Aging Clinical and Experimental Research

, Volume 25, Issue 5, pp 539–544

Brain activation during dual-task walking and executive function among older adults with mild cognitive impairment: a fNIRS study


    • Section for Health Promotion, Department for Research and Development to Support Independent Life of Elderly, Center for Gerontology and Social ScienceNational Center for Geriatrics and Gerontology
  • Hyuma Makizako
    • Section for Health Promotion, Department for Research and Development to Support Independent Life of Elderly, Center for Gerontology and Social ScienceNational Center for Geriatrics and Gerontology
  • Hiroyuki Shimada
    • Section for Health Promotion, Department for Research and Development to Support Independent Life of Elderly, Center for Gerontology and Social ScienceNational Center for Geriatrics and Gerontology
  • Hyuntae Park
    • Section for Physical Functioning Activation, Department of Functioning Activation, Center for Gerontology and Social ScienceNational Center for Geriatrics and Gerontology
  • Kota Tsutsumimoto
    • Section for Health Promotion, Department for Research and Development to Support Independent Life of Elderly, Center for Gerontology and Social ScienceNational Center for Geriatrics and Gerontology
    • Department of Rehabilitation ScienceKobe University Graduate School of Health Sciences
  • Kazuki Uemura
    • Section for Health Promotion, Department for Research and Development to Support Independent Life of Elderly, Center for Gerontology and Social ScienceNational Center for Geriatrics and Gerontology
  • Takao Suzuki
    • Research Institute, National Center for Geriatrics and Gerontology
Original Article

DOI: 10.1007/s40520-013-0119-5

Cite this article as:
Doi, T., Makizako, H., Shimada, H. et al. Aging Clin Exp Res (2013) 25: 539. doi:10.1007/s40520-013-0119-5


Background and aims

Dual-task walking (DTW) is thought to involve activation of the prefrontal cortex in healthy adults and to be affected by cognitive impairment. However, it is unclear whether prefrontal cortex activation is involved in DTW in older adults with mild cognitive impairment. This study examined brain activation during DTW among older adults with mild cognitive impairment using functional near-infrared spectroscopy.


Sixteen older adults (aged 75.4 ± 7.2 years, women n = 6) performed gait experiments under normal walking and DTW conditions. We used a design with 60-s blocks consisting of a 10-s rest standing as pre-resting period, a 20-s walking task period, and a 30-s rest standing as post-resting period. Walking speed was measured during a 20-s walking task. Changes in oxy-hemoglobin were measured in the prefrontal area during gait experiments.


Walking speed was slower during DTW compared with normal walking (p < 0.001). The oxy-hemoglobin level during DTW was higher than during normal walking (p < 0.001) and was correlated with executive function, as measured by Stroop interference (p < 0.05).


Our findings indicate that DTW is associated with prefrontal activation among older adults with mild cognitive impairment. The brain activation during DTW was correlated with executive function. Additional studies are necessary to elucidate the effects of cognitive impairment on the association between prefrontal activity and walking under various conditions.


Dual-task walkingMild cognitive impairmentFunctional near-infrared spectroscopyOlder adults


The act of walking is an inherently complex and unstable movement task. Because the motor and cognitive systems act reciprocally, walking is not regulated automatically. Executive function is thought to be primarily mediated by the prefrontal lobe and plays an important role in successful locomotion. The association between executive function and walking has been widely investigated using the dual-task paradigm [13]. In dual-task walking (DTW), participants walk while performing a cognitive task. Gait performance during DTW has been found to correlate with executive function measured by neuropsychological assessment, and lower executive function is reported to be associated with lower gait performance during DTW [4, 5].

The effects of DTW on gait performance are accentuated in cognitive impairment [710]. This phenomenon is considered a clinical characteristic of older adults with Alzheimer’s disease (AD) [6, 7] and mild cognitive impairment (MCI) [79], which is a prodromal state between normal aging and AD [10]. The maintained ability to walk is thought to have potential beneficial effects in the prevention of cognitive decline and dysfunction [11]. DTW has been used as an intervention to enhance not only physical performance but also cognitive function among healthy older adults [12]. The efficacy of this technique for stimulating brain function is supported by evidence that DTW induces prefrontal activation among healthy older adults [13]. However, it is currently unclear whether prefrontal lobe activity is induced during DTW among older adults with cognitive impairment. Elucidating the association between brain activation and DTW among older adults with cognitive impairment is important for the better understanding of the utility of dual-task walking as an intervention and assessment tool.

Several neuroimaging techniques have been developed to identify patterns of brain activity during walking. These include the measurement of glucose metabolism during actual walking using positron emission tomography with [18F] fluorodeoxyglucose [14], and single-photon emission tomography with technetium-99m hexamethylpropylene amine oxime or 99mTc-ethyl cysteinate dimer to measure fixed regional cerebral blood flow [15]. In addition, functional near-infrared spectroscopy (fNIRS) has been used to measure cerebral hemodynamics in walking experiments [13, 16]. Whereas positron emission tomography and single-photon emission tomography cannot measure walking and brain activation at the same time, fNIRS can be used to measure cerebral hemodynamics while participants are actively engaged in walking. In addition, fNIRS is robust to dynamic body motion and can be used to effectively analyze brain function during a variety of movement tasks without the gamma exposure involved in other methods. Thus, the portable fNIRS system can be used to measure prefrontal activation during walking under unrestricted conditions.

The present study investigated brain activation during DTW among older adults with MCI. MCI involves a relatively high risk of progression to AD [10], and effective interventions to improve cognitive function in MCI are urgently required. Confirming the effects of DTW on brain activation could aid the development of effective intervention methods. In the present study, we used fNIRS, which has been previously used to measure brain activation during dynamic movement [13, 16], to monitor brain activation during DTW. We focused our examination on the prefrontal area, because DTW is reported to be associated with executive function in the prefrontal lobe [4, 5]. We hypothesized that DTW would elicit activity in prefrontal cortex even among older adults with MCI and that activation during DTW would be correlated with executive function.

Subjects and methods


Subjects in our study were 16 right-handed older adults (6 females). Inclusion criteria were 65 years or older, living independently in the community (i.e., no impairment of activities of daily living), and Japanese speaking, with sufficient hearing and visual acuity to participate in the examinations, and general cognitive functioning (Mini-Mental State Examination (MMSE) score >23 [17]). The subjects fulfilled the diagnostic criteria of MCI. The process we used to define and select MCI patients was described in detail in a previous study [18]. We used the diagnostic criteria for MCI outlined by Petersen [10], selecting individuals who exhibited the following characteristics: independent daily living, no dementia, and objective cognitive impairment. None of the subjects had a previous history of major psychiatric illness (e.g., schizophrenia or bipolar disorder), other serious neurological or musculoskeletal diagnoses, or clinical depression. In addition, we recorded the subjects’ age, body mass index, education history, and grip strength. This study was approved by the ethics committee of the National Center for Geriatrics and Gerontology.

Neuropsychological assessment

All neuropsychological assessments were conducted by licensed and well-trained speech therapists. Global cognitive function was examined using the MMSE. Executive function was assessed using the Stroop color and word test [19] and we used the modified Stroop test involving control and incongruent conditions [20]. In the control condition, the subjects were instructed to read aloud color-name words printed in black ink (e.g., the word blue printed in black ink). In the incongruent condition, color-name words were presented in colored ink, and the ink color did not match the color name (e.g., the word red printed in blue ink). There were 24 stimuli for each condition, and the time taken to complete each condition was measured. Stroop interference is calculated as the outcome of the modified Stroop test, the time difference between the incongruent and control conditions. Higher value of Stroop interference means lower function in executive function.

Walking condition

Two walking conditions were used: normal walking (NW) and DTW. In the NW condition, subjects were instructed to walk at their preferred speed. In the DTW condition, subjects walked while performing a verbal letter fluency task [21]. The letter fluency task is most commonly used in experiments regarding DTW [22]. Subjects completed three trials in each walking condition, and the block order was alternated (NW-1, DTW-1, NW-2, DTW-2, NW-3, DTW-3). Each block lasted 60 s, consisting of a 10-s period of standing at rest as a pre-resting period, a 20-s walking task period, and a 30-s period of standing at rest as a post-resting period. Subjects were instructed to walk along a 10-m corridor during the walking task period. At the end of corridor, subjects turned and kept walking until the end of walking period. The distance subjects walked during the 20-s walking task period was measured, and walking speed was calculated based on the distance covered. Walking speed was calculated for each condition (NW and DTW) separately.

fNIRS measurements

The hemodynamic response in PFC was measured using a 16-channel Spectratech OEG-16 system (Spectratech Inc., Yokohama, Japan). The fNIRS system used two wavelengths of near-infrared light (approximately 770 and 840 nm) whose absorption was recorded to estimate hemoglobin levels. The temporal resolution was set at 650 ms. The emission probes were located 3.0 cm apart from the detector probes. Six emission and six detector probes were arranged in a 6 (wide) × 2 (long) matrix on the participant’s forehead (Fig. 1). Thus, cortical responses were obtained from a total of 16 locations. The center of the probe matrix was placed on Fpz (midpoint between Fp1 and Fp2) in accordance with the international 10/20 system used in electroencephalography. This system is able to measure changes in oxygenated hemoglobin (oxy-Hb), de-oxygenated hemoglobin and total hemoglobin levels (mM-mm). fNIRS enables the investigation of regional cerebral blood oxygenation during physiological brain function [23]. Previous studies have found that oxy-Hb values are reliable and sensitive to locomotion-related changes in cerebral blood flow [13, 16]. Thus, we recorded changes in oxy-Hb levels while participants performed the NW and DTW tasks.
Fig. 1

Location of fNIRS probes and channels. The fNIRS system in this study consists of 6 emitters (white) and 6 detectors (black), resulting in 16 channels of source-detector pairs. Regions of interest for fNIRS data were arranged into two regions: right inferior frontal gyrus (channels 1, 2, 3, and 4), and left inferior frontal gyrus (channels 13, 14, 15, and 16)

Data analysis

For fNIRS data, signals were filtered with a 0.05-Hz low-pass filter to reject artifacts caused by minor movement of the subject, and a baseline correction was performed using linear fitting based on two types of baseline data: the mean across the first 10-s period in the pre-resting period, and the mean across the final 10-s period of the post-resting period. Then, signal averaging was conducted in each walking condition and the mean value of oxy-Hb within the walking period was calculated separately for each walking condition.

Walking speed was compared between the NW and DTW conditions using a paired t test. Using a linear mixed-effects model with walking condition (NW vs. DTW) and channel (ch1-16) as factors, and participant as a random effect, the effects of walking condition on oxy-Hb were determined. To examine the relationship between activation during the walking task and executive function in specific regions of PFC, regions of interest for NIRS data were arranged into two regions: (1) right inferior frontal gyrus (RIFG: channels 1, 2, 3, and 4) and (2) left inferior frontal gyrus (LIFG: channels 13, 14, 15, and 16) [24]. We analyzed NIRS data for these measurement points, labeled as RIFG and LIFG, which covered the anterior and lateral PFC [24]. Changes in oxy-Hb levels in ROIs were averaged for each subject. The correlation between oxy-Hb in regions of interest and executive function (Stroop interference) was analyzed using Spearman’s correlation coefficients. The level of statistical significance was set at 0.05 for all analyses, which were carried out using SPSS for Windows, version 20.0 (SPSS, Chicago, IL, USA).


Table 1 summarizes the subjects’ demographic characteristics and the results of neuropsychological assessments. Walking speed during DTW was significantly slower compared with NW (NW 0.98 ± 0.17 m/s, DTW 0.86 ± 0.19 m/s, t = 6.71, p < 0.001). The mean oxy-Hb values during walking task period are shown in Table 2. The task condition (NW vs. DTW) had a significant main effect on oxy-Hb, F (1, 463) = 26.44, p < 0.001, with higher values during DTW compared with NW. The channel factor did not have a significant effect on oxy-Hb, F (15, 463) = 0.78, p = 0.701. The correlation analysis revealed a correlation between oxy-Hb at each ROI and executive function represented by Stroop interference. The oxy-Hb in LIFG (ρ = −0.663, p = 0.005) but not RIFG ρ = −0.261, p = 0.328) was significantly correlated with executive function during DTW, while the oxy-Hb during NW in both ROIs were not significantly correlated with oxy-Hb (RIFG: ρ = −0.402, p = 0.123, LIFG: ρ = −0.479, p = 0.061) (Fig. 2).
Table 1

Subject characteristics


M ± SD

Age (years)

75.4 ± 7.2

Sex, women subjects (%)

55 (50)

Body mass index (kg/m2)

23.4 ± 2.7

Educational history (years)

10.8 ± 2.7

Mini-Mental State Examination, score

26.4 ± 2.0

Stroop interference (s)

18.25 ± 12.8

Grip strength (kg)

26.7 ± 11.3

Values are mean ± SD and numbers (proportion) for sex

Table 2

Oxy-Hb (mM-mm) during normal walking and dual-task walking


Normal walking

Dual-task walking


0.085 ± 0.116

0.168 ± 0.188


0.025 ± 0.140

0.125 ± 0.177


0.083 ± 0.161

0.181 ± 0.245


0.082 ± 0.135

0.184 ± 0.220


0.052 ± 0.129

0.131 ± 0.187


0.083 ± 0.142

0.200 ± 0.247


0.097 ± 0.163

0.194 ± 0.230


−0.007 ± 0.331

0.068 ± 0.378


0.089 ± 0.179

0.195 ± 0.290


0.093 ± 0.178

0.183 ± 0.290


0.058 ± 0.150

0.147 ± 0.193


0.081 ± 0.177

0.188 ± 0.290


0.072 ± 0.134

0.172 ± 0.252


0.048 ± 0.138

0.126 ± 0.223


0.086 ± 0.141

0.175 ± 0.296


0.068 ± 0.127

0.148 ± 0.259

Values are mean ± SD

Oxy-Hb oxy-hemoglobin
Fig. 2

The correlation between executive function (Stroop interference) and oxy-hemoglobin (oxy-Hb) in each ROI (RIFG and LIFG) during a normal walking and b dual-task walking. Higher value of Stroop interference reflects lower executive function


The current results revealed that DTW increased brain activation in prefrontal areas in older adults with MCI. Walking condition had a significant impact on oxy-Hb concentration during walking. In addition, gait velocity was decreased under DTW when compared with normal walking. Importantly, executive function was correlated with activation in prefrontal areas during DTW but not normal walking.

Several studies have reported a strong relationship between executive function and walking, particularly DTW [13]. Executive function is thought to be predominantly mediated by the prefrontal lobe, and is correlated with gait performance in DTW, with lower executive function associated with increasing deterioration of gait variables during DTW [4]. The prefrontal lobe function is thought to play a crucial role in DTW. However, most previous studies investigating the relationship between the prefrontal lobe and DTW have assessed executive function with neuropsychological testing alone [13]. The fNIRS studies have indicated that hemodynamic activation is linked with neurovascular coupling and increased oxy-Hb [25, 26]. Thus, the current finding of PFC activation during DTW has important implications, confirming the relationship between DTW and prefrontal lobe function.

To our knowledge, the present study is the first evidence that hemodynamic activity in the prefrontal area among older adults with MCI is more activated during DTW than normal walking. Holtzer et al. [13] suggested that DTW activates prefrontal regions and that activation in prefrontal regions during DTW exhibits age-related changes. The current results extend the previous findings from healthy subjects to older adults with MCI. A number of previous studies have shown that dual tasking during gait consistently alters gait performance among older adults with and without cognitive impairment [13]. The deterioration in gait variables caused by dual-task performance has been confirmed among MCI as well as cognitively intact individuals, although the degree of change and the specific characteristics of DTW among MCI individuals remain contentious [79, 27]. The MCI subjects in the present study exhibited dual-task-related changes in aspects of gait performance, such as walking speed and the prefrontal area was more activated during DTW than normal walking in MCI individuals. These results may support the idea that DTW requires prefrontal activation even among MCI subjects.

The prefrontal lobe plays a crucial role in executive function, which involves components of sustained and selective attention, response inhibition and working memory. The Stroop test is a neuropsychological measure used to examine executive function [19], based on the evidences that Stroop interference requires activation in lateral PFC in fNIRS studies [28, 29]. The current findings of the correlation between the executive function represented in Stroop interference performance and brain activation in prefrontal areas during DTW confirm, for the first time, a relationship between prefrontal lobe function and DTW among MCI subjects. These results confirm the clinical relevance of DTW, e.g., as an intervention using the DTW among MCI subjects, and the usefulness of fNIRS as a neuroimaging assessment tool in walking tasks.

The present study involved several limitations. First, brain activation during the task may have been affected by task difficulty and the level of cognitive function of subjects [30]. DTW in our study used the verbal fluency task, consistent with other studies of DTW using fNIRS [13]. However, it would be useful to investigate other cognitive tasks in the DTW paradigm [1]. The specific effects of different tasks should be further investigated. The sample size in our study was relatively small, and the heterogeneity of individuals with MCI warrants caution in interpreting the results. Furthermore, this experimental study should be conducted in both MCI and healthy age-matched subjects. The characteristic hemodynamic changes during DTW should be investigated. Further studies with large sample sizes should be conducted to verify the current findings. Finally, to confirm whether cortical activation during DTW can improve cognitive function among MCI subjects, an intervention study using DTW among older adults with MCI should be conducted.


The current findings support the idea that DTW involves brain activation in the prefrontal area among older adults with MCI. We found that prefrontal activation was increased during DTW, and this increase was correlated with executive function. These results indicate that exercise involving dual tasks may be useful for stimulating brain activation in older adults with MCI. Additional studies are necessary to elucidate the effects of cognitive impairment on the role of the prefrontal lobe during walking under various conditions and to determine the possible effects of dual-task exercise intervention on cognitive function through increased brain activation in elderly adults with MCI.


We would like to thank the Obu city office for help with participant recruitment. This work was supported by a grant from the Japanese Ministry of Health, Labour and Welfare and Grant-in-Aid for Research Activity Start-up (22800093) to T.D. in Japan.

Conflict of interest


Copyright information

© Springer International Publishing Switzerland 2013