Secondary Logo

Effects of Different Concurrent Cognitive Tasks on Temporal-Distance Gait Variables in Children

Huang, Hsiang-Ju MS, PT; Mercer, Vicki Stemmons PhD, PT; Thorpe, Deborah Ellen PhD, PT, PCS

doi: 10.1097/01.PEP.0000067886.96352.6B

Purpose The study was designed to investigate the influence of three different concurrent cognitive tasks on gait characteristics in five- to seven-year-old children.

Methods Twenty-seven subjects, 16 boys and 11 girls, between the ages of 5.0 and 7.8 years (M = 6.4 ± 0.8 years) participated in the study. A dual-task paradigm was used. Each subject performed the following single and dual walking tasks: walking alone, walking while identifying pictures of common objects (visual identification), walking while identifying sounds (auditory identification), and walking while retaining a series of numbers in memory (memorization). Gait speed, cadence, and step length were compared among the four gait tasks using repeated-measures analysis of covariance.

Results Gait speed was lower under all dual-task conditions compared to the single-task walking condition. Both cadence and step length decreased with concurrent performance of the visual and auditory identification tasks, while the memorization task affected only cadence. The interference effects of the cognitive tasks on gait were largest for the auditory identification task and smallest for the memorization task.

Conclusion These results indicate that children may have difficulty maintaining motor performance while simultaneously processing cognitive information.

This study of the influence of three different concurrent cognitive tasks on gait characteristics in five- to seven-year-old children indicate that they may have difficulty maintaining motor performance while simultaneously processing cognitive information.

Division of Physical Therapy, University of North Carolina at Chapel Hill, Chapel Hill, NC

Address correspondence to:Vicki Stemmons Mercer, PhD, PT, Division of Physical Therapy, CB#7135, Medical School Wing E, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7135.

Back to Top | Article Outline


Cognitive processing plays an important role in motor performance. 1 A series of cognitive processes are proposed to be involved in executing a motor skill, including stimulus identification, response selection, and schema retrieval from long-term memory. 1,2 Attention, described as the capacity or resources for processing information, 2 is one cognitive factor that is thought to affect motor performance. 3–6 Many situations in our everyday lives involve simultaneous performance of cognitive and motor tasks, and therefore require us to divide attention between two or more tasks. Researchers have used a dual-task paradigm to investigate the attentional demands of motor tasks 7–9 and the effects of concurrent cognitive or motor tasks on motor performance. 10–13

The work of Lajoie et al 8 is an example of a study in which the attentional demands of performing motor skills were of interest. Subjects performed an auditory reaction time task as the secondary task in combination with the primary tasks of sitting, standing, and walking. Under dual-task conditions, the attentional demands of the primary tasks were inferred from changes in secondary task performance when compared to single-task conditions. The greatest increase in auditory reaction times occurred during the walking task. Therefore, the researchers concluded that walking has higher attentional demands than maintaining upright standing or sitting.

The effects of concurrent tasks on motor performance may be of even greater relevance to physical therapists than the attentional demands associated with performance of particular motor skills. 14 These effects provide insights into the changes in performance that may be expected when an individual is required to do two things at once. Under these circumstances, attention is divided or shared between the primary motor task and an additional cognitive or motor task. When motor skills become highly learned or automatic, concurrent tasks are expected to produce little interference, and to have little effect on motor performance. 11,15 For example, expert soccer players slowed less in running than novice players when they needed to dribble a ball and identify shapes simultaneously. 11 By progressively adding secondary tasks to the primary task and observing the effects on performance, one can test the automaticity of a motor skill. 2,11 This type of information may be useful to therapists, teachers, or coaches in determining the number and types of tasks the learner can perform concurrently with a primary motor task without significant interference.

Although dual-task methodology has been used in different ways to investigate the relationship between attention and motor performance, most previous research has been focused on adults. Little research has been conducted to examine how attention influences motor performance in children. Most previous dual-task studies involving children examined developmental changes in certain cognitive skills, such as memorization and reasoning. 16–19 The extent to which different concurrent cognitive tasks interfere with children’s performance of more complex motor activities, such as walking and maintaining balance, is unclear.

Although walking is a well-practiced motor skill for young adults, results of previous research suggest that walking involves significant attentional demands in this age group. 8 The processing involved in performing a cognitive task while walking may influence walking patterns in children who are developing typically. Furthermore, different cognitive tasks may produce different amounts of interference. Whitall 12 investigated the effects of concurrent cognitive tasks on two locomotor skills, running and galloping, in subjects in five age groups (2.5- to 3.5-year-olds, 3.5- to 4.5-year-olds, six- to seven-year-olds, nine- to 10-year-olds, and 18- to 34-year-olds). The results showed that speeds of running and galloping were affected by simultaneously performed memorization or singing tasks, but interlimb coordination was not. The memorization task had a greater effect than singing on these two locomotion skills. Changes in gait speed associated with the memorization task tended to result from changes in step length, while the singing task tended to produce changes in both step time and step length. Interference effects decreased with age, so that nine- to 10-year-olds appeared to time-share the locomotor and cognitive tasks in a similar fashion to adults. Whitall 12 concluded that operation of the variables of step time and step length involves attentional resources and is both age- and task-related.

Children frequently encounter situations involving simultaneous performance of cognitive tasks while walking. For example, children may need to identify traffic signals or respond to verbal instructions while walking. If teachers and therapists have information about how concurrent cognitive tasks influence walking performance, they may be able to structure the environment and their instructions in a manner that is more consistent with a child’s abilities. Evaluation of a child’s performance under dual-task conditions may reveal subtle deficits that would otherwise go unnoticed. In addition, an understanding of the effects of different cognitive tasks may assist with development of guidelines for the progression of motor learning activities.

The purpose of this study was to investigate the influence of three different concurrent cognitive tasks on gait in five- to seven-year-old children who are developing typically. This age range was selected to insure that the children would be old enough to follow instructions in the study protocol but young enough to demonstrate interference effects distinct from those of adults. 12 A dual-task paradigm was used, with comparison of gait characteristics under single- and dual-task conditions. The cognitive tasks involved identification and processing of visual and verbal stimuli similar to those that children encounter during their daily activities. The three cognitive tasks were 1) visual identification, 2) auditory identification, and 3) memorization.

Three specific research questions were asked. 1) Do walking speed, cadence, and step length differ between single-task walking and a dual-task condition, in which walking is combined with performance of a cognitive task? 2) Are there differences in the amount of change in gait speed, cadence, or step length associated with different concurrent cognitive tasks? 3) Does cognitive task performance change from single- to dual-task conditions?

On the basis of previous research 11,12 and the results of pilot testing with three children, we expected that subjects would demonstrate decreases in gait speed, cadence, and step length under dual-task conditions. Interference effects were expected to be largest for the visual identification task and smallest for the memorization task. Expectations concerning differential effects were based on previous research indicating the importance of visual information for locomotor control 20,21 and on the possibility that children might attempt to increase their walking speed during the memorization task to minimize the time interval between stimulus presentation and verbal recall. On the basis of the results of pilot testing, performance of the cognitive tasks, in terms of the percent correct, was not expected to change significantly from single- to dual-task conditions.

Back to Top | Article Outline



Twenty-seven five- to seven-year-old children who were developing typically (16 boys and 11 girls) participated in the study. The children’s mean age was 6.4 ± 0.8 years (range 5.0–7.8 years). The subjects were recruited through day-care centers and elementary schools in central North Carolina, and through contacts with faculty members and graduate students at the University of North Carolina at Chapel Hill (UNC-CH). The inclusion criteria, determined by parental report, were 1) full-term birth (>38 weeks); 2) English speaking; 3) normal or corrected normal vision and hearing; 4) no diagnosis of orthopedic, neurological, or cardiovascular disorders; 5) no diagnosed learning disabilities; 6) no diagnosed mental retardation; 7) began ambulating independently by 16 months of age; 8) no diagnosed speech or language disorders; and 9) no diagnosed attention or behavioral disorders.

Parents who were interested in having their child participate in the study were contacted by telephone for an initial screening interview. The children who met the inclusion criteria were brought to the Center for Human Movement Science at UNC-CH for testing. The Peabody Picture Vocabulary Test (PPVT) 22 was used as a measure of vocabulary acquisition. This test screened for the ability to understand instructions and to identify the pictures included in the visual identification task. The subjects’ standard scores on the PPVT were all within one standard deviation of the mean for their age, with a mean standard score of 113.0 ± 13.5 for the group as a whole.

Back to Top | Article Outline


The experimental set-up is diagrammed in Figure 1. A walkway measuring 30 feet long by five feet wide was delineated by blue tape on the floor of the laboratory. Yellow cardboard footprints were taped to the floor five feet in front of the start of the walkway to allow the children adequate distance to accelerate and achieve their typical gait patterns before reaching the walkway.

Fig. 1

Fig. 1

A video camcorder was used to record the walking performance of each subject. The camcorder was placed eight feet, seven inches above the ground at one corner of the laboratory. The placement of the camcorder was adjusted to allow viewing of the entire length of the walkway, so that the number of steps taken by each child when walking from one end of the walkway to the other could be determined from the videotape. Infrared sensors connected to a millisecond timer were placed at either end of the walkway for gait speed measurement. Light emitting diodes (LEDs) were connected to the sensors at both ends of the walkway to allow the investigator to determine from videotape which steps to include in each trial. When the sensors were activated as the subject reached the start and the end of the walkway, the respective LEDs were illuminated in the field of view of the video camcorder.

Back to Top | Article Outline

Experimental Task

The children were asked to perform three cognitive tasks under single-task conditions and in combination with walking (dual task). A small microphone connected to a microcassette recorder was attached to each subject’s clothing before data collection for recording of the subject’s verbal responses on the cognitive tasks during dual-task conditions. This recording allowed for later verification of any responses that were not clearly heard by the principal investigator during testing.

One cognitive task was a visual identification task, in which subjects were asked to identify pictures of common objects or toys adapted from Photo Cue Cards 23 and picture-vocabulary books for the first grader. 24,25 Fifteen pictures were selected as representative of objects that children commonly see in their daily lives, such as a dog and a dress (Table 1). The vocabulary words represented by these pictures all contained one or two syllables. The pictures were saved on a laptop computer and projected on a 20-inch color video monitor at the far end of the walkway.



A second cognitive task was auditory identification. The children were asked to identify environmental sounds adapted from sound effect compact discs in the Teleconference Studio at UNC-CH. All of the sounds were less than two seconds in duration and were representative of sounds commonly experienced by children, such as a car horn and a doorbell (Table 1). The auditory identification task included a total of 15 sounds. These sounds were saved on the laptop computer and projected through two speakers at the far end of the walkway.

The third cognitive task was memorization. Subjects were asked to recall a series of random numbers. The random numbers for the task in the single-task condition were adapted from the number recall subtest of the Kaufman Assessment Battery for Children. 26 Forty-eight numbers were generated from a table of random numbers for use in the dual-task condition. Before the study, all numbers for the memorization task were recorded on audiotape by a male speaker at a rate of one number every second. A microcassette recorder was used to play the audiotape for the subjects at the start of single- and dual-task trials of the memorization task.

Back to Top | Article Outline


Before the test, the parent was asked to complete a second, more detailed screening questionnaire and to sign an informed consent form approved by the Committee on the Protection of the Rights of Human Subjects at UNC-CH. The principal investigator measured and recorded the child’s height and weight using a standard physician’s scale and administered the PPVT. 22 The child then was asked to perform two practice trials of walking as fast as possible from one end of the walkway to the other. The average time required for these two walking trials was used to determine the delay between presentation of numbers and request for recall during single-task testing of the memorization task (see below for description of procedures for the memorization task).

During testing, the subjects wore comfortable clothing and their customary footwear. Distractions in the room were minimized as much as possible. The room had no windows, and the walls, carpeting, and room dividers were all muted shades of gray, brown, or black. The principal investigator, the child’s parent, and a research assistant were present with the child during testing. The research assistant who controlled presentation of the visual and auditory stimuli by means of the laptop computer was screened from the subject’s view (Fig. 1). During single-task testing of cognitive tasks, subjects were asked to sit in a chair without arms at the beginning of the walkway. They performed the three cognitive tasks, visual identification, auditory identification, and memorization, in random order. Five trials of each visual and auditory task were conducted, including one practice trial. For each visual identification trial, three pictures were presented at a rate of one per second on a video monitor at the end of the walkway. For each auditory identification trial, three sounds were presented at a rate of one every two seconds through speakers at the end of the walkway. The presentation rates for the pictures and sounds were based on the results of pilot testing. These rates allowed for presentation of three stimuli during each walking trial for all three pilot subjects. The presentation rate for the sounds was lower than that for the pictures because the pilot subjects had difficulty identifying sounds that were less than one second in duration. Different presentation rates for the visual and auditory stimuli were necessary, therefore, to provide for an equal number of stimuli per trial and a similar level of difficulty for the two secondary tasks. Subjects were instructed to identify as quickly as possible every picture and sound presented to them by saying the name of the picture or sound out loud. The number of pictures and sounds that each child correctly identified was recorded for each trial.

For the memorization task, subjects were asked to remember the numbers presented to them on audiotape from the microcassette recorder held by the principal investigator. Subjects were asked to repeat the numbers in exactly the same order after a short time delay. The delay between presentation of the numbers and request for recall was based on the average amount of time (to the nearest second) required for that subject to walk from one end of the walkway to the other during the practice walking trials. On the first two trials of the memorization task, the principal investigator presented two digits between one and 10 and recorded the child’s response after the time delay. The number of digits presented was then increased by one digit every other trial, to a maximum of eight digits. Testing on the memorization task ended when the child failed to achieve 100% accuracy (all digits in the correct order) on at least one of the two trials. The child’s single-task performance was operationally defined as the maximum number of digits that the child was able to recall with 100% accuracy on at least one of the two trials.

After a rest period of approximately five minutes, subjects performed the following gait tasks in random order: walking alone (Walk), walking with visual identification (Walk+V), walking with auditory identification (Walk+A), and walking with memorization (Walk+M). Each gait task included one practice trial and four test trials. One additional trial was conducted when the subject fell, failed to reach the end of the walkway, failed to identify two of the three pictures (Walk+V) or sounds (Walk+A) presented, failed to identify >50% of the digits presented (Walk+M), or experienced an unusual distraction, such as someone opening the door to the room. The criteria for acceptable secondary task performance were selected to ensure that the subjects were attending to the secondary task. Walking performance during all tasks was videotaped for later analysis. Only two subjects, ages 5.6 and 5.8 years, fell during testing. Both falls occurred when the subjects appeared to try to walk extremely fast, without running, during the single-task walking (Walk) condition.

During all gait tasks, subjects were asked to stand on the yellow footprints that were five feet in front of the start of the walkway. They were instructed to walk as fast as they could toward the wall that was approximately nine feet, six inches past the end of the walkway. Instructions were designed to encourage subjects to give equal priority to gait and cognitive tasks under dual-task conditions. In Walk+V and Walk+A, children were asked to name the pictures or sounds presented to them on each trial as fast as possible while walking as fast as possible. The same 15 pictures and sounds that were presented in the single-task visual and auditory identification conditions were used in the dual-task conditions, but in different sequences (Table 1). The presentation rates and number of stimuli for each trial for the pictures and sounds were the same as those under single-task conditions. In Walk+M, subjects listened to the series of numbers before they started walking and then were asked to recall the numbers after they reached the end of the walkway and stopped walking. The number of digits presented to each subject at the start of Walk+M trials was based on his/her performance under single-task conditions (maximum number of digits subject was able to recall with 100% accuracy on at least one trial). The principal investigator recorded walking time for all four gait tasks, as well as the number of correctly identified pictures (Walk+V), the number of correctly identified sounds (WALK+A), and the number of digits correctly recalled (Walk+M).

Back to Top | Article Outline

Data Analysis

Four trials were analyzed under each condition. For situations in which the subject showed poor secondary task performance (failed to identify at least two pictures/sounds or failed to recall at least 50% of the digits presented) on more than one of the test trials, selection of trials for analysis was based on gait speed, with the faster trial(s) included in the analysis. For two subjects, only three trials were analyzed for the Walk+A condition. A technical problem occurred during data collection for one of these subjects, and the other subject stopped walking before reaching the end of the walkway on two trials.

Gait speed was calculated (in m/s) by dividing the distance walked (30 feet = 9.15 m) by the time in seconds. Cadence was calculated as the number of steps divided by the time in minutes. Step length was calculated as the distance walked (9.15 m) divided by the number of steps. The number of steps in each trial for each subject was determined from the videotape, with the steps on which the two LEDs were first illuminated included in the count. To examine intrarater reliability in counting the number of steps per trial, a sample of 64 videotaped trials was reanalyzed by the principal investigator approximately nine weeks af-ter the initial analysis. Exact agreement was obtained for 62 of the 64 trials, with the two trials that were not in agree-ment differing by only one step per trial.

The percentage of correct responses for the visual identification and auditory identification tasks was calculated as the number of correct responses divided by the total number of stimuli for the four trials (12 stimuli). The percentage of correct responses for the memorization task of each trial was obtained by dividing the number of digits that were recalled in the correct order by the number of digits presented. The average percentage of correct responses on the four trials of the memorization task was entered into the analysis of secondary task performance.

Descriptive statistics of mean, standard deviation, and range were calculated for each of the gait parameters under every walking condition and for percentage of correct responses on the cognitive tasks. The level of significance for all statistical comparisons was set at 0.05. The means of the four trials for each subject for each condition were entered into a spreadsheet for analysis using repeated measures analysis of covariance (ANCOVA) with age as the covariate. To address the first two research questions concerning differences in temporal-distance measures among the tasks, separate ANCOVAs were performed for each of the dependent variables of gait speed, cadence, and step length. The independent variable for these analyses was gait condition (a repeated measures factor with four levels: Walk, Walk+V, Walk+A, and Walk+M). The a priori alpha level of 0.05 was adjusted according to the method described by Legendre and Legendre. 27 This method is an alternative to the Bonferroni correction, which is viewed as overly conservative in a study such as this. With the use of this method, the p values resulting from the analyses were first ordered from smallest to largest, and then adjusted p values (p ′) were computed according to the formula p ′ = (ki + 1) p, where k is the number of p values and i is the rank of a p value. Working from the largest original p value, each adjusted p value was compared to the one below it. If the adjusted p value below was larger, it was made equal to the one above. In the final step, the adjusted p values were compared to the a priori alpha level to determine significance. Significant F values for the repeated measures ANCOVAs were followed by paired t tests, using identical procedures to adjust the alpha level, to identify significant differences among the gait tasks. Three additional paired t tests, again with adjustment of the alpha level as described above, were performed to address the third research question concerning differences in secondary task performance between single- and dual-task conditions.

Back to Top | Article Outline


Descriptive statistics for the temporal-distance measures of gait speed, cadence, and step length are reported in Table 2. Within-subject variability was low, as indicated by the coefficient of variation of the four trials for each subject under each gait condition. The mean coefficients of variation were 0.053, 0.065, 0.079, and 0.065 for Walk, Walk+V, Walk+A, and Walk+M, respectively. Results of the three repeated-measures ANCOVAs indicated differences among the four gait tasks for speed (Wilks’ lambda = 0.667; F = 3.663;df =3, 23;p = 0.027;p ′ = 0.042), cadence (Wilks’ lambda = 0.705; F = 3.208;df = 3, 23;p = 0.042;p ′ = 0.042), and step length (Wilks’ lambda = 0.629; F = 4.514;df = 3, 23;P = 0.012;p ′ = 0.036). Paired t tests indicated that, consistent with our expectations, gait speed was significantly lower under all dual-task conditions compared to the single-task walking condition. Cadence was also significantly lower for all dual-task conditions compared to single-task walking, but step length was significantly shorter than for single-task walking during Walk+V and Walk+A only. Our expectation that interference effects would be largest for the visual identification task and smallest for the memorization task was only partially supported. The effects were largest for the auditory identification task and smallest for the memorization task. Among the three dual-task conditions, speed and cadence were lower during Walk+A than during Walk+V or Walk+M, with the latter two conditions not different from each other on either gait measure. Step lengths differed among all three dual-task conditions, with the smallest values for Walk+A, the largest for Walk+M, and intermediate values for Walk+V.



Secondary task performance differed between single- and dual-task conditions for both the auditory identification (t = −2.85;df =26;p = 0.008;p ′ = 0.016), and memorization tasks (t = 6.93;df = 26;p = 0.000;p ′ = 0.000) (Fig. 2). Therefore, our expectation that secondary task performance would not change from single- to dual-task conditions was not supported. For the auditory identification task, subjects demonstrated an increase in percentage of correct responses from 74.0% at baseline to 82.3% under dual-task conditions. For the memorization task, the percentage of correct responses fell from 100% at baseline to 73.3% under dual-task conditions. The 100% accuracy for single-task performance resulted from the study design, in which each child’s single-task performance was defined as the maximum number of digits that the child was able to recall with 100% accuracy on at least one trial. Performance of the visual identification task did not change from single-task (88.3% correct) to dual-task conditions (91.4% correct).

Fig. 2

Fig. 2

Thirteen of the 27 subjects did not respond immediately after presentation of the visual or auditory stimuli during single- or dual-task conditions or both. In these instances, subjects reported the stimuli they had seen or heard during the trial en masse, rather than reporting each stimulus after its presentation. Delayed responses occurred during 15 trials in single-task auditory identification, 14 trials in dual-task auditory identification, and two trials in single-task visual identification. Among those children who demonstrated delayed responses, eight were five years old, two were six years old, and three were seven years old.

Back to Top | Article Outline


The children in this study demonstrated slower gait speeds during simultaneous performance of three different cognitive tasks than during a single-task walking condition. These findings are in accord with previous research results indicating interference effects of concurrent cognitive tasks on locomotor performance in children. 11,12 While previous research with children focused on running and galloping tasks, 11,12 results of the present study reveal interference effects for a walking task. Secondary cognitive tasks, therefore, can affect the performance of a very basic and highly practiced locomotor skill in children who are developing typically.

Changes in gait speed as a function of changes in cadence and step length differed for the different cognitive tasks. When performing visual and auditory identification tasks concurrently with walking, children slowed their walking speed by decreasing both cadence and step length. During performance of a concurrent memorization task, decreases in walking speed were primarily associated with decreased cadence. Interference effects for cadence but not step length during the memorization task might have resulted from mental rehearsal during walking. Although subjects were not instructed to rehearse the digits to be recalled, they might have done so spontaneously. Audible or silent recitation might have imparted a slower rhythm to the children’s gait patterns.

The magnitude of the decrease in walking speeds resulting from the addition of secondary cognitive tasks to single-task walking ranged from 0.18 m/s for the memorization task to 0.43 m/s for the auditory identification task. These values are lower than the mean difference of 0.68 m/s between ordinary and fast walking speeds predicted by Norlin et al 28 for children in the age range corresponding to our sample. These results suggest that our subjects were able to maintain a relatively fast walking speed while simultaneously executing a cognitive task. The children apparently did not revert to a more comfortable, “ordinary” walking speed under these conditions. The interference effects were substantial, however, considering that the expected increase in fast walking speed is 0.24 m/s as children mature from 5.0 to 7.8 years of age. 28

If interference effects of this magnitude occur in children with typical development, then we might expect even larger effects in children with disabilities. Although results of our study cannot be directly applied to children with disabilities, this information can be useful for pediatric physical therapists in several ways. As noted in our review of dual-task methodology, 29 increased understanding of the attentional demands of different cognitive and motor tasks should enable physical therapists to make more informed decisions about how they structure evaluation and intervention activities. By adding concurrent cognitive or motor tasks during clinical examinations and measuring changes in simple temporal or distance variables, therapists may be able to determine children’s abilities to perform motor tasks “automatically” or, at least, to divide attention between tasks.

Therapists also can be more cognizant of the demands they are placing on children when they provide instructions, demonstrations, or simply conversation that must be processed at the same time as on-going motor performance. Results of the present study suggest that, in children with typical development and normal vision and hearing, interference effects on walking are larger for tasks requiring processing of concurrent auditory as compared to visual information. Tasks like our memorization task that do not require continuous processing of new afferent information and generation of immediate responses, produce the least interference.

The finding that young children have difficulty maintaining motor performance while processing concurrent information has implications for therapists’ use of instructions, feedback, and secondary tasks. If a high level of motor performance is desired, children may benefit more from instructions or feedback given before or after attempting the task. On the other hand, if the goal is to further refine a relatively skilled motor performance, then providing concurrent feedback or imposing a secondary cognitive task may be beneficial. 2 If information is given during the execution of a motor task, children’s motor performance may be less affected by visual than verbal information. However, because the secondary tasks in the present study were identification tasks rather than tasks involving processing of verbal and visual feedback information, further investigation is needed to support this application of our results.

The results of the present study are in contrast to those presented by Whitall 12 with respect to interference effects on cadence and step length. In Whitall’s study, subjects in all age groups tended to decrease their step lengths but to maintain constant step times when simultaneously executing a memorization task and running or galloping. The discrepant findings may be attributed to differences in the primary motor and secondary cognitive tasks used in the studies. Although fast walking and running may be similar skills, children may modify their locomotor performance differently for these two skills under conditions of divided attention. In addition, the memorization task in Whitall’s study involved concurrent presentation of letters during running and galloping tasks, while the numbers to be recalled in the present study were presented before the subjects started walking. Whereas Whitall presented a specified number of letters for recall for each age group, we were able to modify the number of digits presented during the memorization task in accordance with each individual’s single-task performance.

Our results are consistent with those reported by Whitall, 12 however, with regard to greater interference effects on locomotor performance for tasks requiring concurrent processing of sensory information. In our study, both the visual identification and the auditory identification tasks involved processing of sensory stimuli during walking, whereas Whitall’s memorization task required concurrent auditory processing of letters spoken by the experimenter while the subject ran or galloped along the runway. The other cognitive task used by Whitall involved asking subjects to sing a familiar song while running or galloping. This task, like our memorization task, did not require processing of visual or auditory stimuli during locomotion. In both our study and Whitall’s, larger decreases in locomotor speed were found for secondary tasks involving concurrent processing of sensory stimuli than for those involving other types of cognitive processing.

Although the auditory identification and memorization tasks do not appear to involve the types of cognitive processing necessary for control of walking, these tasks interfered with walking performance in the present study. Previous researchers indicated that postural control interfered particularly with spatial memory tasks. 3,30 Although interference effects in adults are only observed for memory tasks involving the same domain of working memory, 31,32 this may not hold true for children. Researchers have suggested that, compared with older children, children under the age of eight years demonstrate more interference effects that are not specific to the domain of working memory involved in performance of the primary task. 33 Such nonspecific interference effects may at least partially account for the observed auditory identification and memorization task results.

In contrast to our expectations, the auditory identification task, not the visual identification task, produced the greatest interference effects on gait. Perhaps auditory stimuli require more attention because they are transient or time locked 34 and less concrete and familiar to children than visual stimuli. Previous research results suggest that young children rely more on and allocate more attention to visual information than other modalities when performing perceptual and memory tasks. 35,36 Young children also do not perform as well as older children on auditory monitoring tasks, but may match the performance of older children on visual monitoring tasks. 37 In the present study, the incidence of delayed responses was much higher for auditory stimuli (29 trials) than for visual stimuli (two trials). The delayed responses also occurred more frequently in the five-year-old children than the six- or seven-year-old children. These results are consistent with Tallal’s 38 suggestion that the ability to respond to rapidly presented nonverbal auditory stimuli improves with age.

Changes in secondary task performance from single- to dual-task conditions provide additional information about interference effects. Subjects performed the visual and auditory identification tasks as well or better under dual-task as under single-task conditions. Because these results indicate that the subjects were attending to the cognitive tasks, the observed changes in gait parameters may be attributed to the need for divided attention. For the auditory identification task, subjects demonstrated improved performance under dual-task compared with single-task conditions. These improvements may be related to the subjects’ increasing familiarity with the auditory stimuli. Because the same sounds were presented in a different order under dual-task conditions, the children may have been able to respond faster and to recognize more of the sounds on the second presentation. The decline in performance on the memorization task from single- to dual-task conditions suggests that subjects may have used a different strategy for this task than for the visual and auditory identification tasks. In the absence of a need to process and immediately respond to stimuli presented during the walking task, subjects may have chosen a strategy that optimized gait performance at the expense of recall performance.

One of the limitations of this study is that response time was not considered in evaluating secondary task performance. The delayed responses exhibited by some children may have reflected difficulty in attending to secondary cognitive tasks. In addition, the sequence of pictures or sounds might have influenced performance. For example, if the first sound presented on a trial was difficult or unfamiliar to a child and therefore required additional processing time, the child might have missed the two subsequent stimuli. We were not able to insure equal levels of difficulty for the sequences of stimuli used under single- and dual-task conditions. Careful selection and measurement of secondary tasks should enhance future research efforts aimed at identifying the effects of divided attention on children’s motor performance.

Back to Top | Article Outline


Children in the age range of five to seven years decrease their gait speeds while performing concurrent cognitive tasks. The amount and pattern of interference varies with different cognitive tasks. Visual and auditory identification tasks affect both cadence and step length, whereas a memorization task affects only cadence. The auditory identification task produces the largest decreases in the temporal-distance gait measures, whereas the memorization task produces the smallest.

The finding that children’s gait performance declines with the addition of a secondary cognitive task has implications for those involved in teaching motor skills. The results of this study indicate that a progression of secondary cognitive tasks from least to most difficult would include the use of memorization, visual identification, and auditory identification tasks, in that order. Further research is needed to determine the effects of concurrent cognitive tasks on performance of various motor tasks in children of various ages who are developing typically and in children with disabilities.

Back to Top | Article Outline


1. Mulder T. A process-oriented model of human motor behavior: toward a theory-based rehabilitation approach. Phys Ther. 1991; 71: 157–164.
2. Schmidt RA. Motor Control and Learning: a Behavioral Emphasis, 2nd ed. Champaign, Ill: Human Kinetics Publishers; 1988.
3. Camicioli R, Howieson D, Lehman S. Talking while walking: the effect of a dual task in aging and Alzheimer’s disease. Neurology. 1997; 48: 955–958.
4. Persad CC, Giordani B, Chen H, et al. Neuropsychological predictors of complex obstacle avoidance in healthy older adults. J Gerontol. 1995; 50B: P272–P277.
5. Shumway-Cook A, Woollacott MH. Motor Control: Theory and Practical Applications. Baltimore, Md: Williams & Wilkins; 1995: 216–217.
6. Shumway-Cook A, Wollacott M, Kerns KA, et al. The effects of two types of cognitive tasks on postural stability in older adults with and without a history of fall. J Gerontol. 1997; 52A: M232–M240.
7. Bardy BG, Laurent M. Visual cues and attention demand in locomotor positioning. Percept Mot Skills. 1991; 72: 915–926.
8. Lajoie Y, Teasdale N, Bard C, et al. Attentional demands for static and dynamic equilibrium. Exp Brain Res. 1993; 97: 139–144.
9. Wright DL, Kemp TL. The dual-task methodology and assessing the attentional demands of ambulation with walking devices. Phys Ther. 1992; 72: 306–315.
10. Ebersbach G, Dimitrijevic MR, Poewe W. Influence of concurrent tasks on gait: a dual-task approach. Percept Mot Skills. 1995; 81: 107–113.
11. Smith MD, Chamberlin CJ. Effect of adding cognitively demanding tasks on soccer skill performance. Percept Mot Skills. 1992; 75: 955–961.
12. Whitall J. The developmental effect of concurrent cognitive and locomotor skills: time-sharing from a dynamic perspective. J Exp Child Psychol. 1991; 51: 245–266.
13. Yap RL, van der Leij A. Testing the automatization deficit hypothesis of dyslexia via a dual-task paradigm. J Learn Disabil. 1994; 27: 660–665.
14. Huang HJ, Mercer VS. Dual-task methodology: applications in studies of cognitive and motor performance in adults and children. Pediatr Phys Ther. 2001; 13: 133–140.
15. Geurts ACH, Mulder TW, Nienhuis B, et al. Dual-task assessment of reorganization of postural control in persons with lower limb amputation. Arch Phys Med Rehabil. 1991; 72: 1059–1064.
16. Bjorklund DF, Harnishfeger KK. Developmental differences in the mental effort requirements for the use of an organizational strategy in free recall. J Exp Child Psychol. 1987; 44: 109–125.
17. Guttentag RE. The mental effort requirement of cumulative rehearsal: a developmental study. J Exp Child Psychol. 1984; 37: 92–106.
18. Guttentag RE. Age differences in dual-task performance: procedures, assumptions, and results. Dev Rev. 1989; 9: 146–170.
19. Halford GS, Maybery MT, Bain JD. Capacity limitation in children’s reasoning: a dual-task approach. Child Dev. 1986; 57: 616–627.
20. Assaiante C, Amblard B, Carblanc A. Peripheral vision and dynamic equilibrium control in five to twelve year old children. In: Amblard B, Berthoz A, Clarac F, eds. Posture and Gait: Development, Adaptation, and Modulation. New York: Excerpta Medica; 1988: 75–82.
21. Lackner JR, DiZio P. Visual stimulation affects the perception of voluntary leg movements during walking. Perception. 1988; 17: 71–80.
22. Dunn LM, Dunn LM. Peabody Picture Vocabulary Test, 3rd ed. Circle Pines, Minn: American Guidance Service; 1997.
23. Kerr JYK. Photo Cue Cards: 300 Meaningful Pictures for Oral Language Practice. Tucson, Ariz: Communication Skill Builders; 1979.
24. MacKinnon D. My World of Spanish Words. Haupauge, NY: Barron’s Educational Services; 1995.
25. Wilkes A. My First Word Book. New York: Dorling Kindersley; 1991.
26. Kaufman AS, Kauman NL. K-ABC: Kaufman Assessement Battery for Children.Administration and Scoring Manual. Circle Pines, Minn: American Guidance Service; 1983.
27. Legendre P, Legendre L. Numerical Ecology. Amsterdam; New York: Elsevier; 1998.
28. Norlin R, Odenrick P, Sandlund B. Development of gait in the normal child. J Pediatr Orthop. 1981; 1: 261–266.
29. Huang HJ, Mercer VS. Dual-task methodology: applications in studies of cognitive and motor performance in adults and children. Pediatr Phys Ther. 2001; 13: 133–140.
30. Maylor EA, Wing AM. Age differences in postural stability are increased by additional cognitive demands. J Gerontol. 1996; 51B: P143–P154.
31. Fry AF, Hale S. Processing speed, working memory, and fluid intelligence: evidence for developmental cascade. Psychol Sci. 1996; 7: 237–241.
32. Hale S, Myerson J, Rhee SH, Weiss CS, Abrams RA. Selective interference with the maintenance of location information in working memory. Neuropsychology. 1996; 10: 228–240.
33. Hale S, Bronik MD, Fry AF. Verbal and spatial working memory in school-age children: developmental differences in susceptibility to interference. Dev Psychol. 1997; 33: 364–371.
34. Gillet P. Auditory Processes. Novato, Calif: Academic Therapy Publications; 1993.
35. Guttentag RE. A developmental study of attention to auditory and visual signals. J Exp Child Psychol. 1985; 39: 546–561.
36. Hitch GJ, Woodin ME, Baker S. Visual and phonological components of working memory in children. Mem Cognit. 1989; 17: 175–185.
37. Markham R, Howie P, Hlavacek S. Reality monitoring in auditory and visual modalities: developmental trends and effects of cross-modal imagery. J Exp Child Psychol. 1999; 72: 51–70.
38. Tallal P. Rapid auditory processing in normal and disordered language development. J Speech Hear Res. 1976; 19: 561–571.

child; gait/physiology; attention

© 2003 Lippincott Williams & Wilkins, Inc.