Motor Learning and Working Memory in Children: The Role of Cognitive-Motor and Motor-Motor Dual-Task Training

The present study aims to examine the effects of two dual-task training methods (motor-motor and cognitive-motor) as well as a single task method on static and dynamic balance and also on the working memory in children. Forty-five children (all male; mean age 8.82 ± 0.83 years) were selected and randomly assigned into three experimental groups. In the pretest, posttest, and retention stages, the participants took static, and dynamic balance tests, as well as an n-back task. In the training stage, the participants practiced selected balance training tasks under dual-task motor-motor and motor-cognitive conditions as well as a single task over a period of four weeks at two sessions per week. The results of the paired samples t -test indicated that children in all groups improved their balance performance and working memory ( p ≤ .05). The results of ANCOVA showed that the balance training group under motor-motor dual-task conditions outperformed the other two groups in terms of the dynamic balance scores ( p ≤ .05). An improved performance was also observed for the cognitive-motor dual-task conditions compared to the single-task group ( p ≤ .01). In terms of static balance and working memory, both dual-task groups, regardless of the types of their tasks, outperformed the single task group ( p ≤ .05).

• Cognitive-motor dual-task training promotes dynamic balance of children.
• Dual-task training promotes statice balance and working memory of children. • Children benefit from dual-task balance training regardless of the type of the secondary task.
People in their everyday lives often need to perform more than one task at a time (Plummer & Eskes, 2015). The challenge of performing two tasks simultaneously, often referred to as "dual task", has been the subject of a wide range of studies in the existing literature (Nebes et al., 2001). The dual task paradigm involves a combination of two motor tasks with different goals or a single motor task combined with a cognitive task (Broeder et al., 2014). This paradigm is generally used to study the ability to perform simultaneous tasks in different populations, providing useful information including on how to prevent falls in the elderly (Zijlstra et al., 2008) or enhancing learning processes in children (Reilly et al., 2008). Cognitive and motor aspects of a secondary task both can be helpful in selecting optimal task parameters to be used in instructional settings (Remaud et al., 2012).
Dual-task paradigms are also used to identify developmental changes in different components of working memory (Jaroslawska et al., 2018). Working memory is responsible for creating complex cognitive coordination needed to perform daily tasks (Doherty et al., 2019). The stronger working memory is, the more extensive is the cognitive capacity, and therefore individuals with stronger working memory are more successful in multitasking (Hitch et al., 2001). In other words, enhanced working memory and the resulting improvement in cognitive function seem to be effective in improving performance and learning in children (Seidler & Anguera, 2012).
Studies to assess executive functions and motor performance using cognitive-motor dual-task training have shown that this type of intervention has positive impact on executive and balance functions (Wang et al., 2016). However, when it comes to motor-motor dual-task training, there is no strong evidence to suggest effectiveness of this type of intervention (He et al., 2018). Shin and An (2014) showed significant improvement in balance in elderly women following an intervention based on motor-motor dual-task training, although a more recent study indicated that performance quality of dual tasks linked to postural control will not reach an adult level by at least the age of fifteen or sixteen (He et al., 2018). However, other studies suggest that cognitive-motor dual-task training outperforms motor-motor dual-task training in this area (Nourozi et al., 2019). Thus, it is very important to examine the extent to which various paradigms of motor-motor and cognitive-motor dual-task training influence motor learning, and performance in children in order to identify the impact of these techniques compared against the single task settings. More importantly, this can be combined with an assessment of children's working memory as a major factor contributing to cognitive performance.
To the best of our knowledge, no previous investigation analyzed the application of the dual-task paradigm in improving the cognitive performance in children, because available evidence on dual-task training is often linked to enhancing cognitive and motor performance in the elderly and injured individuals (He et al., 2018;Nourozi et al., 2019), further assessments are needed to identify how and to what extent dual-task training interventions may improve cognitive and motor performance and learning in children.
In other words, there is important evidence suggesting improved balance and working memory following dual-task training interventions (Bustillo-Casero et al., 2020). However, types and ways of presenting these dual tasks may influence how these interventions affect balance and working memory. Therefore, the present study attempts to address the question of whether different balance training methods based on motor-motor and cognitive-motor dual-task training alongside with single task training can influence motor performance and motor learning as well as working memory in children.

Method Participants
Our sample consisted of 45 children (all male; mean age 8.82 ± 0.83) selected from the local schools through convenience sampling and randomly assigned into three groups of 15 for balance training based on the motor-motor dual-task group, balance training based on the cognitive-motor dual-task group, and balance training based on the single-task group. Inclusion criteria were; 1) being 8 to 10 years old; 2) completing consent form by the parents for participation in the study; and 3) absence of any skeletal-muscular injuries or disease. Individuals were excluded from the study if they; 1) were not motivated to continue participation; and 2) experienced any disease or injury that emerged during the study and might affect performance. Research design was developed based on the Declaration of Helsinki and approved by the Research Ethics Committee of Shahid Chamran university of Ahvaz. The participants' parents completed the informed consent form before the test started.

Stork Balance Test
This test is used to measure static balance. The subject is required to place his/her hand on the waist (over the iliac crest) with the non-preferred foot pressed against the inside knee of the other leg. Then, keeping this position, the subject should balance on the ball of the preferred foot. A stopwatch is used to record the time (s) to a hundredth of a second from the moment when the subject stands on the ball of the foot to the moment when he or she loses balance, i.e., when hands come off the waist, the knee or the upper body is bent, the supporting foot swivels, or the non-supporting foot loses contact with the knee. Previous studies have reported good reliability of this test (Latorre Román et al., 2017).

Y-balance Test (YBT)
This dynamic balance test is a modified form of the star excursion balance test for static balance. The test targets the preferred leg. The participant is required to stand on one leg while reaching for the ground in three different directions. The participants in this study first completed six trials to warm up and minimize learning effect in any of these directions. The procedure involves three trials in the anterior reach direction, three trials in the posteromedial reach direction, and three trials in the posterolateral reach direction. The maximum reach distance (cm) is recorded for each successful trial. A trial is considered invalid if the reach indicator is tapped onto to increase the distance, the reach indicator is relied on for a longer period for further support, the foot touches the ground, or balance is lost. Invalid trials are excluded from calculations. Studies have reported a good reliability for this test (Yam et al., 2019).

N-back Test
In this test, participants receive a sequence of consecutive visual stimuli. A participant should indicate whether the current stimulus matches the one from N steps earlier. The larger N is, the more difficult is the task. Generally, N is a number from 1 to 3. In this study, we used N = 2. Once the data for a participant was recorded and the type of the test (2-back test) was selected, the participant entered a 30-second initial test stage where he or she could see the results after pressing each button. The initial stage could be repeated until the participant was ready for the main task. The 3-minute main task involved stimuli in the form of the numbers 1 to 9 shown consecutively for 1 second each. The participant would start the comparison from the third stimulus on; that is, the participants were required to compare the third stimuli with the first one (two earlier steps), and then press the "Yes" button if the two stimuli matched and the "No" button otherwise. This process would continue by comparing the fourth stimulus with the second one, the fifth one with the third one, and so on. The output of the program for each participant indicated the number of correct responses, the number of incorrect responses, the number of no-response cases, success rate (percentage) and the mean response time (s) (Ciesielski et al., 2006;Pelegrina et al., 2015).

Procedure
First, the participants were selected through convenience sampling. All participants were novice in terms of skill level, taking the tests, and performing the tasks as they had no background of being familiar with the procedure. Prior to the tests and tasks, the participants were given some information on the procedure. After receiving the initial information, the examiner provided the participants with information on the process involved in the tests and different dual-task trainings. Next, all participants attended the pretest session at an indoor hall to take the stork balance test, YBT, and the n-back test. Immediately after the pretest session, the participants were randomly assigned into three groups; Group 1: balance training based on motor-motor dual-task paradigm; Group 2: balance training based on cognitivemotor dual-task paradigm; and Group 3: balance training based on single-task paradigm and were included in the training session. The study consisted of four stages: pretest, training, posttest, and retention test. The participants took static and dynamic balance tests as well as the n-back test in three stages of pretest, posttest (immediately after the training session), and retention test (one week after the end of training session). In the training stage, the participants practiced balance training under motor-motor, and cognitive-motor dual-task, as well as single-task settings over a period of four weeks with two 45-minute sessions per week. The balance training used here was a modified version of the training used by Shin and An (2014). As training sessions moved forward, the training program was developed into a more complex program as classified by Gentile (1987). The children were then asked about their preferred foot and hand, the foot with which they hit the ball or the hand with which they threw the ball. In this study, only right-handed, and right-footed children were tested. The protocol used for the group with balance training based on motor-motor dual-task paradigm involved standing on a balance pod while catching and throwing a sports ball (with the right, left, and both hands), with two individuals standing face to face 1.5 m apart performing a trial under all three conditions. The distance was increased to two meters if 80 percent of the trials were successful in terms of keeping balance and properly receiving and throwing the ball. Trial sets were interspersed by 2-minute rest periods. The protocol used for the group with balance training based on cognitive-motor dual-task paradigm involved standing on a balance pod, with eyes open, while performing a cognitive task developed by Shumway-Cook (2001) that involved counting down numbers, counting down weekdays, and months of the year, spelling proper names backwards, and recalling the order of a word list. As for the group with balance training based on single-task paradigm, the participants only practiced the balance training protocol during the sessions with no secondary task involved.

Data Analysis
The assumptions of equal variance and normality of the data were tested using the Levene's test of equality of error variances and Shapiro-Wilk's test respectively and acceptably confirmed for all data. One-way analysis of covariance (ANCOVA), one-way analysis of variance (ANOVA), paired samples t-test and Bonferroni post hoc tests were used to compare the groups. The data were analyzed at α ≤ .05 in SPSS 24.

Results
The assumptions of equal variance and normality of the data were tested and acceptably confirmed for all data. Table 1 reports descriptions of individual characteristics and research variables. As seen in this table, according to the result of one-way ANOVA for initial comparison, all groups had similar scores in terms of static and dynamic balance as well as on working memory in the pretest stage.

Balance Performance
In the present study, balance performance was analyzed through two indicators: static balance and dynamic balance.

Working Memory
In the present study, working memory was analyzed through five indicators: incorrect response, no response, correct response, success rate, and mean response time.

Discussion
Our findings showed eight sessions of balance training under motor-motor dual-task training resulted in significant improvement in children's dynamic balance in both acquisition, and retention stages compared to cognitive-motor dual-task, and single-task groups. Similar positive impacts were observed in cognitive-motor dual-task conditions compared to the single-task group. As for other measures assessed by this study, including static balance and working memory, balance training under motor-motor dual-task settings was almost the same as cognitive-motor dual-task settings, and both of these groups outperformed the single task group. In other words, training under dual-task settings, regardless of the type of the secondary task, resulted in a significant improvement compared to the single-task settings in terms of variables such as static balance and working memory.
Studies have shown that motor training can induce positive changes in the brain structure to enhance motor learning and performance by activating a number of brain regions like the prefrontal cortex (Brehmer et al., 2012;Theill et al., 2013). This improved performance can continue through cognitive trainings (Theill et al., 2013). In addition, this improvement can be amplified when trainings are provided simultaneously using the dual-task paradigm. Such a dual-task group, regardless of the type of the secondary motor or cognitive task, can engage higher-level neural centers to enhance not only motor performance in tasks that involve balance performance but also cognitive performance and working memory in individuals (Norouzi et al., 2019). However, a number of studies have reported greater improvements in motor and cognitive functions as a result of applying cognitive-motor dual-task settings compared to motor-motor dual-task settings (Norouzi et al., 2019). This line of studies attributes this better performance to the neural efficiency resulting from training under cognitivemotor dual-task settings as well as growing brain connections. However, our findings were different, particularly in terms of dynamic balance performance as these findings suggest that it was the motor-motor dual-task group that outperformed both the cognitive-motor dual-task group and the single-task group. One reason behind these different findings may lie in the age of the participants, i.e., the fact that they were children.
Most studies in this area targeted adults or the elderly (Choi et al., 2015;Norouzi et al., 2019). It is well known that children have less and more variable motor capacity and cognitive function than young and older adults (Mickeviciene et al., 2019), so, it is not surprising that our results are different from the results found for adults. On the other hand, the existing evidence suggests that greater improvement in motor performance can be achieved by using primary and secondary tasks that are more similar to each other (Silsupadol et al., 2009). It seems that the primary tasks used in the balance trainings in our study were more similar to such secondary tasks as catching a ball than to cognitive secondary tasks like counting down numbers and this similarity, in turn, has resulted in a greater improvement in dynamic balance on the participants. PSIHOLOGIJA, 2022, OnlineFirst, 1-17 Furthermore, there are other studies that suggest identical impact of motormotor and cognitive-motor groups (Akin et al., 2021), a finding which is largely consistent with a major part of our findings. For example, by comparing these two groups, Akin et al. (2021) found that both methods, i.e., motor-motor and cognitive-motor groups, contributed to the improvement in balance and motor performance of the elderly to the same degree.
Our findings in connection to further improvement in balance performance as a result of dual-task training regardless of the type of the secondary task used compared to the improvement attained through the single-task group are consistent with major studies in this area (Goh et al., 2012;Jahanbakhsh et al., 2020;Jehu et al., 2017;Norouzi et al., 2019;Roche et al., 2007;Segev-Jacubovski et al., 2011;Silsupadol et al., 2009). For instance, in a study on dual-task trainings focused on balance and cognitive tasks, Choi et al., (2015) found that these training may improve balance and cognitive performance compared to ordinary (single-task) settings for patients with acute brain stroke. In another study, Demirdel & Erbahçeci (2020) demonstrated improved stance for the amputees in the dual-task group in a stork balance test. These findings suggest that dual-task training is effective in improvement and automaticity of static balance. In other words, they indicate that the attentional demand needed to perform a task was reduced according to the automaticity hypothesis (Silsupadol et al, 2009). Greater improvement in static balance for the dual-task group shows that the dual-task group is effective in reducing the risk of fallings and automaticity of balance control needed in everyday activities. Therefore, it seemed that in the present study, the children in the dual-task group managed to enhance their automaticity in performing motor skills while improving their working memory, and so experienced better conditions that the children in the single-task group.
Furthermore, previous studies also pointed to improved balance performance under dual-task and single-task training settings for patients with multiple sclerosis and brain stroke (Kim et al., 2013;Monjezi et al., 2017). However, the greater post-intervention improvement in performance for the dualtask group suggests that dual-task training is a more effective way for enhancing coordination ability to simultaneously perform two tasks. In other words, regardless of the type of the secondary task, balance trainings based on dualtask group can improve static and dynamic balance by reinforcing sensorimotor integration (McNevin et al., 2003). This sensorimotor integration causes this improvement through greater interaction between voluntary and reflex actions. Training under dual-task conditions can result in using more degrees of freedom by an individual and this, in turn, can facilitate the performance of motor skills (McNevin et al., 2003). Similarly, in the present study, both motor-motor and cognitive-motor dual-task groups resulted in a better dynamic and static balance. The participants seemed to improve their balance by enhancing integration of vision, proprioception and vestibular system function under dual-task conditions regardless of the type of the task. These results are in line with the task integration hypothesis which states that dual-task training improves task coordination skills (Silsupadol et al., 2009 The results concerning improved postural and cognitive performance found by Jahanbakhsh et al. (2020), and Hocking et al. (2020) are consistent with our findings, providing further evidence confirming that dual-task group can be more effective than single-task group in enhancing balance and working memory in children. However, most studies in this area have used dual-task group for the elderly, patients, or adults, largely ignoring younger ages including children. Variations in postural control during different stages of life, for example during childhood, adolescence, adulthood, or old age may point to the fact that some motor learning techniques are more effective in some periods of life compared to other periods (Hytönen et al., 1993).
Another potential reason behind the improved motor and cognitive performance under dual-task settings can be attributed to the way in which the external focus of attention is employed during dual-task training (Wulf, 2013). Under dual-task conditions, due to simultaneously concentrating on primary and secondary tasks, individuals may automatically shift their attention to external cues. A large body of the existing studies confirms that using external focus of attention results in a greater improvement in motor, and even cognitive, learning and functions, including working memory, compared to cases where internal focus of attention is used (Razaghi et al., 2020;Wulf, 2013). In this regard, Wulf et al. (2013) put forth the constrained action hypothesis, arguing that, while using external attention, individuals improve their motor learning and performance through facilitating automatic processes in their motor systems while in cases where the external focus of attention is employed, learners often disrupt these automatic processes, undermining their motor performances. It seems that in the present study, children in the single-task settings used internal attention more frequently than those training in the dual-task settings, and therefore they experienced a greater drop in their motor and cognitive performance compared to the dual-task groups. It is important to note that the attentional patterns employed by the participants were not directly assessed in the present study and this represents a potential research limitation. Thus, it is recommended that future studies directly measure the attentional focus of participants during training based on single-and dual-task groups in an attempt to identify how these patterns may influence performance. Another limitation of the present study is the lack of a method for measuring and analyzing the dual-task paradigm. Therefore, it is suggested that future research consider this limitation.

Conclusion
Our findings suggest that under motor-motor dual-task conditions, children exhibited better dynamic balance performance compared to the performance observed in single-task cognitive-motor dual-task settings. However, when it comes to other variables like static balance performance and learning as well as working memory, motor-motor, and cognitive-motor dual-task groups outperform the single-task group, since in the present study the real control group (without any intervention) did not exist, it is suggested that future studies compare different dual-task methods along with the real control group (without any intervention). moreover, our results were different from the results performed in old adults (Norouzi et al., 2019). therefore, it is suggested that future studies compare the two motor-motor and motor-cognitive dual-task methods in all three age groups of children, adults and old adults.
In summary, our findings might have important implications for instructional settings, particularly in teaching children. It is recommended that coaches and therapists who work in this area try to integrate multiple motor and cognitive tasks within the dual-task framework into their instruction programs for learning motor skills in order to help learners enhance the learning of motor skills within a shorter time.