1. Introduction
Walking is generally viewed as an automated, smooth, rhythmic motor task. With advancing age, reduced function of different systems used for walking may affect walking ability and increase risk of falling. Studies of the ability to walk while performing simultaneous tasks show that the dependence of sensorimotor control on cognitive processes increases with advancing age [1]. A dual-task situation may either affect walking performance, a second activity, or both, depending on individual resources, priorities, and the complexity of the tasks [2]. Change in gait patterns related to Dual Tasking (DT), identified as a Dual-task Cost (DTC), can be attributed to competition for limited resources in the Central Nerve System (CNS) in older people even when they are healthy [1, 3].
It can also increase gait stability and reduce the risk of a fall by changing the strategies needed to address high-level attention and task demands [1]. To maintain efficiency and smoothness in walking, the Central Nerve System (CNS) should consider changes in control strategies [4]. According to recent studies, walking in older people is not an automatic process. Because if it were an automated process, it would not be possible to make changes in the gait pattern and gait speed in the presence of attention demands and the cognitive load [2, 3]. Instead, many studies have shown that DT affect walking speed and decreases with a cognitive load while walking, walking speed, and gait stability [1, 5, 6].
Secondary tasks taken during walking can be put into general categories of motor and cognitive tasks. Among these tasks, the secondary motor task reduces the protective reflexes of the upper extremities for the purpose of requiring high-level motor coordination. Also, when considering attention needs, the use of secondary tasks that have a purely cognitive aspect seems more appropriate than the motor task, which consists of a combination of motor and cognitive aspects [4].
The cognitive load is the mental resource needed to perform the task. An increase in the cognitive load in different tasks while walking (such as counting) imposes a different cognitive load on the central processing system [7]. According to the theory of capacity sharing in the dual-task paradigm, implementing additional tasks in walking may change walking features or the implementation of the second task or both [8]. Dubost et al. reported an increase in walking variability in the elderly while performing simple computational tasks [9]. Lindenberg et al. found that dual-task disorders increased with age, and disorders such as a decrease in walking speed and an increase in the number of wrong steps during walking seemed to appear [10]. However, changes in the natural level can lead to unstable motion and an increased risk of falling [11].
It seems that DT walking, which in many cases is cognitive, involves the individual. These factors can affect stability while walking and falling in a person. Most daily activities need doing more than one activity, such as walking and talking. Some of the fall cases in the elderly occur when there is a need to focus on controlling posture and performing cognitive tasks simultaneously. Knowledge about motor control in healthy people can be used as a base for comparing the mechanism of disease change. The use of dual cognitive tasks seems useful and can be used both in the identification of the population at risk and in the evaluation of the results of rehabilitation programs for preventing falls in the elderly. However, no studies have examined the effect of different cognitive loads on gait stability in healthy younger and older adults simultaneously. To address this issue, we are looking for the answer to the question “what is the effect of different cognitive load on gait stability”?
2. Materials and Methods
Subjects
This study adopted a causal-comparative design. The statistical population consisted of university staff. Ten younger adults (Mean±SD age of 25.91±3.42 years with Mean±SD BMI of 24.10±2.33 kg/m2) and 10 healthy older adults (Mean±SD age of 66.65±4.28 years with Mean±SD BMI of 26.56±2.35 kg/m2) were healthy volunteers of this study without a history of falls (sample size suggested by Byrne et al. with 80% power efficiency to detect group differences at 0.05 level) [12]. All participants were healthy and able to walk for at least 90 seconds without the help and understanding of the verbal instructions. If they had neurological, musculoskeletal, visual, auditory, vestibular, or neural disorders that affected walking or reducing the cognitive score, they would be excluded from the study. The older adults acquired a mini-mental state examination score of 24 or higher [13]. All the individuals completed the informed consent form before their participation. The approval of the Ethics Committee of Kharazmi University Research was issued under the number 100.1000-K.A.P.
Walking task
A treadmill (HP-cosmos Mercury®Med, Germany) with two piezoelectric force plates of Kistler Company was used for walking. These force plates can output the number of steps, step length, cadence, and walking speed [7]. The participants conducted a treadmill test at the preferred speed with bare feet at speeds of one km/h and a gradual increase of 0.5 km/h to the speed stated by the participant [14]. After recording the speed, the participants were again asked to step on the treadmill and confirm their desired speed at their preferred speed determined by them with a 0.5 km/h speed decrease or increase [15]. This method was repeated four times, and eventually, the preferred speed of each participant was determined. For each participant, walking tasks were done three times at the preferential speed with cognitive load (none, simple and complex). The data were recorded on the treadmill for 90 seconds [4, 16]. For the acquaintance of the participants with the laboratory environment and the test, after a full description of the test, each participant practiced on a treadmill experiment several times. The time needed to rest between the tests was given to the individuals. Safety considerations (Harness) were observed during the test.
Cognitive task
The cognitive task of the study was the mental calculation task. The mental calculation task involves working memory and subtracting three by three (complex cognitive task) or one by one (simple cognitive task) from a random three-digit number, which continued for 90 seconds [17]. In the non-cognitive task, the subjects underwent a treadmill at their preferred speed for 90 seconds [4]. The participants were asked to do a cognitive task while sitting on the chair and or moving on the treadmill, and their points were recorded. Intervals, words, or false calculations were recorded for each of these tests. Meanwhile, all speech tests were recorded with a tape recorder. The participants did not have any previous training on this task. As subjects performed the tasks for 90 seconds, they had little learning effect on subjects. The Cognitive Performance Scale (CPS) was calculated by formula 1. 
Formula (1)




The cognitive task was made to changes in Correct Response Rate (CRR) (Formula 2) because of the dual-task to individual conditions. In this study, the cognitive function score was calculated for each task based on the formula 1 [4, 18].
Formula (2)


CRR= Response rate per second × Percent of correct answers response


Gait stability
Measurements were stated with respect to walking speed, and step length changes were expressed by Cromwell et al. [19]. This size includes GSR, which is a cadence ratio (steps per second) to speed (m/s), whose unit is steps per meter. GSR expresses gait stability, and its increase shows the increase of steps per meter, meaning that the person has a double support level and better gait stability. The researchers stated that this method had good validity and could be used to evaluate and measure gait stability in the dynamic movement. They reported the reliability of the test through intraclass correlation coefficient method for walking speeds as 0.91 and for a cadence as 0.75 [19]. The test-retest reliability using the correlation coefficient for GSR was 0.81 in this study.
Statistical analysis
The Kolmogorov-Smirnov test checked the correlation of the frequency distribution of quantitative variables with normal theoretical distribution. Two-way repeated measures ANOVA evaluated the main and interactive effects of the group and cognitive load on dependent variables with a 95% confidence level and Bonferroni posthoc test.
3. Results
Table 1 presents the individual characteristics of the participants. According to the Kolmogorov-Smirnov test, the data were normally distributed (P≥0.05). Figure 1 shows the mean CPS in adult and elderly groups at preferred walking speeds. Using this score, the speed and precision of a person’s cognitive task are taken into account. The higher CPS indicates the weaker person’s cognitive performance (less correct answer) while dual-task walking and vice versa. The elderly seem to have complicated cognitive and weaker CPS (Figure 1).

Table 2 presents the mean of the parameters involved in gait stability at different cognitive loads. The results of two-way repeated measures ANOVA for GSR show a significant difference between the younger and the older adult groups (F1,2=15.64, P=0.001). There is also a significant difference in the cognitive load for GSR (F2, 2=26.47, P=0.024).

Comparison of means by post hoc test showed that complex cognitive load in the older adults was significantly different from cognitive (simple and none) in younger adults (P≤0.05), and at the time of complex cognitive load, gait stability is more in the older adults. Also, in younger adults, complex cognitive load has a significant difference with/without cognitive load (P≥0.05). The results show a significant difference between older and younger adults in different cognitive loads (P≤0.01) (Table 3). The older adult seems to be more successful in learning gait stability (Figure 2).

4. Discussion
The natural aging procedure renders the elderly with a higher risk of falling mostly during walking [20]. Therefore, understanding the mechanism of different cognitive load effects on gait stability in younger and older adults is essential. The results showed that complex cognitive load exercises during preferential walking in the younger and older adults led to a decrease in walking speed and increase in learning gait stability. It was determined that by different cognitive loads (none, simple and complex), the GSR value in older adults is more than younger adults. Cromwell et al. argued that GSR levels were higher in the elderly, meaning that the elderly took more steps in the same distance; so, their gait stability increased with this mechanism. An increase in gait stability while walking makes the elderly compensate for reduction in balance to some extent.
Thus, by maximizing gait stability, the elderly create a pattern of motion that provides greater resistance to disorder, acting as a mechanism for protection against fall [19]. However, this mechanism reduces the forward speed of the elderly, and seems that the speed is sacrificed for stability while stepping. By controlling the movement of the lower limb joints, CNS adapts the walking pattern and, thus, reduces walking speed [19]. It seems that the speed is being sacrificed for gait stability, and the older adult will pay more attention to walk than the younger adult [2, 6, 9]. Thus, walking turns into a complex multi-tasking behavior from a rhythmic and automated behavior, whose control is more difficult that makes the person more likely to fall [6].
This difference was also more common while complex cognitive tasks are done simultaneously. Hall et al. stated that the pattern of walking was changing during cognitive tasks [18]. To maintain vertical posture, CNS must integrate sensory input and retrieve weight distribution information from different sensory systems (vision, vestibular, and sensory) and continually regulate the neuromuscular system [21]. Although this process is automatic, using dual tasks, various studies have shown that tasks such as walking at different speeds and performing complex cognitive tasks need sources of attention [9, 22, 23]. Thus, there is always a competition on the limited resources for control of posture and implementation of dual functional and cognitive tasks [24].
In old age, the cognitive processes of sensorimotor control become increasingly important, and adaptation in the elderly walking pattern tends towards adopting the necessary strategies to increase gait stability [1]. Although body capacity to move forward reduces, this strategy significantly reduces the likelihood of falling in the elderly by increasing stability [19]. Older adults have less walking speed than the middle-aged and younger groups, and this speed often reduces while dual tasks. There are also variations in walking speeds between the older and younger adults in the normal state and dual-task mode while walking [11]. Speed, step length, and double support are among the significant components in achieving stability and reducing falls among individuals [25].
To maintain efficiency and smoothness in walking, CNS should consider changes in control strategies [4]. As cadence is similar in adults and the elderly, the elderly cover a shorter distance with shorter steps with the same number of steps compared to adults, and this pattern makes the elderly have more time in the double support phase, causing a more stable walking pattern [25]. However, the walking pattern in adults is associated with unstable phases, and this pattern makes adults have more effective steps forward along with the changes in the center of gravity in each step [19], which is consistent with the results of this study.
There are different theories explaining how to respond to dual tasks, including bottleneck theory and the capacity-sharing theory [2, 26]. Some findings suggest that when the subtractions are performed, competition over limited resources becomes more attentive and demands for further coordination and attention from the CNS. This is the ability to focus on the underlying factors in the performance of DT. On the other hand, as the age increases, the ability of the elderly to devote attention decreases, and they have difficulty in implementing DT [2]. Probably the older adult of the present research paid the closest attention to the dual-task and successful implementation of this, which made it possible to increase the amount of time spent on the route and change the main task (gait).
Walking and simultaneously performing complex cognitive tasks, working memory, and sustained attention can explain the implementation of walking along with cognitive tasks (such as three-point deductions) [27]. It seems that the harder the cognitive task, the more the cost of the dual-task for walking and vice versa, which is completely in line with the results of our study [18].
In challenging motion load and DT, control effort increases, and the control system tries to reduce the control effort on the system. This is because the control capacity is limited, and the system uses different mechanisms to compensate for this disorder, which is more tangible among the youth who has better performance. Nonetheless, the older adults create double support while walking by different walking strategies, especially the lowering phase, so that it decreases with increasing in GSR falling [4, 28]. It seems that the adoption of walking exercises, combined with complex cognitive load in older adults, can provide better motor control. Also, in particular situations where complex synchronous tasks are possible, the best performance of the control system is provided and reduces the risk of falling.
Some elderly, who were physically and mentally ill, did not meet the inclusion criteria of the study because of the inability to perform study cognitive and functional tasks. The researcher did not control the sleep, rest, and nutritional status of the subjects before the tests. The treadmill use could also be a potential limitation for the study as treadmill kept speed walking of the individual constant in all conditions that could artificially alter the natural variability and attention needs of walking. However, as each participant should be tested according to the same controlled conditions and their cognitive load and walking speed, one could state that walking on the ground was different a little from the current measured values, which were among the limitations of the study.
This study demonstrated that in older adults, the cognitive load could significantly alter gait strategies during walking. These changes may indicate the high adaptability of the CNS to modify control strategies to maintain the stability, rhythmicity, and smoothness of walking during challenging conditions.
Ethical Considerations
Compliance with ethical guidelines
The research was approved by the Ethics Committee of Kharazmi University under the number of 100.1000-K.A.P.
Funding
This study was part of the PhD. dissertation of the first author, Department of motor Behavior, Faculty of Physical Education and Sport Sciences, Islamic Azad University, Central Tehran Branch, Tehran, Iran.
Authors' contributions
All authors contributed in preparing this article.
Conflict of interest
The authors declared no conflict of interests.


References