Content uploaded by Leonard L Lapointe
Author content
All content in this area was uploaded by Leonard L Lapointe on Oct 20, 2014
Content may be subject to copyright.
Cognitive-linguistic load
Running head: EFFECTS OF COGNITIVE-LINGUISTIC LOAD ON GAIT IN PD
Effects of Cognitive-Linguistic Load on Parameters of Gait in Parkinson Disease
Julie A.G. Stierwalt, Ph.D. 1,3
Leonard L. LaPointe, Ph.D. 1,3
Charles G. Maitland, M.D. 2,3
Tonya Toole, Ph.D. 3,4
1Department of Communication Disorders, Florida State University
2College of Medicine, Florida State University
3Tallahassee Memorial Healthcare Foundation- Florida State University
Neurolinguistic- Neurocognitive Research Center
4Department of Food, Nutrition, and Exercise Science
Address correspondence to:
Julie A.G. Stierwalt, Ph.D.
Florida State University
107 RRC
127 Honors Way
PO Box 3061200
Tallahassee, FL 32306-1200
1
Cognitive-linguistic load
ABSTRACT
Multitasking has become a way of life, from operating multiple software packages
simultaneously on a computer, to carrying on a conversation on a cell phone while
driving. Perhaps one of the most common dual tasks that we perform is talking while
walking. In isolation, neither task would be considered difficult to perform, yet when
coupled, the relative ease of each task may change. In the current investigation we
expanded the research on talking while walking by manipulating the cognitive-linguistic
complexity of verbal tasks during simultaneous walking in individuals with and without
Parkinson disease. Twenty-five participants with Parkinson disease and thirteen
participants without neurological compromise completed gait tasks while conducting
tasks of low (counting by ones), middle (serial subtraction of threes), and high load
(alpha-numeric sequencing). Our results indicated that cognitive-linguistic demand had
an impact on gait, but not on the manipulated parameters of speech. Those effects were
demonstrated in individuals without neurological compromise as well as those with
Parkinson disease. These results suggest that it might be prudent for health care
professionals and caregivers to alter expectations and monitor the cognitive-linguistic
demands placed on elderly individuals, particularly those with neurological compromise
who might be at greater risk for injurious falls.
2
Cognitive-linguistic load
INTRODUCTION
The effects of interference, competition, and distraction on cognitive-linguistic
processing are unclearly understood, though increasing evidence points to deleterious
effects of competition and distraction on a variety of simultaneous tasks (Tun, O’Kane, &
Wingfield, 2002; LaPointe, Heald, Stierwalt, & Kemker, 2007). Competition across
sensorimotor systems is a particularly intriguing target of research, especially when
simultaneous dual or multi-tasks result in performance reduction of one or more of the
competing tasks.
The ability to conduct two or more tasks simultaneously is a necessity during
activities of daily living. Multitasking has become a way of life, from operating multiple
software packages simultaneously on a computer, to carrying on a conversation on a cell
phone while driving. Perhaps one of the most common dual tasks that we perform,
seemingly without conscious effort, is talking while walking. In isolation, neither task
would be considered difficult to perform for most individuals. Coupling the two
however, may change the relative ease of talking and simultaneous walking. In fact,
although this dual task may appear to be relatively automatic, there are conditions during
which performance may be compromised (Shkuratova, Morris, & Huxham, 2004;
O’Shea, Morris & Iansek, 2002).
The motor system that regulates ambulation or gait appears to be particularly
vulnerable to compromise. Interestingly, even the process of aging has been found to
affect balance strategies and gait patterns when the healthy young are compared to the
healthy elderly (Shkuratova, et al, 2004). The effects of normal aging combined with
neuropathologies, particularly those that result in neurocognitive or neuromotor
3
Cognitive-linguistic load
disturbance may alter our assumption of the perception of relative ease of walking while
talking.
The importance of impaired gait and balance is a compelling focus of research as
injurious falls are the primary cause of emergency room visits in people over the age of
65. Moreover, the cost of injurious falls is devastating, both in terms of financial as well
as personal cost. A milestone study that examined community-dwelling elderly revealed
that 9% of a sample over 4,000 people reported medical conditions related to falls with a
total direct cost of around $8 billion U.S.dollars (Carroll, Slattum, & Cox, 2005). That
total did not include nondirect medical costs, or costs related to the care of injured
individuals. Falls also have been identified as strong predictors for the need for nursing
home care as well as mortality (Tinetti, & Williams, 1997; Wild, Nayak, & Isaacs, 1981).
Injurious falls are a significant health problem, and can only be expected to increase as a
greater proportion of the population is comprised of the elderly. These data highlight the
need for continuing and expanded investigation of potential influences on gait in any
population susceptible to injurious falls.
One investigation of gait that employed a dual task paradigm (cognitive and
motor) in participants with Parkinson disease was conducted by O’Shea, et al., (2002).
They reported that both groups (PD and Controls) demonstrated changes in gait when a
second task was introduced, whether that task was cognitive or motor in nature. As
predicted the individuals with Parkinson disease showed more of a dual task effect than
the group without neurological compromise. Other researchers have explored the dual
task of walking while talking in other clinical populations. Bowen, et al. (2001) found
that performance on a simple verbal task (responding verbally to a specific target
4
Cognitive-linguistic load
embedded in a continuous word stream) while walking significantly affected the balance
and gait velocity of eleven individuals who had suffered a stroke.
These reports reveal deleterious influences on gait for individuals with
neurological compromise. Certain clinical populations may present with a higher risk of
injurious falls because of the nature of their disease. In a prospective study that followed
over 100 individuals with Parkinson disease, Wood, et al., (2002) found that 68.3% of the
sample reported falls. When the same sample was followed for one year, 50% of those
who had reported falls indicated that they had experienced additional falls. These
researchers were not alone in finding a high prevalence of falls for individuals with
Parkinson disease. In another prospective study the prevalence was reported at 51%
(Bloem, Grimbergen, Cramer, Willemsen, & Zwinderman, 2001). The nature of
neuromuscular, balance, and cognitive involvement for those diagnosed with Parkinson
disease clearly leaves them at an increased risk for injurious falls.
In day to day activities, language is often superimposed on other tasks in a dual
task format. Because dual task activities have been shown to influence gait, it would be
of interest to study the influence that language or verbal tasks might impose on gait. The
documented high prevalence of injurious falls in Parkinson disease makes this a
compelling population to study further. The purpose of the current study, therefore, was
to expand the research on talking while walking by manipulating the cognitive-linguistic
complexity of verbal tasks (talking) during simultaneous walking in individuals with and
without Parkinson disease. The examination of gait under varying linguistic/cognitive
loads can provide important information about fall risk, and perhaps offer leads for
intervention especially for those individuals with Parkinson disease, who already exhibit
5
Cognitive-linguistic load
a heightened risk for injurious falls. In addition to studying gait, a secondary purpose
was to examine the speech of individuals who completed dual tasks under varying
cognitive load.
METHODS
Participants
Twenty-five individuals with Parkinson disease consented to participate in this
study. The mean age of participants was 67.44 years (range of 41-91 years) and included
six women and nineteen men. Hoehn and Yahr (H&Y) stages were distributed as
follows: 18 participants were at H&Y stage II, 4 at H&Y stage III, and the remaining 3
were at H&Y stage IV. In order to examine the effects of cognitive load as uniformly as
possible across the sample, participants were asked to complete the tasks at the peak
portion of their medication cycle which was approximately 1½ -2 hours after
administration. Thirteen volunteers without reported history of neurological impairment,
matched in age (mean 68.07 years), and education served as control participants.
Instrumentation
To analyze parameters of gait, the GAITRite© Portable Walkway System was
employed. The GAITRite© System contains six sensor pads encapsulated in a 14 foot
mat with an active grid comprised of 48 sensors by 288 sensors for a total of 13824
sensors. The mat walkway measures, interprets, and records gait data using the
accompanying GAITRite© software. As participants walk down the walkway, the
program measures and records temporal, spatial, and pressure gait characteristics.
Procedures
All procedures were approved by the appropriate Institutional Review Boards
prior to conducting the investigation. Upon completion of informed consent, participants
6
Cognitive-linguistic load
from each group completed eight walks across the GAITRite© Portable Walkway System
under conditions of Baseline (no dual task, thus, conversation without walking, and
walking without talking), low-load (counting forward by one, while walking), mid-load
(serial subtraction by three, while walking), and high-load (alpha-numeric sequencing;
e.g. D-7, E-8). Each task was completed twice and performance was averaged. Finally,
load conditions were counterbalanced to equally distribute potential order effects across
conditions.
In addition to measuring gait characteristics, participants completed the Dementia
Rating Scale – 2 (DRS-2; Jurica, Leitten, & Mattis, 2001), and the Beck Depression
Inventory. These measures were collected in order to identify additional factors that
might influence gait performance and to serve as exclusionary criteria for participants
who might be severely demented. The group with Parkinson disease obtained a mean
score of 136.27 (SD = 7.45) and the control group mean was 136.94 (5.72). Statistical
comparison of group performance revealed that these two groups did not differ on the
DRS-2 (t1,37 = -0.32; p = 0.37). Analysis of performance on the Beck Depression Index
provided another pattern. Although the depression index for the Parkinson group was not
high, it did differ significantly from the Control group (t1,37 = 2.25; p = 0.015).
Analyses
Dependent measures collected for gait in this investigation included the spatial
and temporal measures of stride length, step velocity, and percent double support time.
According to the GAITRite© Manual, stride length is defined as the line of progression
between the heel points of two consecutive footprints of the same foot (left and right
were averaged), measured in centimeters. Step velocity was obtained after dividing the
7
Cognitive-linguistic load
Distance Traveled by the Ambulation time and was expressed in centimeters per second.
Finally, the GAITRite© system obtains a percentage of time spent stabilized on two feet
for each pass on the walkway which is expressed as percent double support time (again
averaged for left and right) To examine potential effects on speech, speech rate (syllables
per second) and task accuracy (rate of errors) were calculated .
To compare performance across groups and conditions, two factor repeated
measures ANOVAs were employed. The between group factors were group (PD and
Control), and the within factors were the load conditions (baseline, low-load, middle-
load, and high-load). The Bonferroni correction for multiple comparisons was applied
when indicated and an alpha of .05 was adopted to determine statistic significance.
RESULTS
Gait Measures
A two-factor repeated measures ANOVA revealed that there were no significant
differences between groups (PD vs. controls) for stride length (F1,38 = 3.64, p =0.065).
Therefore, the data from the two groups were collapsed and a one factor repeated
measures ANOVA was conducted that revealed a condition effect (F3,150 = 4.08, p =
0.008). Post hoc analyses indicated that individual stride length was significantly reduced
in the high-load condition when compared to baseline performance (see Figure 1).
The results for velocity of gait were strikingly similar. As with stride length, there
were no significant differences between groups for step velocity (F1,38 = 4.09, p =0.051),
but once again when the data were collapsed there was a condition effect (F3,150 = 4.08, p
= 0.008) and a significant difference existed between the conditions of baseline and high-
load in the absence of significant findings across other conditions (Figure 2).
8
Cognitive-linguistic load
When percent double support time was analyzed, a different pattern emerged.
The two factor repeated measures ANOVA revealed a group difference (F1,38 = 4.42, p
= .04) with a significant interaction. The interaction reflected that the control group
significantly increased double support time across conditions (Figure 3), while the group
with Parkinson disease group did not.
Speech Measures
Analysis indicated that speech rate, measured in syllables per second, was not
different between the groups (F1,38 =.28, p = .04). Significant differences were found
across tasks, however (F3,150 = 69.35, p < .001). The rate of speech was M= 2.99(.65)
syllables/second during conversation, M= 4.30 (1.25) syllables/second during low-load,
M= 2.07(.82) syllables/second during middle-load, and M = 1.34(.64) syllables/second
during the high-load condition. Post hoc analyses indicated that speaking rate during
conversation was significantly faster than each of the load conditions and rate during the
high-load conditions was significantly slower than the other conditions.
A low occurrence of errors with regard to task accuracy precluded formal
analysis. The group of individuals with Parkinson disease made no errors in the low-load
condition, had a mean error rate of 0.78 (range 0-8 errors) during the middle-load
condition, and a mean error rate of 0.57 (range 0-4 errors) during the high-load condition.
Similarly, there were no errors during the low-load condition for the control group, a
mean of .26 (range 0-1 errors) during the middle-load, and a mean of .13 (range 0-1
errors) during the high-load condition.
DISCUSSION
9
Cognitive-linguistic load
The purpose of the current investigation was to expand the extant research on
talking while walking by manipulating the cognitive-linguistic complexity of verbal tasks
while walking. Specifically we studied the effects on individuals who might be at risk
for injurious falls, in this case, participants who had been diagnosed with Parkinson
disease. Our results indicated that cognitive-linguistic demand did have an impact on
gait, but not on the manipulated parameters of speech. Specific findings are discussed
below.
Gait Measures
Analyses on gait measures revealed that manipulating the complexity of
cognitive-linguistic tasks affected parameters of gait during simultaneous walking and
talking. Those effects were demonstrated in individuals without neurological
compromise as well as those with Parkinson disease, a result not previously reported as
earlier studies did not include a comparison group (Bowen, et al., 2001; O’Shea, et al.,
2002). A priori assumptions might have anticipated differences between the groups due
to the motor implications of Parkinson disease, however. One factor that may have
influence the results may have been that our sample of individuals with Parkinson disease
was not well represented across the spectrum of severity. Over 70% of this sample was
at H&Y stage 2 indicating that individuals’ balance was not impaired (independently
ambulatory) but that features of Parkinson disease had become bilateral (Hoehn & Yahr,
1967). A trend was suggested in the anticipated direction for both stride length and gait
velocity but the difference did not reach statistical significance. It may be that a sample
that includes a more broad range of severity would bear out the trends observed in this
study.
10
Cognitive-linguistic load
Although group differences were not demonstrated for stride length or gait
velocity, we found a significant difference with respect to percent of double support time.
Because double support is a stabilizing feature of gait, the implications for increased
double support time are a need for individuals to stabilize their gait. Visual inspection of
the data reveals that both groups in the current study increased double support across
conditions at similar rates. The notable difference, however, is in the variance on this
measure in the group with Parkinson disease (Figure 4). Clearly there were individuals
in the Parkinson group who responded to the more difficult verbal tasks by stabilizing
gait, but as a group this was not a consistent strategy. It is likely that the proportion of
our sample at H&Y stages 3-4 (around 30%) contributed to the variance. This finding on
the use of double support time in gait may contribute understanding to those findings of
Wood, et al. (2002) who reported the high incidence of falls (~68%) and repeated falls
(50%) in a large sample of people with Parkinson disease. Further, the dependent
measures of gait in the current study were only a small sample of the parameters that can
be collected and recorded with the GAITRite© sytem. Additional influences on gait may
be revealed with inclusion of more analyzed parameters including Functional Ambulation
Profile (FAP), an algorithm of gait parameters that has been shown to predict falls
(Nelson, 1999)
Speech Measures
The effect on gait during manipulation of the complexity of the speaking task was
a primary focus of this study. A secondary interest was the potential reciprocal effect on
the speaking tasks themselves. The finding of significant differences in speaking rate
11
Cognitive-linguistic load
across conditions was not surprising given the nature of the different tasks. What was
unexpected, was the
minimal impact on speech execution and task accuracy. We anticipated that as task
difficulty increased, there would be detrimental effects on both gait and speech. In fact,
there were several individuals in the Parkinson group who demonstrated striking effects,
such as freezing of gait and dramatic changes in respiration (clavicular breathing) during
the middle and high load cognitive-linguistic conditions. Once again, these individuals
were those who were more severely impaired. Perhaps the measures of speech selected
were not sensitive enough to pick up these effects of task complexity, although rate
appeared a better candidate than task accuracy. Measures that might be of interest to
pursue, based on the observation of some of our participants, would be physiologic
measures such as cardiac output and respiratory patterns. Another plausible
interpretation of these observations and findings might be that, when faced the changing
demands of the speaking task, participants focused more of their resources on that task.
Such a strategy may have negative results if it places them at an increased risk for an
injurious fall.
CLINICAL IMPLICATIONS
The results of the current investigation offer important clinical implications. If
increasing cognitive/linguistic load disrupts parameters of gait, it may place elderly
individuals at increased risk for injurious falls. Moreover, those elderly individuals with
neurological compromise may be at a greater risk. These results suggest that it might be
prudent for health care professionals and caregivers to alter expectations and monitor
cognitive-linguistic demands placed on these individuals while they are walking,
12
Cognitive-linguistic load
particularly during increased risk situations such as descending stairs, in low light
conditions, or avoiding obstructions. Although low-load or automatic tasks may not
significantly alter gait parameters, those demands that require formulation, increased
mental manipulation, or taxing cognitive loads may have the potential to affect gait
performance and potentially increase the risk of injurious falls.
REFERENCES
Bowen, A., Wenman, R., Mickelborough, J., Foster, J., Hill, E., & Tallis, R. (2001).
Dual-task effects of talking while walking on velocity and balance following a
stroke. Age and Aging, 30, 4, 319-323.
Bloem, B.R., Grimbergen, Y.A., Cramer M., Willemsen M., & Zwinderman, A.H.
(2001). Prospective assessment of falls in Parkinson's disease, Journal of
Neurology, 248, 950-958.
Carroll N.V., Slattum, P.W., & Cox, F.M. (2005). The cost of falls among the
community-dwelling elderly. Journal of Managed Care Pharmacy, 11, 307-316.
Hoen MM, Yahr MD. Parkinsonism: Onset progression and mortality. Neurology
1967;17:427–442. 21.
Jurica, P., Leitten, C., & Mattis, S. (2001). The Dementia Rating Scale, 2nd ed. manual.
Psychological Assessment Resources.
LaPointe, L. L. , Heald, G.R., Stierwalt, J.A.G. & Kemker, B.E. (2007). Effects of
auditory distraction on cognitive processing in young adults. Journal of Attention
Disorders, 10, 4, 398-409.
Nelson, A.J., Certo, L.J., Lembo, L.S., Lopez, D.A., Manfredonia, E.F., Vanichpong,
S.K., & Zwick, D. (1999). The functional ambulation performance of elderly
13
Cognitive-linguistic load
fallers and non-fallers walking at their preferred velocity. NeuroRehabilitation,
13, 141 – 146.
O’Shea, S., Morris, M.E., & Iansek, R. (2002). Dual task interference during gait in
people with Parkinson disease: Effects of motor versus cognitive secondary tasks.
Physical Therapy, 82, 9,888-897.
Tinetti, M.E., & Williams, C.S. (1997). Falls, injuries due to falls, and the risk of
admission to a nursing home. New England Journal of Medicine, 337, 1279-
1284.
Shkuratova, N., Morris, M.E., & Huxham, F. (2004). Effects of age on balance control
during walking. Archives of Physical Medicine and Rehabilitation, 85, 4, 582-
588.
Tun, P.A., O’Kane, G., & Wingfield, A. (2002). Distraction by competing speech in
young and older adult listeners. Psychology and Aging, 17, 453-467.
Wild, D. Nayak, S.L., & Isaacs, B. (1981). Prognosis of falls in old people at home.
Journal of Epidemiology and Community Health, 35, 200-2004.
Wood B.H, Bilclough J.A, Bowron A., Walker R. (2002). Incidence and prediction of
falls in Parkinson's disease - a prospective multidisciplinary study. Journal of
Neurology, Neurosurgery & Psychiatry, 72, 721-725.
14
Step Length
124.61(19.23)
123.37 (20.04)
117.02 (23.70)
111.99 (23.54)
0
20
40
60
80
100
120
140
160
Baseline Low Medium High
Condition
Centimeters
Cognitive-linguistic load
Figure 1. Mean stride length (standard deviation), with data collapsed across groups (Parkinson disease, and control group).
15
Step Velocity
86.25* (31.66)
93.18 (30.84)
104.28 (27.21)
106.48 (24.34)
0
20
40
60
80
100
120
140
Baseline Low Medium High
Condition
cm/second
Cognitive-linguistic load
Figure 2. Mean gait velocity (standard deviation) with data collapsed across groups (Parkinson disease, and control group).
16
Double Support
30.30 (3.3)
29.98 (2.4)
28.64 (2.1)
27.47 (2.6)
36.9 (11.7)
34.46 (10.0)
32.66 (7.5)
32.32 (7.1)
0
10
20
30
40
50
60
Baseline Low Medium High
% Double Support
Cognitive-linguistic load
Figure 3. Mean percent double support (standard deviation) for the two groups (Parkinson disease group is depicted with gray, and
the Control group with black)
17
