Working memory is a core executive function supporting dual-task locomotor performance across childhood and adolescence

Highlights

• Children prioritize walking over cognitive performance when challenged in a dual-task situation.

• Working memory capacity is a significant predictor of dual-task gait costs.

• Working memory moderated the relation between age and dual-task costs.

• Extends previous dual-process accounts by highlighting influence of working memory.

Abstract

Most daily-life ambulatory tasks involve dual tasking, for example, talking while walking. In children, the evidence supporting the effects of age on dual tasking is confounded by the difficulty of the cognitive task and lack of adjustment to suit individual cognitive abilities. To address this issue, the current study examined the effects of age, cognitive load, and executive functioning on the degree of dual-task gait interference across childhood and adolescence. We tested 120 typically developing children aged 6–11 years, adolescents aged 12–16 years, and young adults aged 18–25 years. Participants were asked to walk while performing a visuospatial working memory task at two levels of cognitive load (easy and difficult) adjusted to suit each participant’s cognitive ability. Spatiotemporal characteristics and intra-individual variability of gait were measured using a GAITRite electronic walkway. Irrespective of the cognitive load level, children aged 6 to 11 years showed greater dual-task gait interference for selective spatiotemporal gait characteristics; however, the younger children showed a trade-off pattern in gait variability whereby they prioritized gait stability at the expense of cognitive performance. Our results also showed that age and working memory capacity were significant predictors of dual-task interference for a range of complementary gait parameters in the combined sample. Importantly, working memory capacity was part of a moderating relationship between age and dual-task gait interference. These findings emphasize the importance of dual-task prioritization strategies in younger children and highlight the role of individual differences in working memory capacity in performance in dual-task gait situations.

Introduction

Many everyday tasks require the ability to prioritize our attention between cognitive and motor domains simultaneously. Examples include crossing the room while carrying an object and crossing the street while talking to a friend. Children experience pronounced decrements when engaging in two tasks at the same time, referred to as dual tasking or multitasking, even during seemingly automatic or well-practiced skills such as walking and maintaining balance (Kraan, Tan, & Cornish, 2017). It is generally assumed that children show greater difficulty in dual tasking than young adults because the regulation of motor tasks requires cognitive processes that continue to develop across childhood into adolescence (Hausdorff et al., 1999, Reilly et al., 2008, Reilly et al., 2008). However, in comparison with the literature on dual tasking in older adults, there is very little research on examining the interplay between cognitive and motor tasks during the period from childhood into adolescence in typically developing children.

There is a considerable body of evidence to support a substantial relationship between motor and cognitive skills across childhood (van der Fels et al., 2015), adolescence (Rigoli, Piek, Kane, & Oosterlaan, 2012), and older adulthood (Voelcker-Rehage, Godde, & Staudinger, 2011). A review of the experimental evidence has shown activation of the prefrontal cortex and cerebellum during both voluntary movement and cognitive tasks (Ito, 2008). While the cerebellum has been traditionally viewed as primarily involved in sensory-motor function, the prefrontal cortex is viewed as the controller of higher order cognition (Diamond, 2013). It has been proposed that both structures show a protracted developmental trajectory and form part of an integrated system that mediate the relation between cognitive and motor skills (Diamond, 2000). Given the extended period of changes in brain structures and cognitive functions underlying cognitive-motor couplings in children, it is possible that dual-task motor performance might proceed along a similarly prolonged period of development. However, there has been very little attention invested into age-dependent differences in the interaction between cognitive and motor domains across childhood and adolescence.

Dual-task paradigms that involve the simultaneous execution of a cognitive task while completing a motor task are a common approach to investigate the interrelationships between cognitive and motor domains (Woollacott & Shumway-Cook, 2002). These paradigms commonly involve concurrent performance of a secondary cognitive task (e.g., digit recall, spatial orientation) while completing a gait or postural control task (Huang & Mercer, 2001). There is accumulating evidence that gait is adversely affected by increased dual-task costs when typically developing children are asked to walk and perform secondary cognitive tasks in comparison with young adults (Schaefer, 2014). This decline in gait performance when walking with a concurrent cognitive task is typically reflected in worse or more variable spatiotemporal gait parameters, referred to as dual-task interference. However, dual-task interference can occur in the opposite direction when the gait or postural task is detrimental to cognitive task performance and reflected in increased errors or prolonged reaction time (see Reilly, van Donkelaar, et al., 2008, and Schaefer, Krampe, Lindenberger, & Baltes, 2008, for dual-task interference on a concurrent cognitive task in children). However, the extent to which children show dual-task interference in gait and/or cognitive performance and its relationship to demands on developing cognitive functions is, as yet, unclear in the literature.

Yet, the increased dependency on cognitive functions during walking in children has been taken to indicate that regulation of gait requires the prolonged development of executive functions, a set of complex mental operations that organize and regulate goal-directed behavior in novel and challenging situations (Diamond, 2013). It has been proposed that the core components of executive function include the ability to control attention to override a prepotent response (i.e., inhibition), the capacity to shift attention between tasks or mental sets (i.e., shifting), and the ability to update information in working memory (i.e., updating) (Miyake et al., 2000). Furthermore, working memory (i.e., the ability to store and manipulate information over brief periods of time) has been shown to exhibit the largest rates of change, and therefore the most maturation, between childhood and adolescence relative to other cognitive domains, including mental flexibility, episodic memory, and processing speed (Gur et al., 2012). Importantly, the maturation of performance is reached much later for working memory tasks with high executive demands (e.g., backward span tasks, self-ordered search tasks) compared with tasks with low executive demands (e.g., forward digit, spatial span) (Conklin et al., 2007, Luciana et al., 2005). What remains unknown is the extent of individual differences for this rate of change and the impact of working memory growth on the development of cognitive–motor dual-tasking performance across childhood into adolescence.

The extant studies examining age-related changes in the ability to coordinate cognitive and gait tasks in children have shown that dual-task costs are dependent on the type and level of difficulty of the secondary cognitive task (Hagmann-von Arx et al., 2015, Huang et al., 2003, Krampe et al., 2011, Schaefer et al., 2015, Schaefer et al., 2010, Schott and Klotzbier, 2018). For example, Huang et al. (2003) investigated the effect of concurrent visuospatial and verbal working memory tasks on locomotor control in 5- to 7-year-old typically developing children. Dual-task gait interference such as reduced step length was more pronounced when performing a concurrent visuospatial working memory task (visual identification of common objects) than when performing a verbal working memory task (digit span). Furthermore, a study by Schaefer et al. (2010) showed that 9-year-old children had greater dual-task variability in stride time when performing a concurrent visuospatial working memory task with high cognitive load using the n-back task (i.e., 3-back and 4-back) when compared with young adults, whereas both children and adults reduced their dual-task gait variability under low cognitive load (i.e., 1-back and 2-back) (see also Schaefer et al., 2015, for similar findings). Together, these studies are consistent with the importance of modality and cognitive load of the secondary cognitive task in revealing dual-task gait decrements in children.

Studies investigating age-dependent differences in dual-task gait interference have shown mixed findings, however, with some studies showing that effects of concurrent cognitive tasks on dual-task performance were more pronounced in younger typically developing children compared with young adults (Krampe et al., 2011, Schaefer et al., 2015, Schott and Klotzbier, 2018). In contrast, other studies have shown that dual-task decrements in younger children are either comparable to (Anderson, Bucks, Bayliss, & Della Sala, 2011), higher than (Chauvel et al., 2017), or lower than (Boonyong, Siu, van Donkelaar, Chou, & Woollacott, 2012) those in older children, adolescents, or young adults. However, it is important to consider that dual-task decrements are apparent for both secondary cognitive and gait tasks and that dual-task effects may vary as a function of age-related differences in trade-off patterns between motor and cognitive domains. Thus, an important consideration when interpreting findings from dual-task gait studies is the extent to which attention is prioritized to one task at the expense of the other.

Prominent theories of dual-task prioritization suggest that motor and cognitive capabilities, as well as individual factors, determine the extent to which one task is prioritized over another (Yogev-Seligmann, Hausdorff, & Giladi, 2012). According to the integrated dual-task prioritization model, individuals may adopt a “posture first” strategy, whereby the allocation of attention is prioritized toward postural stability in order to avoid hazards and risk of falling, or a “posture second” strategy, whereby attention is prioritized to the cognitive task at the expense of the motor task. In dual-task gait research, attention allocation has been commonly assessed by examining the extent to which performance on one task has improved but performance of the other task is unaffected or deteriorated relative to single-task performance. However, it has been suggested that dual-task prioritization during walking is not invariant but rather involves a range of factors, including postural reserve (i.e., the ability to respond effectively to a postural threat), compensatory capacity of both motor and cognitive domains, self-awareness for hazard estimation, expertise in skilled performance, and complexity of the secondary cognitive task (Yogev-Seligmann et al., 2012).

Although task prioritization between motor and cognitive domains has received more research attention in older adults (Li et al., 2018, Lövdén et al., 2008, Maclean et al., 2017), there have been relatively few studies examining trade-offs between these functional domains in the period from childhood to adolescence. The extant evidence suggests that school-aged children (9- to 11-year-olds) prioritize their postural control at the expense of decrements on cognitive dual tasks under high cognitive load when compared with young adults (Schaefer et al., 2008). This trade-off pattern of prioritizing postural control may relate to the higher ecological relevance of this domain for children (i.e., potential for falls or injury) alongside lower cognitive capacity limits, meaning that they reach their stability boundaries sooner than young adults.

Another prominent theory is the dual-process account of sensorimotor interactions that predicts a secondary task low in cognitive demand will shift attention away from the motor task and thus lead to improvements in balance control. According to this account, the stabilizing effects of an easy cognitive task can be attributed to a shift from an internal focus of attention to an external one that reduces reliance on top-down cognitive control and enables the system to smoothly carry out the automatized motor task (Huxhold, Li, Schmiedek, & Lindenberger, 2006). When the secondary cognitive task is high in cognitive demand, motor performance may decline through cross-domain competition, and the turning point at which this occurs is assumed to be related to an individual’s sensorimotor and cognitive resources. The evidence supporting the dual-process account in young and older adults is generally consistent with a U-shaped relationship between dual-task motor performance and different cognitive loads with age (Lövdén et al., 2008, Verrel et al., 2009). What remains unclear, however, is whether trade-off patterns and the influence of cognitive load on dual-task improvements or costs extend to gait performance across the period from childhood through adolescence.

In a systematic review, Ruffieux, Keller, Lauber, and Taube (2015) found only weak evidence to support the effects of age on dual-task gait interference across childhood into adolescence and discussed several methodological issues that should be taken into account when conducting dual-task gait research in children. These included the lack of studies that (a) compared dual-task performance of typically developing children with that of young adults, (b) assessed single-task performance of the concurrent task, and (c) examined dual-task effects for both the concurrent cognitive and motor tasks. Another systematic review by Saxena, Cinar, Majnemer, and Gagnon (2017) examined age-dependent differences in dual-tasking ability in children and concluded that there was a dearth of studies that equate one or both single tasks for ability level across age groups and a lack of consistency in calculating proportional dual-task costs.

More recently, a study by Saxena, Majnemer, Li, Beauchamp, and Gagnon (2019) found no age-related differences in dual-task costs when single-task difficulty was adjusted for level of difficulty to suit individual cognitive ability in 5- to 8-year-old typically developing children. The design included a concurrent auditory n-back task at varying cognitive loads during walking or when performing a trail making task. The authors attempted to overcome previous limitations by calculating proportional dual-task scores on each single component task. However, the restricted age range of the sample and lack of measures of spatiotemporal characteristics of gait were limitations in understanding the effect of age on dual-task gait performance during childhood in this study.

In the current study, we aimed to explore the relationship between visuospatial working memory and locomotor control and whether age-related variation in working memory capacity could account for changes in dual-task gait interference in typically developing school-aged children, adolescents, and young adults. Given the previous lack of research into the relationship between specific aspects of executive functioning and dual-task interference in children, we also aimed to establish the extent to which other subcomponents of executive function (e.g., inhibition, shifting, working memory) might contribute to dual-task gait effects and the extent to which they can account for age-related variation.

To this end, the change detection task (Luck & Vogel, 1997) was selected for the current study as a secondary cognitive task to provide further insight into interference between visuospatial working memory and locomotor control across childhood into adolescence (Isbell, Fukuda, Neville, & Vogel, 2015). This decision was based on evidence that visuospatial processing requiring working memory is more important for dual-task gait interference than verbal cognitive tasks in both children (Huang et al., 2003, Schaefer et al., 2010) and older adults (Menant et al., 2014, Qu, 2014).

In light of previous recommendations by Saxena et al. (2017), we used performance on the change detection task while seated to titrate the level of difficulty to ensure that task demands were customized for each participant’s cognitive capacity by manipulating the array and number of stimuli on each trial. The titration phase was then followed by the dual-task conditions, where baseline accuracy percentage for each participant on the change detection task was used to determine high or low cognitive load of the secondary cognitive task during walking trials. We designed the study to ensure that proportional dual-task costs were calculated for both single-task and dual-task performance on the cognitive task (accuracy on change detection task) and gait task (walking at preferred speed), allowing us to explore trade-off patterns between the cognitive and motor tasks in the dual-task situations.

Based on evidence that dual-task gait costs were more pronounced under high cognitive load in children when compared with young adults (Krampe et al., 2011, Schaefer et al., 2015, Schott and Klotzbier, 2018), and that children prioritize postural stability over cognitive task performance when attentional demands increase under dual-task conditions (Schaeffer et al., 2008), we hypothesized that primary school-aged children would show greater costs in the cognitive component of the dual task than in the gait component of the dual task with increasing attentional demands relative to adolescents and young adults. Given that we equated difficulty level of the cognitive task according to participants’ ability level, we also predicted that selective subcomponents of executive function would be associated with dual-task gait interference across the combined sample; however, no predictions were made about specific subcomponents that might relate to dual-task costs or moderate the relationship between age and dual-task interference due to a lack of previous research.