1995; K. Z. H. Li & Lindenberger, 2002; Wingfield, Tun, & McCoy, 2005). This reduced capacity hypothesis suggests that a given motor task exerts a higher demand on the attentional resources of older than of younger adults if performance level is to be comparable between age groups.
A recent review suggests the importance of assessing motor and cognitive performance in understanding age-related changes in posture and gait (Woollacott & Shumway-Cook, 2002). To our knowledge, few dual-task paradigms have been used to determine how changes in motor and cognitive function may affect an older adult's control of grasping forces (Voelcker-Rehage, Stronge, & Alberts, 2006).
Few attempts have been made to examine the effects of motor practice on dual-task performance in older adults (Baron & Matilla, 1989; Kramer et al., 1995; Kramer, Larish, Weber, & Bardell, 1999; Roenker, Cissell, Ball, Wadley, & Edwards, 2003; Shinar, Tractinsky, & Compton, 2005; Tsang & Shaner, 1998), particularly when an upper extremity task is paired with a cognitive task. McDowd (1986) investigated dual-task performance in a visual tracking task (joystick pursuit tracking; participants were asked to keep the chaser in contact with the target sprite) and an auditory choice reaction time task (participants pressed three buttons corresponding to three auditory tones as rapidly as possible). These data suggest that younger and older adults benefit from practice. However, it is not clear if the age-related differences remain constant after practice or if the age-related decrements in dual-task processing can be reduced.
We investigated the age-related difference in and the practice effects on motor–cognitive dual-task performance by using a fine motor task that required accurate control of digit forces and a working memory task. The precise control and coordination of digit forces is important in the performance of most activities of daily living (Gloss & Warde, 1981). Impairments in the control and coordination of digit forces are likely responsible for the diminished manual dexterity of older adults. Working memory is an important underlying component in the performance of higher level cognitive processes (Dobbs & Rule, 1989; Engle, 2002); a person needs it, for example, to hold a conversation while opening a container, dressing, or eating. Simultaneously examining motor and cognitive performance under conditions that resemble activities performed daily will provide greater insights into how advances in age may be affecting dexterous function.
We hypothesized that age-related declines in attentional resources affect older adults' ability to precisely control grasping forces. Additionally, we addressed the question of adaptations to dual-task requirements by practice. We hypothesized that if the motor task becomes more automatic then that means the task requires less attentional resources (Fitts, 1964), and, as a result of practice, then dual-task performance might improve in younger and older adults. Furthermore, because of a tighter coupling between cognitive–motor performance (caused by the higher demands of motor processes on attentional resources) in older adults, we hypothesized that older adults would make more errors than younger adults in the cognitive task and that errors in the cognitive task (under dual-task conditions) would further disrupt their motor performance.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
We gave all participants a demographics and health questionnaire to determine characteristics of the sample and self-reported health status. All participants were healthy without any neurological disorders and had normal hearing and vision. We also gave participants a battery of three cognitive tests. They performed the Vocabulary subtest of the Shipley Institute of Living Scale (Shipley, 1986), the Digit Symbol Substitution Test, and the Reverse Digit Span subtest of the third edition of the Wechsler Adult Intelligence Scale (Wechsler, 1997). Whereas the Digit Symbol Substitution Test (a measure of perceptual speed) and the Reverse Digit Span subtest (a measure of working memory) represent the fluid dimension of intelligence, the Vocabulary subtest of the Shipley Institute of Living Scale (a measure of acquired knowledge) represents an aspect of crystallized intelligence (e.g., Baltes, Dittmann-Kohli, & Kliegl, 1986; Baltes, Lindenberger, & Staudinger, 1998). The fluid intelligence or cognitive mechanics (Baltes, 1987, Baltes Dittmann-Kohli, & Dixon, 1984) reflects the neurophysiological architecture of the brain and captures, for example, the speed and accuracy of basic information processing. The crystallized intelligence or cognitive pragmatics can be understood as the knowledge-based facet of the mind, such as vocabulary or professional expertise. We observed the typical decrease in fluid intelligence with advancing age in the Digit Symbol Substitution Test and the Reverse Digit Span subtest (see Table 1<--?2-->). In the Vocabulary subtest, an increase in crystallized intelligence was shown; older adults performed significantly better than younger adults. Clinical manual dexterity (Purdue Pegboard Test; Model 32020, Lafayette Instruments, Lafayette, IN) was greater for younger than for older adults.
|
Force modulation task
The target sine wave (6 cycles/30 s) ranged from 5% to 25% of the maximum voluntary force (Voelcker-Rehage & Alberts, 2005). To provide real-time visual feedback, we displayed the target force level, a range of 5% above and below the target force, and actual grip force produced on a 21-in. (53.34 cm) monitor approximately 46 to 61 cm (18–24 in.) directly in front of the participants. We asked participants to match their grip force to the target force line as accurately as possible for each 30-s trial. We tested participants in two sessions with durations of approximately 2.5 hr; they performed a pretest (15 trials to ensure task requirements and 5 test trials), 10 blocks of 10 practice trials (50 trials in each session), and a post-test (5 trials). A prior study indicated that this training protocol is suitable to induce significant motor improvements in the force-tracking performance of young and older adults (Voelcker-Rehage & Alberts).
n-Back Task
The n-back task requires participants to repeat the nth item back (n-back, one-back, two-back) in a sequentially presented list of items (Dobbs & Rule, 1989; S. C. Li & Sikström, 2002). The experimenter manipulates the difficulty of the task by having participants remember items further back in the list. In this study the number of intervening letters varied from zero to two. We presented the letters at a rate of approximately one item every 1.5 s. Experimenter 1 read aloud the letter sets of the n-back task. If the participant made an error, Experimenter 2 said "start over"; then Experimenter 1 presented a new set of letters. At the beginning of each 30-s trial, the experimenter waited 3 s before reading the first letter; this delay allowed the participant to achieve the target force. Participants performed five trials at each difficulty level of the n-back task during pretesting and post-testing, with three practice trials to ensure task understanding.
Dual-Task Conditions
Participants simultaneously performed the motor and n-back task, five trials for each n-back condition in pretest and post-test. The motor task was performed in combination with each of the three n-back conditions (zero-back, one-back, and two-back). An experimenter instructed participants to perform both tasks as accurately as possible and to give equal importance to each task. The experimenter gave participants at least 30 s of rest between each trial.
Data Analysis and Reduction
Force modulation
We analyzed force data from the moment the participant achieved the target force range until completion of the trial. We calculated grip force variability by using the mean standard deviation (SD) of the participants' force, the coefficient of variation, or CV [CV = (SD/M) x 100], and the time within the target range (TWR; this is ±5% of target force). The CV represents the variability in the target force relative to the mean interval of the target force. The TWR indicates the time the participant stayed within the range of 5% above or below the target force. We measured SD, CV, and TWR of the target force every 500 ms within each trial to examine variability within a trial. We aggregated the measures and created a mean SD, CV, and TWR.
n-back task
We used the percentage of correctly repeated letters before the first error (mean of the five trials per n-back difficulty) as an indicator for the working memory capacity of the participants.
Dual-task conditions
In dual-task conditions we measured SD, CV, and TWR as described for the force modulation task. Additionally, if participants made an error in the cognitive task, then we analyzed a time window of 500 ms prior to and 500 ms after the error separately.
Statistical Analysis
We analyzed data by using mixed factor analysis of variance (ANOVA) and corresponding post hoc tests (Bonferroni adjustment). We conducted analyses separately for the motor and cognitive performance. Because of a lack of variance of the zero-back results in the single- and dual-task context, we did not use zero-back results in the ANOVA of the cognitive task. Results indicated no effect of gender on any of the variables for either group, so we collapsed the data across gender.
To determine whether an error in the cognitive task influenced force variability, we compared the mean variability (SD) around the error with the mean variability of that condition. Younger adults did not make a sufficient number of errors in the zero- or one-back dual-task conditions, so we compared performances of younger and older adults by t tests for independent samples. In addition, we calculated the intercorrelation among all pretest and post-test conditions by using bivariate correlations.
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Younger adults displayed a nearly smooth modulation of grip force in single- and dual-task conditions during all n-back difficulty levels. In contrast, the grip force performance of older adults was more variable, particularly as the complexity of the task increased (see Figure 1<--?1-->). Means and standard deviations of the data are presented in Table 2; results of the ANOVAs are presented in Table 3<--?6-->. The 2 (younger, older adults) x 4 (force only, force at zero-back, one-back, two-back difficulty) x 2 (pretest, post-test) ANOVA showed a significant main effect of task difficulty, with force modulation variability increasing with task difficulty. The age and time effect was also significant, as younger adults outperformed older adults and performance improved from pretest to post-test with practice (see Tables 2 and 3 and Figure 2). After practice, older adults' motor performance was similar to the pretest performance of younger adults (see Figure 2, force only). A significant Age x Time interaction indicated a different pretest-to-post-test development for younger and older adults. Time x Task Difficulty and Task Difficulty x Age interactions were also significant. Whereas older adults' force performance decreased from single- to dual-task context, younger adults' performance stayed nearly constant (see Table 3 and Figure 2). The Time x Task Difficulty x Age interaction was not significant (post hoc contrasts showed these results for pretest and post-test). Results for CV and TWR confirmed the results described for SD (see Table 3).
|
|
|
|
Cognitive Task
The 2 (younger, older adults) x 2 (one-back, two-back) x 2 (single-, dual-task) x 2 (pretest, post-test) ANOVA showed a significant main effect of n-back difficulty, F(1, 24) = 87.34, p <.01, 
2 = 0.79, context, F(1, 24) = 9.55, p =.01, 2 = 0.29, age, F(1, 24) = 10.89, p <.01, 2 = 0.31, and time, F(1, 24) = 18.98, p <.01, 2 = 0.44. Participants recalled a lower percentage of letters correctly as the difficulty of the n-back task increased and as the context changed from single- to dual-task condition (cf. Figure 3 and Table 2). Older adults performed generally worse than younger participants. Overall, participants improved performance from pretest to post-test. A significant n-Back Difficulty x Age interaction, F(1, 24) = 4.59, p =.04, 2 = 0.16, indicated that younger adults showed less decrements in n-back performance with increasing n-back difficulty than did older adults. All other interactions were not statistically significant. Post hoc contrasts showed that contrary to pretest, in post-test the context effect was not significant. Hence, in post-test, participants did not recall a lower percentage of letters correctly as the context changed from single- to dual-task.
|
Within-Person Stability and Intraindividual Variability
Next, we examined the correlation among the variability of all test blocks for younger and older adults. Table 4 contains the results of the correlations among the intraindividual variability (SD). Intraindividual variability was strongly and positively correlated over the eight blocks for younger and older adults with one exception. Post-test dual-task performance in the one-back task of the younger adults was unrelated to pretest performance blocks.
|
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Age-Related Differences in Dual-Task Performance
Age differences in dual-task performance were predicted because motor performance may require an increasing amount of cognitive resources, control, and supervision as one ages. According to this prediction, the observed age differences in dual-task performance might primarily reflect processing conflicts at the level of executive control processes or attentional resources (McDowd & Shaw, 2000). As we expected, performing the more difficult cognitive tasks led to greater variability in older adults' motor performance compared with the single-task condition. The performance variability observed in older adults might be a reflection of age-related declines in perceptual speed (or related constructs), as reflected in the Digit Symbol Substitution Test. Older adults showed a significant correlation between pretest and post-test dual-task performance in the two-back task and the performance in the Digit Symbol Substitution Test, rpre(12) =.83, p <.01, and rpost(12) =.69, p =.01. Reduced performances in the Digit Symbol Substitution Test reflect a general slowing of cognitive processes (e.g., Salthouse, 1992a) that might suggest processing conflicts under dual-task conditions. We did not find effects for the other cognitive tasks. In contrast to the compromised motor performance under dual-task performance of older adults, younger adults' motor performance was nearly stable across task complexity levels. Younger adults' performance indicates their ability to successfully divide and allocate attentional resources. Aging seems to be associated with decreases in the ability to manage and coordinate multiple processes, skills, and tasks (Kramer et al., 1995; Salthouse, 1992b). In light of the four-dimensional multiple-resource model (Wickens, 2002), the low interference between motor and cognitive performance in young adults suggests that both tasks share different sensory modalities such as visual, auditory, and tactile, whereas in advanced age, sensory and motor performances require greater cognitive and attentional resources. This may result in higher interference between motor and cognitive resources.
Effects of Motor Practice on Dual-Task Performance
Motor practice improved force modulation in both age groups. Older adults' post-test performance was comparable with the pretest performance of young adults. Thus, older adults are able to improve the modulation of grasping forces after practice. Somewhat unexpectedly, motor practice failed to reduce performance losses under the dual-task conditions (Figure 2). Motor performance decrements from single- to dual-task and age differences in the force modulation task were neither removed nor reduced by intensive motor practice. These data suggest that older adults may not be able to divide their attentional resources successfully even after extensive practice. Frontal regions of the brain are implicated in tasks that require executive functions and task coordination (Cabeza, 2001). These regions are most affected by the normal aging process (Cabeza, 2001; Cabeza et al., 2004). Compromised functioning of these frontal regions most likely underlies the diminished dual-task performance of older adults.
It is important to mention that motor practice did improve force-tracking performance in both groups. As one can see in Figure 2, younger and older adults' post-test performance under dual-task conditions (also at the two-back condition) was as precise as their pretest performance under single-task conditions. Motor practice alone did not improve decoupling of motor and cognitive performance in older adults. The higher absolute post-test performance of younger and older adults, however, points to the importance of well-directed intervention programs to improve manual dexterity of older adults—also in multitask situations. Everyday life, for the most part, consists of settings in which concurrent tasks must be coordinated (Shinar et al., 2005). The tracking of a sinusoidal force wave is not a typical daily activity. However, many daily tasks such as transporting objects throughout the workspace to a specific destination require a modulation of grasping forces (Flanagan & Tresilian, 1994; e.g., picking up a glass of water, transporting it to the mouth, returning it to the table, and releasing hold of the glass).
Motor practice did reduce the decline in cognitive performance under dual-task conditions for older adults, particularly in the one-back condition [Figure 3(b)], suggesting that the benefit of motor training on the reduction of dual-task working memory costs in younger and older adults was actually comparable. Motor practice seems to free up cognitive resources that were previously monitoring motor performance. Older adults seemed to use these resources to improve their cognitive performance under dual-task conditions. Recently, Dault and Frank (2004) examined whether dual-task practice of a postural control and a cognitive task could modify the changes seen in postural sway under dual-task conditions. Practice did not impact motor performance but did improve cognitive performance under dual-task conditions. It is unknown whether continued exposure to the dual-task situation itself (and not motor practice alone) would have produced similar changes in motor or cognitive performance or even higher performance benefits. In this vein, one should consider whether practice on a single task is the best way to improve dual-task performance.
Errors and Motor Variability
Further evidence for limited cognitive–motor resources in older adults and the impact of cognitive resources on force control may be the variability in force tracking at the moment the participant made an error in the cognitive task. In many cases, older adults exhibited a dramatic increase in force variability when they made an error in the cognitive task (e.g., a sudden decrease in grip force; see Figure 1). An error in the cognitive task may require the older adults to "reset" cognitive processing, which consequently requires a resetting of motor commands as a result of this disruption. It appears that younger adults have sufficient attentional resources (or no process channel sharing) to perform both tasks regardless of complexity and error, as they exhibited consistent force profiles, even when making an error on the cognitive task, whereas limited cognitive–motor resources in older adults may necessitate a resetting of cognitive and motor performance. The results suggest the relative differentiation of the motor and cognitive domains in younger adults, and the dedifferentiation of these domains with advancing age (Baltes & Lindenberger, 1997; Lindenberger, Markise, & Baltes, 2000). The present results are consistent with previous studies suggesting that the coupling between sensory and motor performance may be greater in older adults. (K. Z. H. Li, Lindenberger, Freund, & Baltes, 2001; Lindenberger et al.).
Within-Person Stability and Intraindividual Variability
Examination of intraindividual variability has grown in the field of cognitive aging. Intraindividual variability is defined as the variability in a single person on a single task on multiple occasions (Hultsch & MacDonald, 2004; Hultsch, MacDonald, Hunter, Levy-Bencheton, & Strauss, 2000) or on the same occasion over many trials. We found a high consistency of intraindividual variability across all force modulation conditions in younger and older adults; across condition correlations were generally positive and modest to strong. These results are consistent with previous work on intraindividual variability in cognitive tasks: within a particular measure, intraindividual variability tends to be consistent (e.g., Hultsch, MacDonald, & Dixon, 2002; also see Allaire & Marsiske, 2005), indicating a high stability of the motor requirements across conditions. We observed, however, no correlation between younger adults' post-test dual-task performance in the one-back task and their pretest performance. This might indicate a change of performance strategies, as for example discussed by S. C. Li, Aggen, Nesselroade, and Baltes (2001) and Allaire and Marsiske (2005). Although not notable in the absolute performance measures (Table 2 and Figure 2), these results might reflect a practice-related reorganization of dual-task performance from pretest to post-test.
The current data suggest that cognitive performance and force control are interconnected in older adults and that older adults need more attentional resources than younger adults do to perform a dual-task, even after intensive practice. These results underscore the potential benefit of intervention programs to improve manual dexterity in older adults under multitask situations. A more productive intervention strategy may be to develop programs that simultaneously work to improve motor and cognitive performance.
|
Acknowledgments |
|---|
We thank Emiko Suzuki for her assistance in data collection and Erik M. Norrell for his assistance in data analysis.
|
Footnotes |
|---|
Received for publication April 27, 2006. Accepted for publication October 5, 2006.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. L. Alberts, C. Voelcker-Rehage, K. Hallahan, M. Vitek, R. Bamzai, and J. L. Vitek Bilateral subthalamic stimulation impairs cognitive-motor performance in Parkinson's disease patients Brain, December 1, 2008; 131(12): 3348 - 3360. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
