| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
SPECIAL SECTION |
1 Centre for Studies in Aging, Sunnybrook Health Sciences Centre, Toronto, Canada.
2 Institute of Medical Science, 3 Institute of Biomaterials and Biomedical Engineering, and 4 Department of Surgery, University of Toronto, Canada.
5 Department of Kinesiology, University of Waterloo, Canada.
6 Toronto Rehabilitation Institute, Canada.
Address correspondence to Brian E. Maki, PhD, PEng, Centre for Studies in Aging, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Toronto, Ontario, Canada M4N 3M5. E-mail: brian.maki{at}sri.utoronto.ca
 |
A |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. Unpredictable platform perturbations were used to evoke rapid forward stepping reactions in healthy young and older adults. Challenging obstacles and/or step targets increased demands for accurate foot motion in some trials. A concurrent tracking task was performed in half of the trials.
Results. Although participants looked down more frequently in the absence of the tracking task, the ability to clear the obstacle or land on the step target and other spatiotemporal features of the stepping reactions were largely unaffected. There was, however, one notable exception: In older adults, the duration and amplitude of the "anticipatory postural adjustment" that preceded foot lift were reduced in tracking trials, resulting in increased lateral center-of-mass motion.
Conclusion. Impaired attention switching apparently compromised the control of lateral stability during stepping reactions in older adults, and may be an important contributor to increased risk of falling.
Key Words: Aging • Balance • Environmental constraints • Eye movements • Postural control • Stepping • Triggered reactions • Vision • Visual attention
Recent dual-task studies, by the authors, of both young adults (YA) and older adults (OA) (9,10) demonstrated that online visual attention was not required to guide perturbation-evoked stepping reactions, even when obstacles and/or step targets intensified demands for accurate foot movement. Participants performed a concurrent visuomotor (tracking) task that required them to look straight ahead, and they typically continued to look straight ahead during the execution of the step reaction. This finding indicates that the step was guided using visuospatial information that was acquired and stored prior to perturbation onset. An apparent switching of attention (inferred from sudden onset of significant tracking error) occurred after perturbation onset in >95% of trials, but was not related to changes in gaze direction. This "attention switching" almost always occurred prior to foot lift, suggesting that planning and execution of the foot lift and/or step trajectory required significant cognitive involvement. None of these previous studies, however, have examined whether the dual-task competition for cognitive resources affects the capacity to plan, initiate, and execute the step reaction effectively.
The present study aimed to extend the previous work by determining which, if any, features of the perturbation-evoked stepping reactions were affected by the dual-task competition for attention and other cognitive resources. To accomplish this, we compared step reactions evoked while participants did or did not perform the concurrent visuomotor tracking task, within a subset of the previously described cohort of YA and OA (9,10). Gaze direction was also monitored to determine if the competing visuomotor task may have affected the step reactions by inhibiting redirection of gaze toward the floor. We hypothesized that the concurrent task would not affect the step reactions in YA, because of their ability to rapidly and effectively switch attention and reallocate cognitive resources to the task of balance recovery. It appears that OA may have a diminished ability to do this (11), yet may be more dependent on using cognitive resources to control balance (1,2,4,5,12–15). We therefore hypothesized that some features of their step reactions would be affected when OA performed the concurrent task.
|
METHODS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
1 in past year), diabetes, neurological or sensory disorders, recurrent dizziness or feelings of unsteadiness, use of medications that may affect balance, joint replacement, medical conditions interfering significantly with daily activities, or functional limitations of limb use. Participants provided written informed consent to comply with ethics approval granted by the institutional review board.
Protocol
A large (2 m x 2 m) computer-controlled motion platform evoked step reactions via sudden unpredictable horizontal support-surface movement. Participants assumed a standardized comfortable foot position at the center of the platform at the start of each trial. For safety, participants wore a harness suspended from a track mounted on the ceiling. See (9) for details.
Blocks of trials were performed for four environmental-constraint conditions: (i) no constraint, (ii) obstacle only, (iii) target only, and (iv) obstacle plus target (see Figure 1B and caption for details). In 50% of trials, participants also performed a visuomotor "thumb-tracking" task, which involved pursuit tracking of pseudorandom target motion displayed for 20 seconds on a computer screen (3,11); see Figure 1A and caption.
|
|
30 seconds); however, after a trial had started, it was not possible to see the obstacle or step targets (in central or peripheral fields) without redirecting gaze downward, away from the computer screen. To encourage accurate performance on the tracking task, participants were told that the individual with the smallest tracking errors would receive a monetary reward of $50. They were also informed that their tracking scores would be penalized if they failed to clear the obstacle or land on the step target.
Measurements
A bilateral lightweight eye tracker (model 501; Applied Science Labs, Bedford, MA) superimposed the point-of-gaze (accurate to 0.5° of visual angle) on the video images recorded by a head-mounted "scene" camera (sampled at 60 Hz). These images were reviewed to determine the onset of any downward gaze shifts, defined to occur when the point-of-gaze dropped below the bottom edge of the monitor screen and continue to move downward for 50 ms (typical saccade duration) (9). All participants were found to maintain gaze on the monitor during the preperturbation portion of each trial (as instructed); hence, the analysis focused entirely on downward gaze shifts that occurred after perturbation onset.
Two forceplates (AMTI model OR6-5, Watertown, MA; Kistler model 9281, Winterthur, Switzerland) were used to determine step timing and displacement of the center of foot pressure (COP) and center of mass (COM), as detailed previously (16,17). Time of foot-off and foot-contact were determined from the vertical ground-reaction force; step onset was determined from the mediolateral (m-l) COP (excursion > 4 mm). The reaction was deemed to include a preparatory weight-shift phase, commonly referred to as an "anticipatory postural adjustment" (APA), if the initial m-l COP displacement was toward the swing leg, and was characterized by the duration and amplitude of this COP excursion (see schematic inset in Figure 4). This APA propels the COM toward the stance leg, thereby countering the tendency of the body to fall toward the unsupported side during swing phase; hence, the stabilizing effect of the APA was characterized by determining m-l COM displacement at time of foot-off and foot-contact.
|
The "automatic postural response" (APR) that preceded step initiation was quantified by surface electromyographic (EMG) activity recorded bilaterally in medial gastrocnemius. Onset of activation was detected by a computerized algorithm (11) and confirmed by visual inspection. EMG magnitude was quantified by: i) computing the time integral of the rectified signal (first 100 ms of activity); ii) "normalizing" these values by dividing by the participant's average in no-constraint/no-tracking trials; and iii) averaging the left and right normalized values.
Statistical Analysis
For each of the measures described above, a repeated-measures analysis of variance (ANOVA) was performed within each age group to assess the effect of performing the tracking task (tracking vs no-tracking). To determine whether the effect of the tracking task was dependent on the level of demand for accurate control of the foot trajectory, constraint condition (no constraint, obstacle only, target only, obstacle plus target) and Constraint x Tracking interaction were included in the ANOVA, and Tukey post hoc tests were performed. Where necessary, rank transformations were used to normalize the data and/or stabilize the variance, prior to performing the ANOVA (18).
The analysis focused on the forward stepping reactions evoked by the LBTs. Although multiple steps were sometimes required to restore equilibrium (7% of trials in YA, 13% in OA), we focused on the initial step in each trial, as this was the step that had to be modulated to deal with the constraints on foot trajectory. All timing variables were defined relative to onset of platform acceleration (0.1 m/s2), and all spatial variables were expressed as a percentage of participant height. For the frequency analyses, the dependent variable was the percentage of trials (determined for each task condition, within each participant) in which downward gaze shift, successful obstacle clearance, or successful step to target occurred. All other analyses involved the data from individual trials.
|
RESULTS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
1 cm when tracking [0.036 ± 0.037 m vs 0.026 ± 0.034 m; F(1,11) = 14.2, p =.003].
|
Effect of Tracking on Other Features of the Postural Reactions
Table 2 summarizes the results of the statistical analysis of the effect of performing the tracking task on the spatiotemporal features of the perturbation-evoked stepping reactions, as well as the preceding ankle-muscle activation that characterizes the early APR. For most variables, performance of the tracking task resulted either in no statistically significant effect (e.g., step distance) or an effect that was small in magnitude (e.g., 15 ms decrease in foot-off and foot-contact times in OA; 19 ms delay in step onset time in OA; 5 ms delay in step onset in YA).
|
Performing the tracking task led to changes in lateral COM motion, in the OA, consistent with the APA changes. Specifically, the increased APAs in the no-tracking trials acted to propel the COM farther toward the stance-leg side by time of foot-off, which then reduced the degree to which the COM fell laterally during the swing phase (Figure 5, Table 2). As indicated by significant Tracking x Constraint interactions, these effects were largely limited to trials that involved the obstacle.
|
|
DISCUSSION |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
OA also exhibited some small (15–20 ms) but statistically significant changes in step timing when performing the tracking task. Tracking-related reduction in foot-off and foot-contact times was likely caused, in large part, by the tracking-related reduction in the duration of the preceding APA, whereas the delay in step-onset times could possibly reflect dual-task interference with the planning of the step initiation.
The absence of dual-task interference effects, for other features of the postural reactions, implies that either: i) control of the features in question does not require attentional resources, or ii) the switching of attention that occurs after perturbation onset effectively reallocates the resources that are required. Although it is possible that we failed to see such effects because the secondary task was not sufficiently difficult, this seems unlikely. Participants found the tracking task to be very challenging, and required considerable practice to achieve high levels of tracking accuracy. Furthermore, the tracking involves visuospatial processing, and is therefore more likely than nonvisuospatial tasks (e.g., counting, mental arithmetic, verbal fluency) to compete for the cognitive resources required to plan the step. The lack of effect on the APR supports previous reports that this early component of the postural reaction requires little or no cognitive processing, and that cognitive resources are not engaged until later phases of the response (7,11). Consistent with this conclusion, the onset of attention switching in the present cohort [as reported previously (9)] consistently occurred well after APR onset. Conversely, the fact that the attention switching usually occurred prior to foot-off (88% of trials) suggests that cognitive resources were required to plan and execute at least some phases of the stepping reactions, and other dual-task studies have provided further evidence to support this (1,3,4).
Based on the dual-task interference observed in the OA, it appears that effective scaling of the APA is one aspect of the stepping reaction that requires cognitive resources. This conclusion agrees with past work that has attributed anticipatory postural control to supraspinal structures (19). Potentially, the APA disruption observed in the OA, and the absence of such disruption in the YA, can be attributed to age-related impairment in the capacity to rapidly reallocate the required cognitive resources. The similarity, in the YA, in mean onset timing of the APA (210 ms) and attention switching [230 ms (9)] is consistent with this proposition. In a previous dual-task study, OA exhibited a substantial delay in attention switching [120 ms (11)]. The mean delay in the present cohort was smaller [30 ms (7)], and it is not clear whether such a delay could, in itself, cause the impaired APA scaling; however, it is possible that other age-related impairments in ability to reallocate the required cognitive resources may have also contributed. The susceptibility of OA to dual-task interference could thus be caused not only by increased reliance on cognitive control of balance, as suggested previously (12,20,21), but also impaired "attentional dynamics" (5,21).
The effect of the concurrent task on gaze behavior did not appear to be an important factor. Although the tracking task inhibited downward redirection of gaze during the step reaction, participants still stepped without looking down in the majority of trials, and were often able to step over an obstacle and/or land the foot on a target without looking down. Thus, it appears that "stored" visuospatial information was commonly used to guide the stepping reactions, even when participants were free to look down during the reaction, and the increased frequency of downward gaze shifts that occurred in the no-tracking trials provided little benefit (other than slightly improving step-to-target accuracy). The attentional demands of performing the concurrent visuospatial task apparently did not compromise the ability to store, retrieve, and use the salient visuospatial information needed to guide the step, in either age group.
A potential limitation of the study pertains to the relatively small number of OA tested. The tested numbers of participants and trials did, however, provide sufficient statistical power to detect very small task-related differences (e.g., mean differences of 15 ms in foot-off and foot-contact times, in comparing tracking and no-tracking trials; see Table 2). Moreover, the most important new finding of the study, regarding APA scaling, was robust. For example, all six OA exhibited a substantial decrease in mean APA duration in tracking trials, with percent decreases in individual participants ranging from 22% to 45%. The consistency across participants suggests that this impairment may be a common occurrence in senior populations; however, further testing with larger samples is needed to confirm this.
Although the influence of environmental constraints on stepping behavior was not the primary focus of the present study, the findings raise questions about the ability of OA to modulate preparatory weight shifts effectively. Whereas YA increased APA frequency, duration, and amplitude when faced with the challenge of stepping over the obstacle (16,17), such up-regulation was much less evident in the OA, and these deficiencies were exacerbated when the OA were distracted by the concurrent visuomotor task. Moreover, the age-related deficiencies in APA control impacted functional stability (i.e., increased lateral COM motion), and likely also contributed to a diminished ability to land the foot on the target. Previous studies of unobstructed stepping reactions have also demonstrated impaired control of lateral stability in OA (22), and associations with balance impairment (23), fall history (24), and future fall risk (25). APA control is critical when the step must be prolonged to step over objects (16,17); hence, the lateral instability problem is likely to be exacerbated in the "cluttered" environments of daily life. Given the links between lateral instability and risk of hip fracture (26), the current findings may thus have important implications for prevention of such injuries, indicating a need to improve APA control and suggesting that interventions to target "attentional dynamics" may also merit investigation.
|
A |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Dr. Zettel is currently with Human Health and Nutritional Sciences, University of Guelph, Ontario, Canada.
|
F |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Received December 21, 2007
Accepted August 16, 2008
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  N. B. Alexander and J. M. Hausdorff Guest Editorial: Linking Thinking, Walking, and Falling J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2008; 63(12): 1325 - 1328. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
