Older adults with fear of falling show deficits in motor imagery of gait

Abstract

Objective

Understanding of the underlying mechanisms of Fear of Falling (FoF) could help to expand potential treatments. Given the nature of motor performance, the decline in the planning stage of motor execution may be associated with an expression of FoF. The aim of this study was to assess the planning/prediction accuracy in motor execution in people with FoF using gait-related motor imagery (MI).

Design

Cross-sectional case/control study.

Setting

Three health centers in Japan.

Participants

Two hundred and eighty-three community-dwelling older adults were recruited and stratified by presence of FoF as FoF group (n=178) or non-FoF group (n=107).

Measurements

Participants were tested for both imagery and execution tasks of a Timed Up and Go (TUG) test. The participants were first asked to imagine the trial (iTUG) and estimate the time it would take, and then perform the actual trial (aTUG). The difference between iTUG and aTUG (Δ TUG) was calculated.

Results

The FoF group was significantly slower in aTUG, but iTUG duration was almost identical between the two groups, resulting in significant overestimation in the FoF group. The adjusted logistic regression analysis showed that increased Δ TUG (i.e., tendency to overestimate) was significantly associated with FoF (OR = 1.05; 95% CI = 1.02–1.10). Low frequency of going outdoors was also associated with FoF (OR 2.95; 95% CI: 1.16–7.44).

Conclusions

Older adults with FoF overestimate their TUG performance, reflecting impairment in motor planning. Overestimation of physical capabilities can be an additional explanation of the high risk of falls in this population.

Introduction

Fear of Falling (FoF) is a major health problem among older adults characterized by loss of self-confidence, loss of self-efficacy, and activity avoidance, and future falls (1, 2). Physical, psychological, and social factors have been associated with FoF (3, 4, 5, 6, 7, 8, 9); however, the mechanisms by which FoF pose a higher risk of future falls are poorly known. A better understanding of the underlying mechanisms of FoF could help expand potential treatments.

As previously shown (10) there is evidence to suggest the close relationship between FoF and the decline in the planning stage of motor execution. When older adults with FoF face environmental challenges, they are not able to select the appropriate strategy in terms of visual searching and anticipatory postural adjustments to overcome these fall hazards (10, 11). It has been suggested that this incapability to avoid fall hazards, can be related to deficits in efficiently planning the motor execution of the necessary adjustments to elude falling (12). As such, by assessing the planning stage of motor execution, we may be able to understand the underlying possible mechanisms of FoF.

Motor imagery (MI), the mental simulation of an action without its actual execution (13, 14), is a valid methodology to assess the planning stage and prediction timing of motor execution (14). Imagining a movement relies on neural processes similar to those evoked during real performance of the same movement (13, 15, 16). Similar spatiotemporal/ neurophysiological characteristics between imagined and executed movements are provided by internal forward models that predict the sensory consequences of motor commands and specify motor commands required to achieve a given outcome efficiently and safely (i.e., motor planning) (17, 18, 19).

Based on these findings, we speculate that higher incongruence between imagined actions and actual actions is an expression of deficits in motor planning in older adults with FoF. These deficits may place them at the risk of failure during goal-directed behavior, such as tripping and stumbling during walking. Since MI has been shown to decline with aging (19, 20, 21) and to be associated with dysfunctions in gait related tasks (20), we hypothesize that deficits in MI of the Timed Up and Go (TUG) test will be more evident in older adults with FoF.

Methods

Participants

From 2012 to 2014, 285 community dwelling older adults (mean age = 73.5 years, SD = 5.5; 77.9% were women) were recruited from three health centers in Japan: Tokyo, Kanagawa, and Shiga, and they underwent a comprehensive assessment. The assessments of each cohort took place from January to March. Participants were included if they were 60 years and older, without unstable medical conditions, fully functional in activities of daily living, and without neurological overt diseases that may affect gait performance. Exclusion criteria included having neurological disease with motor deficits (e.g.: previous stroke), depression (Geriatric Depression Scale-GDS->5), cognitive impairment (Mini-Mental State Examination-MMSE-<27), visual deficits, abnormal gait at physical examination, use of assistive devices (e.g., cane walkers), and use of psychoactive medications or tranquilizers.

The study was conducted in accordance with the ethical standards of the Declaration of Helsinki. The research protocol was approved by the Tokyo Metropolitan Institute of Gerontology, and all participants provided written consent.

TUG test for motor imagery and execution

MI ability and actual performance were measured using TUG test (Figure 1). The method for determining MI ability using TUG test has been validated as a mental chronometry task (20, 22, 23, 24), an experimental approach that examines the time required by mental operations to obtain insights into human attention, perception, cognition, and action. These experiments were conducted in a sound-isolated room illuminated with homogeneous white lighting. A marker shaped like a circular cone was placed 3 meters in front of a chair with no armrest. Before the experiment trials, each participant was provided with comfortable walking shoes with rubber soles.

Figure 1.

Schematic illustration of both imagery and actual tasks of a Timed Up and Go test (TUG)

Participants were asked to sit on the chair ensuring that their heels touched the floor and their hands placed on their knees. For the imagery trial condition (iTUG), which was conducted before the actual TUG (aTUG) trial condition, participants were instructed to imagine performing a TUG trial with no actual motor action (conducted when seated on the chair) and estimate the time taken to complete the trial using a chronometer. The display of the chronometer was covered with black sheet so that participants did not see their iTUG time before performing the aTUG trial. Participants were instructed to imagine performing the task by themselves (i.e., first person perspective). For the aTUG trial condition, participants were instructed to stand up from the chair, walk forward to a cone-shaped marker at a distance of 3 meters from the chair, turn around the marker, return to the chair, and sit down. In both, imagine and actual trial, participants were instructed to perform them as fast as possible (brisk walking) and after hearing “ready-set-go”. We have previously demonstrated that “fast walking” is more sensitive to changes in high functional older adults (25). They were allowed to turn either to the right or to the left of the cone. For the actual TUG, the elapsed time from the moment their hips were raised from the chair to sitting down again on the chair was recorded with a chronometer (SEIKO SVAE109, temporal precision of within ±0.0012%). Participants received no feedback for both iTUG and aTUG results at any point during the TUG task. Prior to starting the task, we ensured that participants fully understood how to use the chronometer. Each participant performed a total of 4 trials, 2 trials for each condition; the mean each time was calculated per participant. The difference between iTUG and aTUG (Δ TUG) was then calculated to determine bias error of the MI of TUG. This was then standardized according to the aTUG with the following formula: [(aTUG - iTUG) / aTUG] × 100. A positive Δ TUG represents a tendency to underestimate actual TUG time (i.e., overestimate TUG ability) whereas a negative Δ TUG represents tendency to overestimate it (i.e., underestimate TUG ability). Assessment of fear of falling and covariates FoF was determined based on a previously validated questionnaire (26, 27, 28). Participants were asked to respond “yes/ no” to the question: “Are you afraid of falling?”, and were assigned to the FoF or non- FoF group based on their response. For the covariates, all participants were interviewed by either a physician or physical therapist who assessed their healthrelated characteristics (e.g., demographics, comorbidities, history of hospitalization, and medication). Functional capacity, frequency of going outdoors, depression status, and fall history were also recorded. Functional capacity was evaluated using the Tokyo Metropolitan Institute of Gerontology Index of Competence (TMIG-IC) (29). A questionnaire consisting of three subscales (IADL, intellectual activity, and sociability); scores range from 0 to 13, with higher scores indicating greater functional capacity. The participants' daily practice of going outdoors was examined based on which they were assigned either to a high- (go out every day) or low- (go out every few days or less) frequency of going outdoors group using a valid questionnaire (30). Depression status was assessed by the GDS, ranging from 0–15, with higher scores indicated more depressive symptoms (31). History of falls in the previous display of the chronometer was covered with black sheet so that participants did not see their iTUG time before performing the aTUG trial. Participants were instructed to imagine performing the task by themselves (i.e., first person perspective). For the aTUG trial condition, participants were instructed to stand up from the chair, walk forward to a cone-shaped marker at a distance of 3 meters from the chair, turn around the marker, return to the chair, and sit down.

In both, imagine and actual trial, participants were instructed to perform them as fast as possible (brisk walking) and after hearing “ready-set-go”. We have previously demonstrated that “fast walking” is more sensitive to changes in high functional older adults (25). They were allowed to turn either to the right or to the left of the cone. For the actual TUG, the elapsed time from the moment their hips were raised from the chair to sitting down again on the chair was recorded with a chronometer (SEIKO SVAE109, temporal precision of within ±0.0012%).

Participants received no feedback for both iTUG and aTUG results at any point during the TUG task. Prior to starting the task, we ensured that participants fully understood how to use the chronometer. Each participant performed a total of 4 trials, 2 trials for each condition; the mean each time was calculated per participant. The difference between iTUG and aTUG (Δ TUG) was then calculated to determine bias error of the MI of TUG. This was then standardized according to the aTUG with the following formula: [(aTUG - iTUG) / aTUG] × 100. A positive Δ TUG represents a tendency to underestimate actual TUG time (i.e., overestimate TUG ability) whereas a negative Δ TUG represents tendency to overestimate it (i.e., underestimate TUG ability).

Assessment of fear of falling and covariates

FoF was determined based on a previously validated questionnaire (26, 27, 28). Participants were asked to respond “yes/no” to the question: “Are you afraid of falling?”, and were assigned to the FoF or non- FoF group based on their response.

For the covariates, all participants were interviewed by either a physician or physical therapist who assessed their healthrelated characteristics (e.g., demographics, comorbidities, history of hospitalization, and medication). Functional capacity, frequency of going outdoors, depression status, and fall history were also recorded. Functional capacity was evaluated using the Tokyo Metropolitan Institute of Gerontology Index of Competence (TMIG-IC) (29). A questionnaire consisting of three subscales (IADL, intellectual activity, and sociability); scores range from 0 to 13, with higher scores indicating greater functional capacity. The participants' daily practice of going outdoors was examined based on which they were assigned either to a high- (go out every day) or low- (go out every few days or less) frequency of going outdoors group using a valid questionnaire (30). Depression status was assessed by the GDS, ranging from 0–15, with higher scores indicated more depressive symptoms (31). History of falls in the previous year was recorded; a fall was defined as an unintentional drop or fall to the ground, excluding bicycle accidents, accidental contact with furniture, walls, or other environmental structures, and sudden cardiovascular or central nervous system events (32). Maximum grip strength of the dominant hand and fast gait velocity was also measured as covariates using a handheld Smedley-type dynamometer and the 5-meter fast gait test, respectively (25).

Data analysis

Descriptive statistics of the differences between the FoF and non-FoF groups was analyzed using Chi-square tests for the clinical and demographic variables (i.e., sex ratio, fall history, and frequency of going outdoors); and using multivariate analysis of variance (MANOVA) for anthropometric variables (i.e. age, height, and weight). To examine the difference in accuracy of the MI of TUG between FoF and non-FoF groups, a repeated measures analysis of variance (ANOVA) adjusted with sex and age was performed. A two-sample t-test was also performed to examine the difference in Δ TUG between the FoF and non-FoF groups. The correlations between iTUG and aTUG (i.e., accuracy of MI) were examined separately for the FoF and non-FoF groups. To examine the relationship between decreasing MI ability and FoF, a logistic regression analysis was performed on FoF (i.e., with or without FoF) as a dependent variable. For this, biological variables (i.e., age and sex), functional capacity (i.e., TMIG-IC) and other possible confounding factors for FoF (i.e., comorbidities, fall history, frequency of going outdoors, depression symptoms, and fast gait speed) were used as covariates. All statistical analyses were performed with the PC-compatible version of IBM SPSS version 20.0 (SPSS Inc., Chicago, IL).

Results

Clinical and demographic characteristics of participants, divided by FoF groups, are shown in Table 1. The FoF group was significantly comprised of more women (p < .001), showed lower frequency of going outdoors (p < .001), more history of falls (p < .001), lower grip strength (p = .001), and slower gait speed (p < .001) than those of the non-FoF group.

Table 1.

Demographic characteristics and anthropometric measurements for non-fear of falling (non-FoF) and fear of falling (FoF) groups

Variables	All participants (n = 285)	non-FoF (n = 107)	FoF (n = 178)	p-value
Age, mean (SD)	73.5 (5.5)	73.1 (5.3)	73.7 (5.6)	.368
Female, n (%)	222 (77.9)	72 (67.3)	150 (84.3)	<.001
Hypertension, n (%)	113 (39.6)	37 (34.6)	76 (42.7)	.175
Cerebrovascular disorder, n (%)	11 (3.9)	5 (4.7)	6 (3.4)	.581
Diabetes, n (%)	23 (8.1)	7 (6.5)	16 (9.0)	.463
Arthrosis, n (%)	35 (12.3)	13 (12.1)	22 (12.4)	.958
Body height, mean (SD)	154.0 (7.2)	155.3 (7.1)	153.2 (7.2)	.018
Weight, mean (SD)	53.4 (8.6)	53.3 (8.6)	53.3 (8.7)	.965
TMIG-IC, mean (SD)	12.4 (1.0)	12.3 (1.2)	12.5 (0.9)	.109
Geriatric Depression Scale, mean (SD)	1.5 (2.3)	1.2 (2.2)	1.7 (2.3)	.101
Fall experience within a year, n (%)	55 (19.3)	11 (10.3)	44 (24.7)	<.001
Low frequency of going outdoors, n (%)	37 (13.0)	7 (6.5)	30 (16.9)	.012
Grip strength (kg)	23.8 (6.7)	25.5 (7.8)	22.8 (5.7)	.001
Fast gait speed (m/s)	2.2 (0.4)	2.3 (0.4)	2.1 (0.4)	<.001
iTUG, s, mean (SD)	3.99 (1.22)	4.09 (1.25)	3.95 (1.20)	†
aTUG, s, mean (SD)	4.85 (0.84)	4.58 (0.72)	5.01 (0.87)	†
ΔTUG, %, mean (SD)	16.6 (1.9)	10.3 (15.7)	19.9 (14.5)	<.001‡

Note: TMIG-IC = Tokyo Metropolitan Institute of Gerontology Index of Competence; iTUG = imagery TUG; aTUG = actual TUG; ΔTUG = [(aTUG - iTUG) / aTUG] × 100.

^†A repeated measures analysis of variance was performed.

^‡A two-sample t-test was performed.

Figure 2 shows the results of the two-way ANOVA of FoF (the FoF group and non-FoF group), imagery-actual TUG task (iTUG and aTUG), and the comparison between the two groups on Δ TUG. A mixed-design two-way ANOVA showed significant main effects of imagery–actual TUG task (F1, 281 = 14.6, p < .001) but not for the FoF group (F1, 281 = 0.9, p = .320); there was significant interaction (F1, 279 = 10.6, p < .001) between the two factors. Post hoc analysis showed a significant simple main effect of group for aTUG (p < .001) but not for iTUG (p = .318). A two-sample t-test showed that Δ TUG for the FoF group was significantly larger than that for non-FoF group (t = 3.1, df = 283, p < .001).

Figure 2.

Graphical representation of imagery TUG (iTUG), actual TUG (aTUG), and error of motor imagery of the TUG test for the fear of falling (FoF) group and the non-FoF group

Figure 3 shows the correlations between iTUG and aTUG stratified by groups. Both conditions were significantly correlated for the non-FoF group (r = .437, p < .001) while having low correlation in the FoF group (r = .180, p < .001); the difference between the two correlation coefficients was significant (p < .001).

Figure 3.

Scatter diagrams of the imagery and execution tasks of a Timed Up and Go (TUG) test

The adjusted logistic regression analysis showed that increased Δ TUG (i.e., tendency to overestimate) was significantly associated with FoF (Odds Ratio = 1.05; 95% CI = 1.02–1.10; p = .025). Low frequency of going outdoors also showed significant relationship to FoF (Odds Ratio 2.95; 95% CI: 1.16–7.44; p = .023).

Discussion

Our results demonstrated that older adults with FoF have deficits in the ability to accurately imagine time to perform mobility related tasks. The overestimation of TUG performance by the FoF group suggest a deficit in their motor planning which can explain their high risk of future falls and, potentially, be a new target for interventions. To the best of our knowledge, these results provide the first evidence of a relationship between gait-related MI deficit in older adults with FoF.

Older adults with FoF traditionally show low performance in gait (26) and balance tests (27), and have more occurrence of falls (4, 5), suggesting that FoF simply reflects a rational appraisal of reduced functional abilities (33). Normally, imagined actions must be synonymous with physically capable/realizable actions and this high correspondence and agreement has been shown to be capital for planning or predicting outcomes (14). However, our results indicate that older adults with FoF may overestimate their ability to perform the TUG, which also suggests a lack of awareness of their physical capability. Previous studies have shown that the lack of awareness of age-related physical decline causes seniors to overestimate their physical ability (34, 35). This overestimation can jeopardize capabilities to safely perform daily activities, and consequently, can increases the risk of accidental falls (35).

A possible explanation of why gait-related MI deficit, specifically decline in the planning stage of gait execution, is independently associated with FoF can be related to the low physical activity resulting in/from FoF (i.e., inactive lifestyle or activity avoidance) which may result in lack of awareness of physical abilities and motor actions. For instance, a previous study indicated that motor imagery performance was better in children who frequently practiced vigorous physical activities when compared with those who did not (36). Similarly, physical-education students imagined their movement more clearly than physics/English/surveying students who may be more sedentary (37). Correspondingly, athletes were more clear in motor imagery than non-athletes (37). Regarding older adults, it has been shown that seniors who went outdoors infrequently overestimate their physical ability compared with seniors who did frequent the outdoors (30). These findings suggest that decline in MI ability might be mediated by activity avoidance/inactive lifestyle resulting in/from FoF. In addition to decreasing MI ability, low frequency of going outdoors increases the presentation of FoF by almost 3 times in our study. Further studies are needed to examine the longitudinal relationship between physical activity level, MI deficit, and FoF.

Previous works have shown that cognitive function, particularly working memory, is associated with MI ability in older adults (20, 23, 24). The association between MI ability and cognitive function is also supported by functional imaging studies showing the correlation between activation during imagery of gait and performance in executive function (22). Despite that executive function performance was not measured in our study, our results raise the possibility that deficits in MI can be also mediated by executive dysfunction.

Another possible explanation of the relationship between MI deficit and FoF is that anxiety may influence this relationship because anxiety has been associated with FoF (38) and with gait speed (39). Indeed, a previous work using 33 communitydwelling older adults with a history of falls suggested that depressive symptoms were associated with an underestimation of reach capacity among older adults (40). In the present study, we purposely excluded participants with depression and/or anxiety and, therefore, the probability that MI deficits being mediated by depression or anxiety is low.

Our study has several strengths including one of the largest sample sizes analyzing potential mechanisms in older adults with FoF, and being the first to assess imagined and real motor performance in this population. However, some limitations need to be outlined. The cross-sectional design precludes us from drawing inferences of the likely causal relationships between FoF and MI deficits. Despite the use of a dichotomized question for assessing FoF is reliable, valid, and simple to use in clinics (26, 27, 28), this approach is limited to measure the different level of FoF in various circumstances. Our sample was composed of high functioning community-based volunteers, in terms of grip strength and gait velocity, compared with previous studies (41, 42), which limits the generalizability of our results. Although global cognitive function was evaluated in our participants, we lack information related to cognitive domains, including executive function, which might mediate the association between FoF and poor motor imaginary. A future longitudinal study is needed to examine causal relationships and mediators in these associations using more detailed measures of FoF.

Conclusion

Older adults with FoF could not accurately imagine their TUG and overestimated their performance, implying that FoF might be mediated by deficits in the MI ability of gait. Overestimation of physical capabilities can place seniors with FoF at risk of falls while dealing with environmental challenges. Our findings suggest that decline in the planning/ prediction stage of motor execution may underline the relation between FoF and Motor Imagery. Further research should explore the role of cognitive and emotional status, including executive function and anxiety status, as potential mediators of the relation between gait-related MI deficits and FoF.

Acknowledgement: The study was conducted and draft was written while the first author was a PhD student at the Tokyo Metropolitan University under the supervision of Dr. Imanaka. This study was supported by Grant-in-Aid for JSPS fellows (23-5365 and 26-7168). The authors acknowledge the continued efforts of the members of Unit 412 of the Tokyo Metropolitan Institute of Gerontology. We also gratefully acknowledge Anam Islam and Karen Gopaul (Gait and Brain Lab, Parkwood Institute) for their generous advice on the manuscript. Dr Montero-Odasso program in ‘Gait and Brain' function is supported by grants from the Canadian Institute of Health and Research (CIHR-IA), the Physician Services Incorporated Foundation of Canada (PSI), the Ontario Ministry of Research and Innovation, and by Department of Medicine Program of Experimental Medicine (POEM) Research Award from the University of Western Ontario. He is the first recipient of the Schulich Clinician-Scientist Award and holds the CIHR New Investigator Award.

Conflict of interest: The authors do not have any conflict of interest to disclose.

Ethical standard: All study procedures were conducted in accordance with the ethical standards of the Declaration of Helsinki and approved by the ethics committee of the Tokyo Metropolitan Institute of Gerontology.