The aim of this study was to investigate the combined task of walking and prehension with different levels of manual task difficulty in older adults with and without a history of falls.. Ab
Trang 15 N M RINALDIa,b*AND R MORAESb,c
6 a Ribeirao Preto Medical School, University of Sao Paulo, Brazil
7 bResearch Support Center on Chronic-Degenerative
8 Diseases, University of Sao Paulo, Brazil
9 cBiomechanics and Motor Control Lab, School of Physical
10 Education and Sport of Ribeirao Preto, University of Sao Paulo, Brazil
11 Abstract—Older adults have a greater incidence of falls, and
risk of falls will increase when combining two motor tasks.
Thus, it is interesting to investigate the effect of fall history
on motor performance in older adults when combining
walk-ing with another task such as graspwalk-ing an object The aim of
this study was to investigate the combined task of walking
and prehension with different levels of manual task difficulty
in older adults with and without a history of falls Thirty
older adults participated in this study; groups were
desig-nated as fallers ( n = 15) and non-fallers (n = 15)
Partici-pants were asked to reach-to-grasp a dowel during quiet
standing and during walking Level of manual task difficulty
was manipulated by the type of dowel support and obstacles
located at different distances to the sides of the dowel Fall
history influenced the performance of this combined task
for the most difficult manual conditions Fallers were able
to be identified due to differences in the grasping strategies
used while walking compared to non-fallers In addition,
walking and grasping were mutually modulated due to the
level of difficulty of the manual task Ó 2015 Published by
Elsevier Ltd on behalf of IBRO.
Key words: aging, falls, locomotion, prehension, dual task.
12
14 It is estimated that one-third of community-dwelling
15 people aged 65 and older fall every year (O’loughlin
16
et al., 1993; Perracini and Ramos, 2002) Consequently,
17 fall-related injuries are associated with a poorer quality
18
of life due to restricted mobility and functional decline
19 (Tideiksaar, 1996) In addition, one of the major intrinsic
20 risk factors for falls in older adults is deficit in static and
21 dynamic postural control (Verghese et al., 2007)
Impor-22 tantly, more than 50% of falls occur during locomotion
23 (Barak et al., 2006)
24 Older adults with a history of falls (FOA) present some
25 gait impairments (Hausdorff et al., 2001), such as a
26 decrease in stride length and velocity, and an increase in
27 gait variability and double support time (Kirkwood et al.,
28 2011; Toebes et al., 2012) These changes in the walking
29 pattern are even more evident when two motor tasks are
30 combined (Nordin et al., 2010) FOA have a slower swing
31 time and step velocity than older adults without a history of
32 falls in a dual task paradigm (Springer et al., 2006) These
33 results suggest that FOA may have problems switching
34 their attention between two motor tasks due to
neuromus-35 cular problems (Hawkes et al., 2012) The changes in the
36 walking behavior during a dual task paradigm can predict
37 falls in older adults (Beauchet et al., 2009) Moreover,
38 the level of difficulty of the secondary task can also
influ-39 ence how dual-task-related changes are associated with
40
a history of falls (Chu et al., 2013) Nordin et al (2010)
41 investigated, in FOA, gait changes during dual task
condi-42 tions at different levels of difficulty They found that FOA
43 increased their step width in the two most difficult tasks
44 (task 1: carry a saucer with a coffee cup in one hand; task
45 2: perform serial subtractions by three starting from 50)
46 These results indicated the usage of sensory-motor
47 resources in a flexible manner to decrease the risk of falls
48 (i.e., a protective strategy).Hall et al (2011)investigated
49 the impact of cognitive task level of difficulty on walking
50
of FOA FOA reduced gait speed when cognitive task
51 demand increased, suggesting that the more difficult the
52 secondary task is, the greater the impact on gait
perfor-53 mance Furthermore, FOA performed the alphabet and
54 alternate letters tasks more slowly in walking than in the
55 seated condition With an increase in task difficulty, older
56 adults must allocate more attentional resources to walking
57
to compensate for the reduction in sensory-motor control
58 (Stelmach et al., 1990)
59 The combined task of walking and prehension (i.e.,
60 reach-to-grasp) is widely performed during activities of
61 daily life Older adults exhibit smaller peak wrist velocity
62 and greater movement times than young adults when
63 reaching for an object (Roy et al., 1996) Furthermore,
http://dx.doi.org/10.1016/j.neuroscience.2015.12.037
0306-4522/ Ó 2015 Published by Elsevier Ltd on behalf of IBRO.
*Correspondence to: N M Rinaldi, Faculdade de Medicina de
Ribeira˜o Preto, Programa de Po´s-Graduac¸a˜o em Reabilitac¸a˜o e
Desempenho Funcional, Universidade de Sa˜o Paulo, Avenida dos
Bandeirantes, 3900 Ribeira˜o Preto, SP 14049-900, Brazil Tel:
+55-16-3315-0359; fax: +55-16-3315-0551.
E-mail address: narinaldi@yahoo.com.br (N M Rinaldi).
Abbreviations: ANOVAs, analysis of variances; AP, anterior–posterior;
COM, center of mass; FOA, Older adults with a history of falls; HCs,
heel contacts; MDS, margin of dynamic stability; ML, medial–lateral;
MMSE, Mini Mental State Examination; OA, older adults with no history
of falls; SB, stable base; SLD, stable base with obstacles at long; SSD,
stable base with obstacles at short; UB, unstable base; WT, walking
baseline.
Neuroscience xxx (2015) xxx–xxx
1
Trang 264 older adults have reduced tactile sensitivity and,
conse-65 quently, increase the grip force as a compensatory
strat-66 egy (Gorniak et al., 2011) During walking aging has been
67 found to affect prehension.Diermayr et al (2011)
investi-68 gated the aging effects on grasp control when walking and
69 transporting an object They found an increase in grip
70 force while walking, indicating a decline in manual
dexter-71 ity while performing functional tasks
72 Interestingly,Delbaere et al (2004)found that walking
73 and reaching are the most avoided tasks in older adults
74 with fear of falling When reaching for an object in an
75 upright position older adults adopted a hip strategy to
per-76 form the task, which is different than that of young adults
77 who preferred an ankle strategy (Delbaere et al., 2004)
78 Additionally, Huang and Brown (2015) found that older
79 adults showed a larger center of pressure excursions
80 compared to young individuals when combining upright
81 stance with reach-to-grasp These different strategies
82 are likely to compensate for constraints in balance-related
83 functions Thus, it becomes interesting to combine these
84 two tasks because they have the potential to challenge
85 dynamic stability due to mechanical constraints and, at
86 the same time, increase cognitive load because this
com-87 bined task is also a dual task
88 Many studies have investigated the interference of
89 motor/cognitive tasks on walking and the relationship to
90 fall risk in older adults (Menant et al., 2014) However,
91 most of these studies involving dual task paradigms and
92 FOA investigated primarily the main task (i.e., walking)
93 (Beauchet et al., 2009) Recently, we found modifications
94 in walking and prehension when combining these two
95 tasks in young adults (Rinaldi and Moraes, 2015) We
96 suggested that prehension was superimposed on gait,
97 although the adaptations in motor behavior were global
98 because both motor patterns were modified to guarantee
99 the execution of prehension with different levels of
diffi-100 culty while walking without stopping Then, in this context
101 of dual task and falls, it is important to analyze both tasks
102 to investigate the level of interference between these two
103 motor tasks in FOA Possible changes in the prehension
104 control, such as, reduced movement time, wrist velocity
105 and grip aperture velocity could be related to changes in
106 walking control Changes in gait stability could be part of
107 a compensatory strategy to accommodate the control of
108 upper body movements toward an object in FOA
Further-109 more, this combined motor task is different from other
110 dual task paradigms in the literature (Yamada et al.,
111 2011), because most studies have older adults perform
112 the secondary task during the entire pathway (e.g.,
carry-113 ing a tray) Thus, they do not need to change their motor
114 strategy to perform the secondary task, since they could
115 preprogram their movement from the beginning of the
116 walking task However, to perform daily life activities,
117 older adults are required to change their walking patterns
118 to accommodate other tasks (e.g., prehension) Based on
119 these assumptions, our combined motor task can
con-120 tribute to investigate the motor strategies used by FOA
121 when they have to disrupt the walking pattern to
superim-122 pose a voluntary, discrete task
123 Based on these considerations, this study presents two
124 main research questions: (1) what are the changes in
125 prehension and walking when these tasks are combined
126
in FOA? (2) Do these changes occur as a function of the
127 manual task difficulty? To answer these questions, we
128 analyzed variables based on whole body center of mass
129 (COM) (including stability measures) and spatiotemporal
130 gait parameters to describe the possible changes in
131 walking of the FOA due to manual task difficulty We
132 analyzed two steps before object grasping to investigate
133 the changes in walking during the approach phase In
134 relation to reach-to-grasp, we analyzed the reaching and
135 grasping components, such as reaching duration and
136 velocity, and hand grip aperture and velocity We also
137 investigated prehension variables in the upright stance to
138 identify changes in reach-to-grasp due to the addition of
139 walking Therefore, the aim of this study was to
140 investigate the combined task of walking and prehension
141 with different levels of manual task difficulty in older
142 adults with and without a history of falls
143 EXPERIMENTAL PROCEDURES
144 Participants
145 Thirty individuals participated in this study They were
146 distributed in two groups (n = 15): older adults with no
147 history of falls (OA) (15 females); older adults who
148 experienced at least one fall in the 12-month period
149 prior to data collection (FOA) (15 females) Participants
150 were screened before starting the experimental task by
151 filling out a clinical questionnaire to check the history of
152 falls, health status, physical activity level (Baecke)
153 (Voorrips et al., 1991), cognitive function (Mini Mental
154 State Examination, MMSE) (Folstein et al., 1975) and
bal-155 ance performance (Mini-BESTest) (Maia et al., 2013)
156 Participants were excluded if they had cognitive
impair-157 ment (<24 points in the MMSE), vestibular dysfunction,
158 and/or if they were unable to walk without assistance
159
We invited participants through local media (newspaper,
160 television and radio) Forty-eight older adults returned to
161 our invitation We did an initial contact by phone and we
162 asked them whether or not they experienced a fall in the
163 last 12 months, after explaining to them that a fall was
164
an event in which they came to the ground or to some
165 lower level unintentionally, regardless of the
conse-166 quences of the fall After this screening, 28 older adults
167 reported a recent history of fall and 20 older adults did
168 not experience a fall in the last 12 months However,
169 regarding the FOA, seven individuals did not attend to
170 the inclusion criteria (visual problem [n = 1] and use of
171 assistive devices [n = 6]) Yet, six individuals refused to
172 participate in the study For the OA, five individuals did
173 not attend to the inclusion criteria (neurological disorders
174 [n = 3] and musculoskeletal problems [n = 2])
175 All participants had normal or corrected-to-normal
176 vision and no neurological/musculoskeletal disorders
177 that would affect task performance The local ethics
178 committee approved all procedures and participants
179 signed a consent form before starting the experiment
180 Experimental protocol
181 For data collection, we used an 8-camera motion analysis
182 system (MX-T40S, Vicon) with a sampling rate of 100 Hz
Trang 3183 Passive reflective markers were placed on participants’
184 skin at predefined landmarks according to the
Plug-in-185 Gait Full Body model (Vicon) and two markers were
186 placed on the index finger and thumb, respectively
187 Moreover, we positioned two video cameras (Bonita,
188 Vicon) in front of and on the left side of the participants
189 We used the video camera images to determine dowel
190 contact (visual inspection), defined as the first contact of
191 the fingers with the dowel
192 Participants performed two experimental tasks:
193 reaching-to-grasping a dowel (a wood cylinder,
194 diameter: 4.5 cm, height: 10 cm, mass: 105 g) during
195 quiet standing (stationary) and during walking We
196 positioned the dowel on the top of a base made of wood
197 and placed over a support located approximately 3 m
198 from the starting position We adjusted the height of the
199 support to height of the participants’ greater trochanter
200 We positioned the support on the right side of the
201 walkway, with a distance corresponding to 50% of the
202 participants’ right arm length This distance was also
203 used for the stationary task and, in this task,
204 participants stood behind the object at a distance of
205 30% of their right upper limb length The base of the
206 dowel could be stable (wide base) and unstable (narrow
207 base) We made the manual task more difficult by
208 placing the dowel between two wooden obstacles such
209 that for each type of support there were three obstacle
210 conditions: no obstacle, short distance, and long
211 distance The short and long distances of the obstacles
212 corresponded to three and five times the right hand
213 thickness, respectively The combination of type of base
214 and obstacle resulted in six grasping conditions: stable
215 (SB) and unstable (UB) bases without obstacles, stable
216 base with obstacles at short (SSD) and long (SLD)
217 distances, and unstable base with obstacles at short
218 (USD) and long (ULD) distances (Fig 1) Participants
219 also performed a baseline walking condition (WT)
220 without grasping (control condition) More details about
221 the experimental procedures are available in our
222 previous study (Rinaldi and Moraes, 2015)
223 We asked participants to reach-to-grasp the dowel as
224 they walked, without stopping, at a self-selected speed
225 After grasping the dowel, participants were instructed to
226 hold it and walk normally until the end of the pathway
227 For the stationary task, we instructed them to stand as
228 quietly as possible In both tasks, participants were not
229 allowed to contact the obstacles and knock down the
230 support Participants performed 21 trials for the walking
231 task, and 18 trials for the stationary task We collected
232 these tasks in separate blocks and counterbalanced
233 them within each group Trials were completely
234 randomized within each block We repeated those trials
235 with errors at the end of each block without participants’
236 awareness
237 Data analysis
238 The tridimensional coordinates of the individual markers
239 were digitally filtered using a 4th-order Butterworth filter
240 with a 6-Hz cut-off frequency For the walking task, all
241 variables were calculated for the step at the period of
242 contact with the dowel (N) and two steps before contact
243 (N-2 and N-1) to verify possible gait adjustments during
244 the approach phase For the WT condition, the dowel
245 was kept on the support to use as a reference to
246 identify the region corresponding to steps N-2, N-1,
247 and N
248 Spatiotemporal gait parameters Step width and step
249 length were calculated as the absolute difference
250 between heel markers of the right and left feet at
251 subsequent heel contacts (HCs) in each direction,
252 medial–lateral (ML) and anterior–posterior (AP),
253 respectively Step duration corresponded to the frame
254 difference between each HC divided by the sampling
255 frequency The division of step length by step duration
256 resulted in step velocity
257 COM variables and margin of dynamic stability (MDS)
258 Nexus software (Vicon) computed the tridimensional
259 COM coordinates based on the tridimensional
260 coordinates of the 39 markers, which defined a
261 15-segment model (Winter, 2005) COM velocity
corre-262 sponded to the first derivative of the COM position
263 (central difference procedure) We identified the minimum
264 COM AP velocity during the approach phase until dowel
265 contact for all walking conditions After that, we computed
266 the percentage of COM AP velocity reduction as the
dif-267 ference between minimum COM AP velocity in the WT
268 condition and in the walking combined with prehension
269 conditions (Fig 3A) We also calculated the temporal
dif-270 ference between the time of dowel contact and minimum
271 COM AP velocity (Fig 3A)
272 For the computation of the MDS, we first calculated
273 the extrapolated COM position (XcoM) (Hof et al.,
274
2005) Based on XcoM, MDS was calculated according
275
toRinaldi and Moraes (2015), where metatarsal and heel
276 markers on both feet were used to define the outside
277 boundaries of the foot (foot edge) in the AP and ML
direc-278 tions We calculated the MDS by the difference between
279 foot edge and XcoM position (AP and ML directions) A
280 positive value for the MDS indicates that XcoM is located
281 before the foot edge and the system is dynamically stable
282 Reaching-to-grasping variables We used the interval
283 between reaching onset and dowel contact to calculate
284 the reach-to-grasp variables The methods used to
285 detect reaching onset for walking and stationary tasks
286 were the same as the ones used byRinaldi and Moraes
287 (2015) The reaching variables analyzed were: movement
288 time (temporal difference between reaching onset and
289 dowel contact), peak wrist velocity (maximum value
290 obtained in the resultant wrist velocity curve), and
time-291 to-peak wrist velocity (time of occurrence of the peak wrist
292 velocity adjusted to movement time, %) The peak wrist
293 velocity was calculated based on the relative position of
294 the right wrist to the right iliac crest (i.e., relative to the
295 person’s body position in space, as used by Carnahan
296
et al (1996)) The grasping variables analyzed were:
297 peak grip aperture (maximum distance between the
mark-298 ers on thumb and index finger), time-to-peak grip aperture
299 (time of occurrence of peak grip aperture adjusted to
300 movement time, %), peak grip aperture velocity
Trang 4(maxi-301 mum value obtained in the resultant velocity curve of the
302 distance thumb-finger, which was determined as the first
303 derivative of the thumb-finger distance), and time-to-peak
304 grip aperture velocity (time of occurrence of peak grip
305 aperture velocity adjusted to movement time, %)
306 We also analyzed the relationship between reaching
307 onset and gait events For that, we calculated the
308 temporal difference between the last right HC before
309 dowel contact and reaching onset Negative values
310 indicated that the participant touched the ground before
311 starting the reaching movement
312 Statistical analysis
313 One-way analysis of variances (ANOVAs) were
314 computed to compare age, anthropometric (height and
Fig 1 (A) Transverse view of the experimental set-up for the walking and stationary tasks It shows the three steps (N-2, N-1, and N) selected for data analysis and dowel’s location for walking and stationary tasks (B) Illustration of the six prehension conditions Dowel is shown in white, obstacles in black and base in gray colors ( Rinaldi and Moraes, 2015 ).
Trang 5315 body mass) and clinical characteristics (MMSE,
Mini-316 BESTest, and Baecke scores) between groups For the
317 remaining data, five MANOVAs and three ANOVAs
318 were performed Three-way MANOVAs (group [FOA,
319 OA] conditions [SB, SLD, SSD, UB, ULD, USD,
320 WT] step [N-2, N-1, N]) with repeated measures in the
321 last two factors were carried out for the following set of
322 dependent variables: (1) step length and step width; (2)
323 step duration and step velocity; and (3) MDS in the AP
324 and ML directions Two-way ANOVAs (group
325 conditions) with repeated measures for the last factor
326 were carried out for the following dependent variables:
327 (1) temporal difference between HC and reaching onset;
328 (2) temporal difference between dowel contact and
329 minimum COM AP velocity; and (3) AP velocity
330 reduction For reaching and grasping variables, two
331 three-way MANOVAs were calculated (group
332 conditions [SB, SLD, SSD, UB, ULD, USD] task
333 [walking and stationary]) with repeated measures in the
334 last two factors for the following set of dependent
335 variables: (1) movement time, peak wrist velocity, and
336 time-to-peak wrist velocity; and (2) peak grip aperture,
337 time-to-peak grip aperture, peak grip aperture velocity,
338 and time-to-peak grip aperture velocity MANOVAs were
339 followed by univariate analyses, which revealed main
340 and interaction effects; therefore we focused on group
341 and interaction effects Post hoc tests with Bonferroni
342 adjustments were performed for main and interaction
343 effects We computed the effect size using the eta
344 squared (g2) parameter The cut-off criteria for the effect
345 size (partial eta squared [g2]) were: small effect
346 (0.206 g2
< 0.50), medium effect (0.506 g2
< 0.80),
347 and large effect (g2P 0.80) as suggested by Cohen
348 (1992) The level of significance was set at p6 0.05
350 We present results of the MANOVAs and ANOVAs in
351 Tables 1 and 2 Since there was no interaction between
352 step*condition and task*condition for any of the
353 dependent variables, we exclude the statistical results of
354 these interactions from bothTables 1 and 2
355 Sample characteristics
356 ANOVA showed difference between groups for the
Mini-357 BESTest (F1,28= 112.7, p6 0.001) FOA (22.2 pts)
358 scored lower than OA (27.3 pts), indicating impaired
359 balance However, groups were similar in age
360 (FOA = 71.8 years|OA = 70.1 years), height (FOA =
361 1.55 m|OA = 1.54 m), body mass (FOA = 65.7 kg|
362 OA = 59.9 kg), MMSE (FOA = 27 pts|OA = 28.2 pts)
363 and physical activity level (FOA = 4.2 pts|OA =
364 4.1 pts) In addition, FOA presented a mean of 2.1 falls
365 in the last 12 months
366 Spatiotemporal gait parameters
367 Step width was greater for FOA (0.09 m) than for OA
368 (0.06 m) (p = 0.003) (Fig 2A) Step length was greater
369 in WT (0.58 m) than in all grasping conditions (0.47 m)
370 (p = 0.001) (Fig 2B) Relative to the step effect, step
371 width was greater in step N (0.09 m) than N-2 (0.07 m)
372 and N-1 (0.06 m) (p = 0.001) (Fig 2A) However, step
373 length was greater in step N-2 (0.54 m) than N-1
374 (0.48 m) and N (0.46 m) (p6 0.001) (Fig 2B)
375 Step duration was greater for FOA (0.62 s) than for OA
376 (0.52 s) (p = 0.03) (Fig 2C) FOA (0.64 m/s) presented a
377 lower step velocity than OA (0.93 m/s) in step N
378 (p6 0.001) (Fig 2D) For the effect of condition, step
379 duration was greater for the stable base with obstacles
380
at short distance (0.61 s) and the stable base with
381 obstacles at long distance (0.58 s) conditions than for
382 walking through (0.51 s) (p6 0.001) (Fig 2C) However,
383 step velocity was lower in the grasping conditions
384 (0.88 m/s) than in walking through (1.15 m/s) (p6 0.001)
385 (Fig 2D) Step duration was lower in steps N-2 (0.52 s)
386 and N-1 (0.53 s) than in step N (0.66 s) (p = 0.002)
387 (Fig 2C) Step velocity was greater in steps N-2 (1.04 m/
388 s) and N-1 (0.93 m/s) than in step N (0.78 m/s) (Fig 2D)
389 COM variables and MDS
390 For the percentage of COM AP velocity, FOA (60%)
391 presented a greater reduction in AP velocity than OA
392 (30%) (Fig 3B) For the effect of condition, COM AP
393 velocity reduction was greater in the obstacle (stable
394 base with obstacles at short distance: 56.5%; stable
395 base with obstacles at long distance: 45.7%; unstable
396 base with obstacles at short distance: 59.7%; unstable
397 base with obstacles at long distance: 47.9%) than in the
398 no-obstacle conditions (stable base: 28.9%; unstable
399 base: 32.4%) (p6 0.001) (Fig 3B) In absolute values,
400 the mean COM AP velocity for FOA was equal to
401 0.39 m/s and 0.75 m/s for the OA
402 The temporal difference between minimum velocity
403 and dowel contact was greater for FOA (3.51 s) than
404
OA (2.58 s) (Fig 3B) This result showed that FOA
405 exhibited a minimum velocity earlier than OA before
406 dowel contact (Fig 3B) In addition, the minimum
407 velocity occurred earlier in the stable base condition
408 (2.71 s) than in the stable base at short distance
409 condition (3.31 s) (p = 0.001) (Fig 3B)
410
In both directions, MDS was greater for FOA (AP:
411 0.07 m, ML: 0.04 m) than OA (AP: 0.02 m, ML: 0.01 m)
412 (p = 0.002) (Fig 3C) The MDS AP was greater in the
413 stable base with obstacles at short distance (0.07 m)
414 and the unstable base with obstacles at short distance
415 (0.06 m) conditions than in walking through (0.01 m)
416 (p = 0.002) (Fig 3C) In addition, the MDS AP was
417 greater in step N-1 (0.07 m) than in step N (0.03 m)
418 (p = 0.017) (Fig 3C)
419 Reaching-to-grasping variables
420 FOA presented a greater movement time and a lower
421 peak wrist velocity (1.57 s|27.7 m/s) than OA (1.21 s|
422 31.7 m/s) (p = 0.001) (Fig 4A, B) Movement time and
423 peak wrist velocity were greater for stationary (1.62 s|
424 0.46 m/s) than for the walking (1.16 s|0.38 m/s,
425 respectively) task (p6 0.001) (Fig 4A , B) In addition,
426 the time- to- peak wrist velocity occurred earlier in
427 walking (18.3%) than in the stationary (41.2%) task
428 (p6 0.001) (Fig 4C) Regarding the condition effect, the
Trang 6429 movement time was lower for conditions without
430 obstacles (stable base: 1.25 s|unstable base: 1.28 s)
431 than with obstacles (stable base with obstacles at short
432 distance: 1.55-s|unstable base with obstacles at short
433 distance: 1.56 s) (p6 0.001) (Fig 4A) In addition, the
434 movement time was greater for conditions with
435 obstacles at short (stable base with obstacles at short
436 distance: 1.55-s|unstable base with obstacles at short
437 distance: 1.56 s) versus long distances (stable base with
438 obstacles at long distance: 1.36-s|unstable base with
439 obstacles at long distance: 1.36 s) (p6 0.001) (Fig 4A)
440 Peak wrist velocity and time-to-peak wrist velocity were
441 greater in the stable base (0.45 m/s|32%) than in the
442 stable base with obstacles at short distance (0.39 m/s|
443 26%) condition (p = 0.010) (Fig 4B, C) For the
444 unstable base, peak wrist velocity and time- to- peak
445 wrist velocity were lower for the unstable base with
446 obstacles at short distance (0.39 m/s|27%) than for the
447 unstable base with obstacles at long distance (0.43 m/s|
448 33%) condition (p = 0.008) (Fig 4B, C)
449 The temporal difference between right HC and
450 reaching onset was greater for FOA (0.36 s) than for OA
451 ( 0.14 s) This indicated that FOA touched the ground
452 before reaching onset (Fig 4D)
Table 1 F- and p-values for main and interaction effects (condition, step and condition*step interaction) of the MANOVAs and the univariate ANOVAs for the spatiotemporal gait parameters (step length, step width, step duration and step velocity), margin of dynamic stability in AP and ML directions, temporal difference between minimum velocity and dowel contact, and AP velocity reduction (%)
Variables Group Condition Step Condition*group Step*group
MANOVA Wilk’s Lambda = 0.629,
F 2,27 = 7.95, p = 0.002,
g 2
= 0.371
Wilk’s Lambda = 0.121,
F 12,17 = 10.28,
p 6 0.0001,
g 2 = 0.879
Wilk’s Lambda = 0.232,
F 4,25 = 20.66,
p 6 0.0001,
g 2 = 0.768
Wilk’s Lambda = 0.365,
F 12,17 = 2.46,
p = 0.044,
g 2 = 0.635
Wilk’s Lambda = 0.717,
F 4,25 = 2.47,
p = 0.070,
g 2 = 0.283 Follow-up univariate
Step length F1,28= 3.09,
p = 0.09,
g 2 = 0.094
F6,168= 25.63,
p 6 0.0001,
g 2 = 0.478
F2.56= 28.98,
p 6 0.0001,
g 2 = 0.509
F6,636= 1.88,
p = 0.132,
g 2 = 0.063
F2,56= 4.31,
p = 0.018,
g 2 = 0.283 Step width F1,28= 10.80,
p = 0.003,
g 2 = 0.278
F6,168= 3.39,
p = 0.017,
g 2 = 0.108
F2.56= 8.78,
p = 0.002,
g 2 = 0.239
F6,636= 2.46,
p = 0.061,
g 2 = 0.081
F2,56= 0.55,
p = 0.525,
g 2 = 0.133 MANOVA Wilk’s Lambda = 0.67,
F 2,27 = 6.59,
p = 0.005,
g 2
= 0.328
Wilk’s Lambda = 0.120,
F 12,17 = 10.38,
p 6 0.0001,
g 2
= 0.880
Wilk’s Lambda = 0.163,
F 4,25 = 32.16,
p 6 0.0001,
g 2
= 0.837
Wilk’s Lambda = 0.393,
F 12,17 = 2.19,
p = 0.068,
g 2
= 0.325
Wilk’s Lambda = 0.675,
F 4,25 = 3.04,
p = 0.037,
g 2
= 0.607 Follow-up univariate
Step duration F1,28= 10.63,
p = 0.030,
g 2 = 0.275
F6,168= 4.31,
p = 0.007,
g 2 = 0.133
F2,56= 13.73,
p 6 0.0001,
g 2 = 0.329
F6,168= 1.55,
p = 0.206,
g 2 = 0.053
F2.56= 3.78,
p = 0.060,
g 2 = 0.119 Step velocity F 1,28 = 10.62,
p = 0.020,
g 2
= 0.293
F 6,168 = 52.03,
p 6 0.0001,
g 2
= 0.650
F 2,56 = 64.32,
p 6 0.0001,
g 2
= 0.697
F 6,168 = 1.57,
p = 0.193,
g 2
= 0.054
F 2.56 = 5.08,
p = 0.014,
g 2
= 0.154 MANOVA Wilk’s Lambda = 0.528,
F2,27= 12.06,
p 6 0.0001,
g 2 = 0.472
Wilk’s Lambda = 0.338,
F2,27= 12.06,
p 6 0.0001,
g 2 = 0.662
Wilk’s Lambda = 0.451,
F4,25= 7.61,
p 6 0.0001,
g 2 = 0.549
Wilk’s Lambda = 0.604,
F12,17= 0.93,
p = 0.541,
g 2 = 0.596
Wilk’s Lambda = 0.919,
F4,25= 0.55,
p = 0.698,
g 2 = 0.081 Follow-up univariate
MDS AP F 1,28 = 5.85,
p = 0.022,
g 2
= 0.173
F 6,168 = 4.86,
p 6 0.0001,
g 2
= 0.148
F 2,56 = 6.78,
p = 0.002,
g 2
= 0.195
F 6,168 = 0.559,
p = 0.521,
g 2
= 0.020
F 2,56 = 0.546,
p = 0.582,
g 2
= 0.019 MDS ML F 1,28 = 18.53,
p 6 0.0001,
g 2
= 0.398
F 6,168 = 2.51,
p = 0.061,
g 2
= 0.082
F 2,56 = 1.54,
p = 0.227,
g 2
= 0.052
F 6,168 = 1.292,
p = 0.282,
g 2
= 0.044
F 2,56 = 1.443,
p = 0.245,
g 2
= 0.049 ANOVA
Temporal
difference
F1,28= 9.681,
p = 0.004,
g 2 = 0.257
F2,28= 6.391,
p 6 0.0001,
g 2 = 0.186
p = 0.545,
g 2 = 0.027
–
ANOVA
% AP velocity
reduction
F 1.28 = 14.13,
p 6 0.0001,
g 2
= 0.335
F 5.140 = 43.67,
p 6 0.0001,
g 2
= 0.609
p = 0.360,
g 2
= 0.037
–
Trang 7453 Peak grip aperture was unaffected by group, condition
454 and task (Fig 5A) However, time-to-peak grip aperture
455 was lower for FOA (73.2%) than for OA (86.8%) in the
456 walking task (Fig 5B) Time-to-peak grip aperture was
457 lower for conditions without obstacles than for
458 conditions with obstacles at both short and long
459 distances (stable base: 72.2%|stable base with
460 obstacles at long distance: 78.6%|stable base with
461 obstacles at short distance: 78.7%|unstable base:
462 71.9%|unstable base at long distance: 79.9%|unstable
463 base at short distance: 79.1%) (p6 0.001) (Fig 5B) In
464 addition, time-to-peak grip aperture was greater in
465 walking (79.9%) than in stationary (73.6%) task
466 (p = 0.001) (Fig 5B)
467 The peak grip aperture velocity and the time-to-peak
468 grip aperture velocity were lower for FOA (0.12 m/s|
469 37.2%) than for OA (0.15 m/s|47.3%) (p6 0.001)
470 (Fig 5C, D) The peak grip aperture velocity was greater
471 in the stationary (0.18 m/s) than in the walking task
472 (0.09 m/s) (p = 0.01) (Fig 5C) For the condition effect,
473 the time- to- peak grip aperture velocity was lower in
474 conditions without obstacles than in conditions with
475 obstacles at short and long distances (stable base:
476 31.6%|stable base with obstacles at long distance:
477 45.2%|stable base with obstacles at short distance:
478 48.2%|unstable base: 29.6%|unstable base with
479 obstacles at long distance: 47.7%|unstable base with
480 obstacles at short distance: 52.2%) (p6 0.001) (Fig 5D)
481 DISCUSSION
482
We investigated the motor performance of walking
483 combined with prehension at varying levels of manual
484 task difficulty in older adults with and without a history of
485 falls FOA performed worse than OA in the
486 MiniBESTest, which indicates poor balance and
487 mobility This result supports the findings that FOA have
488 poor balance control (Cebolla et al., 2015) History of falls
489 did influence motor performance during the combined
Table 2 F- and p-values for main and interaction effects (condition, step and condition*step interaction) of the MANOVAs and the univariate ANOVAs for reaching (movement time, peak wrist velocity, time-to-peak wrist velocity, temporal difference) and grasping (peak grip aperture, time-to-peak grip aperture, peak grip aperture velocity, time-to-peak grip aperture velocity) variables
Variables Group Condition Task Task*group
MANOVA Wilk’s
Lambda = 0.618,
F 3,25 = 5.15,
p = 0.007,
g 2 = 0.402
Wilk’s Lambda = 0.110,
F 3,25 = 7.04,
p = 0.0001,
g 2 = 0.883
Wilk’s Lambda = 0.121,
F 3,25 = 60.27,
p 6 0.0001,
g 2 = 0.877
Wilk’s Lambda = 0.880,
F 3,25 = 1.13,
p = 0.354,
g 2 = 0.110 Follow-up univariate
Movement time F1,27= 4.37,
p = 0.046,
g 2 = 0.159
F5,135= 21.78,
p 6 0.0001,
g 2 = 0.452
F1,27= 35.27,
p 6 0.0001,
g 2 = 0.573
F1,27= 0.77,
p = 0.389,
g 2 = 0.030 Peak wrist velocity F 1,27 = 15.57,
p = 0.0001,
g 2
= 0.385
F 5,135 = 6.89,
p 6 0.0001,
g 2
= 0.226
F 1,27 = 21.15,
p 6 0.0001,
g 2
= 0.458
F 1,27 = 1.08,
p = 0.307,
g 2
= 0.053 Time-to-peak wrist velocity F 1,27 = 2.68,
p = 0.113,
g 2
= 0.098
F 5,135 = 7.04,
p 6 0.0001,
g 2
= 0.168
F 1,27 = 120.95,
p 6 0.0001,
g 2
= 0.817
F 1,27 = 2.29,
p = 0.142,
g 2
= 0.049 ANOVA
Temporal difference F1,28= 11.98,
p = 0.002,
g 2 = 0.300
F5,140= 1.92,
p = 0.145,
g 2 = 0.064
MANOVA Wilk’s
Lambda = 0.572,
F 4,24 = 4.48,
p = 0.008,
g 2
= 0.428
Wilk’s Lambda = 0.033,
F 20,8 = 11.84,
p = 0.001,
g 2
= 0.976
Wilk’s Lambda = 0.283,
F 4,24 = 15.22,
p 6 0.0001,
g 2
= 0.717
Wilk’s Lambda = 0.658,
F 4,24 = 3.12,
p = 0.034,
g 2
= 0.342 Follow-up univariate
Peak grip aperture F1,27= 0.18,
p = 0.674,
g 2 = 0.007
F5,135= 1.87,
p = 0.167,
g 2 = 0.065
F1,27= 2.58,
p = 0.120,
g 2 = 0.087
F1, 27= 0.11,
p = 0.746,
g 2 = 0.004 Time-to-peak grip aperture F1,27= 12.43,
p = 0.002,
g 2 = 0.315
F5,135= 9.05,
p 6 0.0001,
g 2 = 0.251
F1,27= 14.45,
p = 0.001,
g 2 = 0.349
F1, 27= 5.77,
p = 0.023,
g 2 = 0.176 Peak grip aperture velocity F1,27= 4.55,
p = 0.042,
g 2 = 0.144
F5,135= 2.77,
p = 0.064,
g 2 = 0.093
F1,27= 52.13,
p = 0.001,
g 2 = 0.659
F1, 27= 6.08,
p = 0.020,
g 2 = 0.184 Time-to-peak grip aperture velocity F 1,27 = 8.38,
p = 0.007,
g 2
= 0.237
F 5,135 = 26.56,
p 6 0.0001,
g 2
= 0.496
F 1,27 = 0.16,
p = 0.695,
g 2
= 0.006
F 1, 27 = 3.16,
p = 0.087,
g 2
= 0.105
Trang 8490 task of walking and prehension, and will be discussed
491 below
492 FOA exhibited a more conservative walking strategy
493 and decoupled the combined task when compared to
494 OA
495 Walking performance in FOA was observed to be
496 impaired These impairments included: reduced step
497 velocity and increased step width and duration when
498 compared to OA These findings are in agreement with
499 other studies on gait changes in FOA (Barak et al.,
500 2006; Kirkwood et al., 2011; Toebes et al., 2012) The
501 slowness of FOA may indicate the need for extra time
502 to pick up necessary sensory information to assist with
503 movement performance (Chapman and Hollands, 2007)
504 Although the spatial-temporal gait parameters could allow
505 inferences about the control of stability, the MDS is a
506 more appropriate measurement to investigate dynamic
507 stability control Our results showed an increase in the
508 MDS of FOA to ensure body stability Consequently,
509 FOA preferred a more conservative strategy to increase
510 their dynamic stability compared to OA
511 When walking was combined with prehension, FOA
512 presented a greater reduction in AP velocity than OA
513 Moreover, FOA exhibited the minimum AP velocity
514 earlier than OA before dowel contact Considering the
515 AP velocity, FOA almost stopped walking to perform the
516 prehension task Altogether, these results suggest that
517 FOA decouple the walking and prehension tasks, which
518 may represent a loss of automaticity to superimpose a
519 discrete motor task on walking Similarly to our task, in
520 activities that require movement transitions in sequence,
521 such as sit-to-walk, FOA have been observed to also
522 divide this motor task into two phases (i.e., sit-to-stand
523 and stand-to-walk) to achieve a more upright position
524 before initiating gait to ensure body stability (Chen et al.,
525
2013)
526 FOA began moving their hand toward the dowel
527
360 ms after HC, whereas OA started at 140 ms before
528
HC The behavior of OA was more similar to young
529 adults (Rinaldi and Moraes, 2015), although hand
move-530 ment in young adults started 460 ms before HC The
531 beginning of hand movement 460 ms before HC
repre-532 sents the appropriate timing to exploit the upper limb
for-533 ward momentum (Rinaldi and Moraes, 2015) The short
534 time for OA could be related to a smaller range of motion
535
of their upper limbs For FOA, the beginning of hand
536 movement after HC is one more indication of the
decou-537 pling of walking and prehension FOA may prefer to slow
538 down, almost to the point of terminating gait, before
initi-539 ating hand movement toward the dowel These changes
540 indicate a conservative motor strategy to achieve the goal
541
of the task Older adults with fear of falling present an
542 extended anticipatory postural adjustment (APA) during
543 gait initiation under dual-task condition (Uemura et al.,
544
2012) to compensate for deficits in balance control Thus,
545 FOA have to compensate for balance deficits when
per-546 forming the combined task by dividing it in two phases:
547
a preparatory phase to gain stability followed by the
548 grasping execution
549 History of falls affected prehension movement
550 Changes in motor performance of FOA were not
551 restricted to walking alone, but also occurred in the
552 prehension task FOA presented slower movement time
553 and lower peak wrist velocity, peak grip aperture
554 velocity, and time-to-peak grip aperture To the best of
555 our knowledge, this is the first study to show
556 modifications in prehension movement in FOA, which
557 indicates a generalized slowing down in movement
Fig 2 Mean and standard error for step width (A), step length (B), (C) step duration and (D) step velocity for fallers and non-fallers for all experimental conditions (SB: stable base, SLD: stable base with obstacles at long distance, SSD: stable base with obstacles at short distance, UB: unstable base, ULD: unstable base with obstacles at long distance, USD: unstable base with obstacles at short distance, WT: walking baseline).
Trang 9558 performance This slowness suggests that FOA need
559 more time to gain sensory information to accomplish the
560 manual task successfully, as has been suggested for
561 walking tasks (Chapman and Hollands, 2007) Although
562 most of studies that investigate the performance of FOA
563 showed changes in walking (primary task) (Hall et al.,
564 2011), we also found changes in the secondary motor
565 task (prehension) We suggest that FOA may have
prob-566 lems in switching attention between two motor tasks due
567 to neuromuscular problems, which may be related to a
568 loss of function in the frontal subcortical pathways
569 (Viswanathan and Sudarky, 2011)
570 Additionally, the time-to-peak grip aperture occurred
571 earlier for FOA than for OA only during the walking task,
572 indicating a strategy to increase the time available for
573 online control of hand configuration before dowel
574 contact The need for this additional time in FOA
575 appeared only during the combined task This may have
576 occurred due to the complexity of planning a movement
577 while concurrently performing another movement Thus,
578 they preprogrammed less their grasping movement in
579 the walking task and relied more on online control for
580 performing grasping successfully
581 Gait and prehension were mutually modified due to
582 the level of difficulty of the manual task in both
583 groups
584 Most of the studies have shown a decline in gait
585 performance of older adults during dual task conditions
586 (Hall et al., 2011) However, some studies did not show
587 when older adults changed their walking pattern due to
588 the addition of a secondary task Our results show that
589 older adults made modifications in the walking pattern in
590 the step of object grasping and the previous step before
591
it Similarly,Rinaldi and Moraes (2015)found that young
592 adults made modifications in their walking due to the
addi-593 tion of prehension one-step before it In this context, the
594 level of difficulty of the prehension task influenced the gait
595
of older adults We found a decrease in both step length
596 and speed with the addition of the prehension task Yet,
597 the different levels of difficulty of the manual task affected
598 more some variables; COM AP velocity reduction was
599 more pronounced for the conditions with obstacles,
600 whereas the MDS increased for conditions with obstacles
601
at short distance as compared to WT Thus, participants
602 adjusted their walking to accommodate the accurate
exe-Fig 3 Time series of COM AP velocity for one faller and one non-faller in walking combined to grasping and walking through conditions (A) The vertical dashed lines indicate minimum velocity and dowel contact, respectively The AP velocity reduction (%) is indicated by the difference between the combined condition (thick line) and the baseline walking condition (thin line) The temporal difference is calculated by the difference between the time of minimum velocity and dowel contact, mean and standard error for AP velocity reduction (left side), temporal difference between minimum velocity and dowel contact (right side) (B) and margin of dynamic stability in AP (left side) and ML direction (right side) for experimental conditions (SB: stable base, SLD: stable base with obstacles at long distance, SSD: stable base with obstacles at short distance, UB: unstable base, ULD: unstable base with obstacles at long distance, USD: unstable base with obstacles at short distance, WT: walking through).
Trang 10603 cution of the manual task that may challenge their body
604 stability, as found in young adults (Rinaldi and Moraes,
605 2015) As a result, they adopted a more conservative
606 strategy to allow them to allocate more attention to the
607 grasping task and avoid errors Previous studies also
608 found changes in the walking behavior of older adults
609 for the secondary tasks with the highest level of difficulty
610 (Nordin et al., 2010; Hall et al., 2011) In synthesis,
611 reaching-to-grasping affected postural control by
chang-612 ing the gait behavior
613 Different conditions and tasks influenced prehension
614 Participants presented a greater movement time and a
615 lower peak wrist velocity for conditions with obstacles
616 (short and long distance) Relative to grasping, the
time-Fig 4 Mean and standard error for movement time (A), peak wrist velocity (B), time-to-peak wrist velocity (C) and temporal difference between heel contact and reaching onset (D) for fallers and non-fallers for all experimental conditions (SB: stable base, SLD: stable base with obstacles at long distance, SSD: stable base with obstacles at short distance, UB: unstable base, ULD: unstable base with obstacles at long distance, USD: unstable base with obstacles at short distance).
Fig 5 Mean and standard error for peak grip aperture (A), time-to-peak grip aperture (B), peak grip aperture velocity (C), and time-to-peak grip aperture velocity (D) for fallers and non-fallers for all experimental conditions (SB: stable base, SLD: stable base with obstacles at long distance, SSD: stable base with obstacles at short distance, UB: unstable base, ULD: unstable base with obstacles at long distance, USD: unstable base with obstacles at short distance).