Journal of Biomechanics xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke Elena Bergamini a,⇑, Marco Iosa b, Valeria Belluscio a, Giovanni Morone b,c, Marco Tramontano b, Giuseppe Vannozzi a a Interuniversity Centre of Bioengineering of the Human Neuromusculoskeletal System (BOHNES), Department of Movement, Human and Health Sciences, University of Rome ‘‘Foro Italico”, Rome, Italy b Clinical Laboratory of Experimental Neurorehabilitation, Fondazione Santa Lucia (Scientific Institute for Research Hospitalization and Health Care), Rome, Italy c Private Inpatients Unit, Fondazione Santa Lucia (Scientific Institute for Research Hospitalization and Health Care), Rome, Italy
a r t i c l e
i n f o
Article history: Accepted 22 July 2017 Available online xxxx Keywords: Stroke Gait stability Acceleration Inertial sensors Locomotion Fall risk
a b s t r a c t The capacity to maintain upright balance by minimising upper body oscillations during walking, also referred to as gait stability, has been associated with a decreased risk of fall. Although it is well known that fall is a common complication after stroke, no study considered the role of both trunk and head when assessing gait stability in this population. The primary aim of this study was to propose a multi-sensor protocol to quantify gait stability in patients with subacute stroke using gait quality indices derived from pelvis, sternum, and head accelerations. Second, the association of these indices with the level of walking ability, with traditional clinical scale scores, and with fall events occurring within the six months after patients’ dismissal was investigated. The accelerations corresponding to the three abovementioned body levels were measured using inertial sensors during a 10-Meter Walk Test performed by 45 inpatients and 25 control healthy subjects. A set of indices related to gait stability were estimated and clinical performance scales were administered to each patient. The amplitude of the accelerations, the way it is attenuated/amplified from lower to upper body levels, and the gait symmetry provide valuable information about subject-specific motor strategies, discriminate between different levels of walking ability, and correlate with clinical scales. In conclusion, the proposed multi-sensor protocol could represent a useful tool to quantify gait stability, support clinicians in the identification of patients potentially exposed to a high risk of falling, and assess the effectiveness of rehabilitation protocols in the clinical routine. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Falls are known to produce physical and psychological consequences imposing a tremendous economic burden on the health care system, bringing the efficacy of the entire rehabilitative pathway into question (Langhorne et al., 2000). Several studies focused on the identification of risk factors that could help recognising and treating patients exposed to a high risk of falling, such as persons who experienced cerebrovascular events (Campbell and Matthews, 2010; Fletcher and Hirdes, 2002). Among the huge number of risk factors that has been reported in the literature (Masud and Morris, 2001), the presence of gait instability has been
⇑ Corresponding author at: Interuniversity Centre of Bioengineering of the Human Neuromusculoskeletal System, Department of Movement, Human and Health Sciences, University of Rome ‘‘Foro Italico”, Piazza Lauro de Bosis 6, 00135 Rome, Italy. E-mail address: [email protected] (E. Bergamini).
acknowledged as one of the most important fall predictors (Campbell and Matthews, 2010; Hamacher et al., 2011). It has been shown, in fact, that individuals with impaired mobility are 1.65 times more likely to experience a fall and that up to 70% of the falls occurs during walking (Fletcher and Hirdes, 2002). As walking is one of the most frequent dynamic activities of daily living (Hamacher et al., 2011), the recovery of gait stability is one of the most important aim of neurorehabilitation (Paolucci et al., 2008). Gait stability can be referred to as the capacity to minimise oscillations, in a progressive way, from the lower to upper levels of the human body, and thus to maintain upright balance during walking (Cappozzo, 1981). In this framework, the development and validation of protocols aimed at objectively quantifying gait stability, and that can be regularly employed in the clinical routine, is crucial. Traditionally, clinical performance scales based on questionnaire checklists or patients’ qualitative observation are used. However, these scales may lack inter-rater reliability and specificity
http://dx.doi.org/10.1016/j.jbiomech.2017.07.034 0021-9290/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
2
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
(Mancini and Horak, 2010; Senden et al., 2012). For these reasons, instrumented gait analysis is used, allowing the estimation of numerous parameters based on biomechanical measures (Hamacher et al., 2011). However, the estimation of these parameters is limited to the laboratory environment and, due to the complexity of their interpretation, their clinical application is still limited (Cimolin and Galli, 2014). Therefore, technology-based protocols relying on the use of wearable Inertial Measurement Units (IMUs) have been recently proposed and flanked to clinical scale routines (Buckley et al., 2015; Iosa et al., 2012a; Senden et al., 2012). These protocols aim at collecting information directly inthe-field and obtaining concise quantitative metrics of the overall quality of gait (namely gait quality indices), that are able to assess patients’ motor ability at person level and that take into consideration features like maintenance of balance during walking and gait symmetry (Cappozzo, 1983; WHO, 2001). Several studies have focused on the assessment of gait stability using a single IMU located at the pelvis level and quantified the amount of accelerations and/or the gait symmetry in patients with Parkinson’s disease (Lowry et al., 2009), stroke (Iosa et al., 2016), cerebral palsy (Iosa et al., 2012b), lower limb amputation (Iosa et al., 2014), and in individuals at risk of falls (Isho et al., 2015; Senden et al., 2012). All these studies agree in associating to decreased gait stability higher values of accelerations and decreased gait symmetry. The use of a single IMU, however, does not allow to obtain information about the role of the whole trunk and head, which, in stroke patients, is crucial both in movement control and postural balance (Isho and Usuda, 2016). It has been reported, in fact, that the control of the head movements during walking allows for the stabilisation of the optic flow, for a more effective processing of the vestibular system signals, and for the consequent control of equilibrium (Berthoz and Pozzo, 1994; Cappozzo, 1981; Mazzà et al., 2008). In addition, the literature suggests that difficulties in controlling the upper body accelerations are also associated with a higher risk of fall (Marigold and Patla, 2008). In this respect, the use of a multi-sensor approach to gain insight on the way accelerations are attenuated from the pelvis to the head, in patients with stroke, may represent an added value for clinicians, supporting them in the definition of patient-specific treatments and in the assessment of rehabilitation programs’ efficacy. The literature provides examples of multi-sensor assessment of gait stability in the elderly (Doi et al., 2013; Mazzà et al., 2008; Menz et al., 2003a) and in patients with Parkinson’s disease (Buckley et al., 2015) or cerebral palsy (Summa et al., 2016), but lacks of studies on patients with stroke in the subacute phase (1-to-6 months post-stroke event). Furthermore, whereas a large number of studies focused on the ability of the abovementioned quantities to discriminate between healthy and pathological populations (Buckley et al., 2015; Iosa et al., 2014, 2012a, 2012b; Summa et al., 2016), no information is available about their capability to discriminate among different levels of walking ability, as defined by currently administered clinical scales, like the
Functional Ambulation Classification scale (Holden et al., 1984). Finally, it is still unclear whether an association exists between these quantities and clinical scale scores, as well as whether the former could be of additional value in current fall risk screening. The primary aim of the present study is, thus, to propose a multi-sensor protocol to quantify the stability of patients with subacute stroke during level walking, using indices based on accelerations measured at head, trunk, and pelvis levels. Second, the association of the estimated gait quality indices with the following aspects was investigated: (i) the level of walking ability; (ii) the scores of commonly administered clinical scales; (iii) the occurrence of fall episodes within the six months following patients’ dismissal. The outcome of the proposed multi-sensor protocol is expected to corroborate the following hypothesis: patients with stroke in the subacute phase present a lack of ability in attenuating accelerations from the pelvis to the head while walking, and thus a deficit in maintaining the head stable. This deficit is assumed to endanger the stabilisation of the optic flow and the physiological processing of the vestibular system signals, exposing the patients to an increased risk of fall. 2. Methods 2.1. Participants Two groups of subjects participated in this study, which was conducted according to the World Medical Association Declaration of Helsinki and was approved by the S. Lucia Foundation Ethics Committee (CE/AG4/PROG.383-11 and successive integrations). The first group was composed of 45 inpatients with subacute stroke (SG, age: 63 ± 13 years, 18 males) (Table 1) complying with the following inclusion criteria: first ever stroke with unilateral hemiplegia, stroke event occurred within the previous six months, and ability to walk without any device or need of continuous physical assistance (Functional Ambulation Classification 3). Exclusion criteria were: cognitive deficits affecting the capacity of patients to understand the task instructions (Mini Mental State Examination >24), severe unilateral spatial neglect, severe aphasia, and presence of neurological, orthopaedic or cardiac comorbidities. The second group was composed of 25 adults without neurological, orthopaedic, or cardiothoracic conditions that may have affected their walking (CG, age: 54 ± 8 years, 19 males). Each participant gave written informed consent. 2.2. Experimental protocol All acquisitions were performed in the rehabilitation gym of the S. Lucia Foundation hospital before patients’ dismissal. First, for each patient, the walking ability, the risk of fall, the presence of impairment in balance function, and the degree of independence in various activities of daily living were assessed, by an expert physiotherapist, using the following clinical scales:
Table 1 Demographic characteristic of the three groups of stroke patients (SG1, SG2, SG3).
Age [years]
Mean Standard deviation
Male sex [%] Time since stroke [days]
Mean Standard deviation
SG1
SG2
SG3
58.0 12.7
65.8 14.0
63.2 9.2
25.0
34.6
63.6
30.3 13.6
46.4 17.7
71.0 29.0
Stroke type, ischemic [%]
Ischemic
85.2
84.6
83.4
Stroke location [%]
Right
45.8
51.6
44.0
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
3
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
Functional Ambulation Classification (FAC) (Holden et al., 1984), Tinetti Balance and Gait (TBG) (Tinetti et al., 1986), Berg Balance Scale (BBS) (Berg et al., 1992), and Barthel Index (BI) (Collin et al., 1988). To codify for the different levels of walking ability, patients were further divided into three sub-groups according to their FAC score: SG1 (8 patients), SG2 (26 patients) and SG3 (11 patients) characterised by a score of 3, 4, and 5, respectively. The demographic characteristics of each subgroup are reported in Table 1. Second, participants were asked to stand still for 5 s and to perform a 10-Meter Walk Test (10-MWT), for three times consecutively, on a straight pathway at their self-selected walking speed, according to the criteria described by Duncan et al. (2003) and Perera et al. (2006). Five IMUs (Opal, APDM Inc., Portland, Oregon, USA) were used to collect 3D linear accelerations and angular velocities during the 10-MWT trials. Each unit embedded threeaxial accelerometers and gyroscopes (±6 g with g = 9.81 ms2, and ±1500 °/s of full-range scale, respectively) and provided the measured quantities with respect to a unit-embedded system of reference. To assess gait stability, three IMUs were secured to the participants’ upper body: one on the occipital cranium bone close to the lambdoid suture of the head (H), one on the centre of the sternum body (S), and one at L4-L5 level, slightly above the pelvis (P) (Fig. 1). The other two units were located on both distal tibiae (lateral malleoli) and were used to perform stride segmentation. To limit the relative movement between the units and the skin, IMUs were secured to the relevant body segment using ad hoc supports (a swim cap with a tailored pocket for the head IMU and elastic straps for the other units).
In order to investigate if any fall event occurred after patients’ dismissal, they were followed-up by phone interviews, made by the same physiotherapist, at two, four and six months after their dismissal (Morone et al., 2014). They were asked if any fall had occurred to them and, if yes, to describe how and why it happened. A fall event was defined as an unexpected event in which the subject come to rest on the ground, floor, or lower level (Lamb et al., 2005). 2.3. Data processing Data were processed using custom algorithms implemented in the MatlabÒ software (The MathWorks Inc., MA, US). To guarantee a repeatable system of reference for all participants, a verticalized local reference frame was used for the three IMUs located on the upper body. To this aim, a rigid transformation was determined, during the initial static window, that rotated the IMU frame so as to have an axis aligned to the gravity vector (Bergamini et al., 2014). The resulting axes could be considered to approximate antero-posterior (AP), medio-lateral (ML), and cranio-caudal (CC) anatomical axes. This time-invariant transformation was applied to the measured accelerations in each instant of time. The beginning and the end of each 10-MWT were identified on the tibiae angular velocity signals and, for each acceleration component, the average of the signal over the 10-MWT was subtracted from the whole data series. Finally, acceleration signals were lowpass filtered using a 4th-order Butterworth filter at 20 Hz, according to the result of a residual analysis (Winter, 1990) and to the guidelines reported in Pasciuto et al. (2017). Stride segmentation was then performed through a peak-detection algorithm applied to the angular velocity signals measured around the leg ML axis. For each 10-MWT, the following spatiotemporal parameters were obtained: average walking speed (WS = 10 m/time to complete the test), average stride length (SL = 10 m/total number of strides), and stride frequency (SF = total number of strides/time to complete the test). For the indices related to the gait stability, only steady-state strides were analysed, excluding the first and the last two strides. The following parameters were then estimated: Root Mean Square (RMS) values of each acceleration component (j) measured by the IMUs located on the upper body (P, S, H). To account for the influence of the participants’ walking speed on this parameter, RMS values were normalised according to Mizuike et al. (2009). Attenuation Coefficients (Mazzà et al., 2008) between each level pair of the upper body, for each acceleration component (j), defined as:
RMSj S 1 RMSj P RMSj H ACPHj ¼ 1 RMSj P RMSj H ACSHj ¼ 1 RMSj S
ACPSj ¼
Fig. 1. Location of the Inertial Measurement Units (IMUs) attached on the participants’ body segments. The axes orientation of the pelvis (P), sternum (S), and head (H) IMUs was the same during the static phase at the beginning of each trial. For the sake of clarity only the orientation of the pelvis unit is depicted (AP, antero-posterior; ML, medio-lateral; CC, cranio-caudal).
ð1Þ
A positive coefficient represents an attenuation of the accelerations from lower to upper body levels, whereas a negative coefficient indicates an amplification. Symmetry indices: improved Harmonic Ratio (iHR) (Pasciuto et al., 2017) for each acceleration component measured at the pelvis level. This index is based on a spectral analysis of the acceleration signals and was calculated as follows: P
iHR ¼ P
Power of intrinsic harmonics P 100 Power of intrinsic harmonics þ Power of extrinsic harmonics ð2Þ
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
4
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
where harmonics characterising a perfectly symmetrical gait are named intrinsic and harmonics leading to deviations from the ideal gait are named extrinsic (Cappozzo, 1982). This index has been recently proposed as an alternative to the commonly used Harmonic Ratio (HR) (Gage, 1967), in order to overcome a number of limitations of the latter (Bellanca et al., 2013; Pasciuto et al., 2017; Roche et al., 2013). It allows indeed a more intuitive interpretation (being expressed in percentage: 0% = total asymmetry, 100% = total symmetry) and relies on a more rigorous mathematical definition. To allow for a direct comparison with the existing literature, however, the HR was also calculated as follows, for each acceleration component measured at the pelvis level:
P Amplitude of intrinsinc harmonics HR ¼ P Amplitude of extrinsic harmonics
ð3Þ
2.4. Statistical analysis For the spatiotemporal parameters, the values corresponding to the 10-MWT trial characterised by the median WS were considered for each participant. For the RMS values, the attenuation coefficients and the symmetry indices, which were obtained over all the steady-state strides performed during the three 10-MWT trials, ~) and inter-quartile range (IQR) values were comthe median (x puted and further considered. Inferential statistical analysis was performed using the IBM SPSS Statistics software (v23, IBM Corp., Armonk, NY, U.S.A.; alpha level of significance = 0.05). The normal distribution of each parameter was verified using the Shapiro-Wilk test. As most of the parameters were not normally distributed, the following nonparametric tests were performed: Kruskal-Wallis H-test on all estimated parameters, to investigate if significant differences existed among the different levels of walking ability (‘‘group” factor). When a significant ‘‘group” effect was found, pairwise comparisons were performed using the independent Mann-Whitney U test with the HolmBonferroni correction. Spearman’s rank correlation coefficient (q) between gait quality indices and clinical scale scores, considering the whole SG. 3. Results 3.1. Clinical scale scores and spatiotemporal parameters The scores of the administered clinical scales and the values of the spatiotemporal parameters are reported in Table 2. A consis-
tent trend was observed for all scale scores and spatiotemporal parameters, which increased with the level of walking ability, i.e. from SG1 (FAC = 3) to SG3 (FAC = 5). Specifically, the Kruskal-Wallis H-test showed a significant effect of the ‘‘group” factor for the spatiotemporal parameters (v2(3) > 39.3, p < 0.001). Pairwise comparisons revealed that significant differences existed between each group pair for WS (p < 0.02, U < 74.5). Similar results were obtained for SL (p < 0.002, U < 64.0), for which the only non-significant difference was between SG3 and CG, whereas for SF, significant differences were found between CG and each stroke subgroup (p < 0.001, U < 41.0). 3.2. RMS values, attenuation coefficients and symmetry indices For what concerns the RMS values, attenuation coefficients and symmetry indices, the Kruskal-Wallis H-test showed a significant effect of the ‘‘group” factor for all parameters except for ACPHML and ACSHML (v2(3) > 11.5, p < 0.001 for the RMS values; v2(3) > 8.3, p < 0.05 for the attenuation coefficients, and v2(3) > 11.9, p < 0.001 for the symmetry indices). The box-plots and the detailed results of the Mann-Whitney U pairwise comparisons for RMS values and attenuation coefficients as well as both symmetry indices are reported in Figs. 2 and 3, respectively (p < 0.04, U < 77.0 for the RMS values; p < 0.05, U < 52.0 for the attenuation coefficients, and p < 0.03, U < 80.0 for the symmetry indices). 3.3. Association of the gait quality indices with the clinical scale scores The results of the correlation analysis (Table 3) show as several indices displayed a significant correlation with the clinical scales scores, particularly with the FAC scale. 3.4. Follow-up interviews and fall episodes According to the follow-up interviews, eight patients reported a fall episode within the six months after dismissal. Two of them reported that the fall was due to postural or gait instability, whereas the remaining falls were due to dizziness and orthostatic hypotension. Specifically, the first patient (P1) was part of SG1, whereas the second patient (P2) was part of SG2. Their scores in the clinical scales and the values of their spatiotemporal parameters are reported in Table 4, whereas their RMS values, attenuation coefficients, and symmetry indices are depicted in Figs. 2 and 3. It can be noted that, for almost all parameters, both patients are characterised by extremely different values (higher RMS values, or lower attenuation coefficients and symmetry indices) with respect
Table 2 Median (~ x) and interquartile range (IQR) values of the clinical scale scores and spatiotemporal parameters for the three groups of stroke patients (SG1, SG2, SG3) and for the control group (CG). Abbreviations: FAC, Functional Ambulation Classification; TBG, Tinetti Balance and Gait scale; BBS, Berg Balance Scale; BI, Barthel Index; WS, walking speed; SL, stride length; SF, stride frequency. SG1
SG2
SG3
CG
TBG
~ x IQR
17.5 3.5
21.0 4.5
25.0 2.0
– –
BBS
~ x IQR
37.5 10.2
45.0 7.0
50.0 4.0
– –
BI
~ x IQR
84.5 18.7
91.5 12.5
100.0 3.0
– –
WS [ms1]
~ x IQR
0.41 0.25
0.66 0.29
0.83 0.28
1.19 0.21
SL [m]
~ x IQR
0.69 0.21
1.00 0.28
1.25 0.32
1.29 0.25
SF [stridess1]
~ x IQR
0.67 0.09
0.67 0.15
0.72 0.10
0.85 0.16
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
5
Fig. 2. (A) Normalized RMS values and (B) attenuation coefficients (AC) for the three groups of stroke patients (SG1, SG2, SG3) and for CG. Medians and interquartile ranges are reported. AP, antero-posterior; ML, medio-lateral; CC, cranio-caudal; P, pelvis; S, sternum; H, head. The horizontal lines indicate statistically significant between-group differences. The values of each parameter for the two patients who experienced a fall within the six months after dismissal are also indicated (triangle: P1 and circle: P2).
Fig. 3. Symmetry indices (iHR, improved Harmonic Ratio; HR, Harmonic Ratio) for the three groups of stroke patients (SG1, SG2, SG3) and for CG. Medians and interquartile ranges are reported. AP, antero-posterior; ML, medio-lateral; CC, cranio-caudal; P, pelvis; S, sternum; H, head. The horizontal lines indicate statistically significant betweengroup differences. The values of each parameter for the two patients who experienced a fall within the six months after dismissal are also indicated (triangle: P1 and circle: P2).
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
6
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
Table 3 Spearman’s correlation coefficients (q) between each estimated parameter and each clinical scale. Statistical significance is indicated by asterisks (*p < 0.05; **p < 0.001). Abbreviations: FAC, Functional Ambulation Classification scale; TBG, Tinetti Balance and Gait scale; BBS, Berg Balance Scale; BI, Barthel Index; WS, walking speed; SL, stride length; SF, stride frequency; RMS, root mean square; AC, attenuation coefficient; HR, Harmonic Ratio; iHR, improved Harmonic Ratio; AP, antero-posterior; ML, medio-lateral; CC, craniocaudal; P, pelvis; S, sternum; H, head.
WS SL SF
FAC
TBG
BBS
BI
0.84** 0.77** 0.72**
0.54** 0.58** 0.20
0.71** 0.75** 0.11
0.44** 0.59** 0.06
RMS P
AP ML CC
0.66** 0.57** 0.44**
0.07 0.35* 0.09
0.16 0.38* 0.11
0.05 0.18 0.11
RMS S
AP ML CC
0.33** 0.68** 0.29*
0.38* 0.56** 0.24
0.45** 0.60** 0.24
0.15 0.40** 0.09
RMS H
AP ML CC
0.65** 0.62** 0.39**
0.37* 0.45** 0.24
0.42** 0.53** 0.26
0.21 0.22 0.07
AC PH
AP ML CC
0.38** 0.27* 0.25*
0.37* 0.29 0.20
0.41** 0.32* 0.13
0.25 0.15 0.19
AC PS
AP ML CC
0.32** 0.32** 0.49**
0.46** 0.39* 0.08
0.44** 0.42** 0.15
0.30 0.31* 0.10
AC SH
AP ML CC
0.66** 0.01 0.31**
0.28 0.01 0.31
0.33* 0.08 0.27
0.18 0.17 0.07
iHR
AP ML CC
0.68** 0.59** 0.69**
0.56** 0.00 0.57**
0.63** 0.10 0.72**
0.39** 0.17 0.49**
HR
AP ML CC
0.61** 0.39** 0.71**
0.60** 0.17 0.54**
0.67** 0.30 0.70**
0.45** 0.01 0.49**
Table 4 Results of the four administered clinical scales and of the spatiotemporal parameters for the two patients (P1 and P2) who reported a fall episode within the six months of followup. Abbreviations: FAC, Functional Ambulation Classification scale; TBG, Tinetti Balance and Gait scale; BBS, Berg Balance Scale; BI, Barthel Index; WS, walking speed; SL, stride length; SF, stride frequency.
P1 P2
FAC
TBG
BBS
BI
WS [ms1]
SL [m]
SF [stridess1]
3 4
15 18
38 45
93 65
0.30 0.44
0.71 0.77
0.42 0.57
to the other patients of the same subgroup (SG1 for P1 and SG2 for P2). 4. Discussion The present study proposed a multi-sensor approach to quantify upright balance during gait in patients with subacute stroke using gait quality indices derived from pelvis, sternum and head accelerations. In addition, the ability of these indices to discriminate between different levels of walking ability and their association with the scores of four commonly administered clinical scales, as well as with fall episodes occurred within the six months after patients’ dismissal, were investigated. The multi-sensor approach allowed to obtain information about trunk and head movements and, thus, about the different strategies patients adopt to control dynamic balance during walking. As expected, the clinical scale scores displayed a consistent increasing trend from low to high walking ability levels (Table 2), proving an interesting consistency among the different abilities/ impairments assessed by the four scales. A similar consistent trend was displayed by the spatiotemporal indices (Table 2), which increased with the level of walking ability. In addition, both WS and SL proved to significantly discriminate between each FAC level and between SG and CG, supporting their important role as infor-
mative and concise parameters to evaluate overall walking ability (Dickstein, 2008). RMS values increased as walking ability decreased in all three directions (AP, ML, CC) (Fig. 2A). This result supports the existing literature which associates high acceleration values with a decreased gait stability and an increased fall risk (Iosa et al., 2012a; Mazzà et al., 2008; Menz et al., 2003b; Summa et al., 2016). In addition, the RMS parameter proved to be a good indicator of the level of walking ability, discriminating not only between SG and CG, but also between most of the walking ability level pairs. Specifically, patients characterised by the most severe impairment (SG1, FAC = 3) displayed remarkably high accelerations, particularly at the pelvis and head. The main contribution of the present work is represented by the analysis of the way upper body accelerations are controlled by patients with stroke. This analysis, which was made possible by the devised multi-sensor protocol, indicates that patients are characterised by a general lack of ability in attenuating accelerations from lower to upper body levels, particularly in the AP direction, from both the pelvis and sternum to the head (Fig. 2B). This behaviour presumably leads to a decreased stabilisation of the head which, associated to the difficulty to properly select the pertinent sensory input when visual, vestibular and somatosensory systems send information simultaneously (Bonan et al., 2004), may cause
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
postural instability and expose patients to a high risk of falling (Marigold and Patla, 2008). On the other hand, when considering the AP component of the pelvis-to-sternum coefficient, SG displayed a higher attenuation with respect to CG. This result is in agreement with a previous study considering individuals with cerebral palsy (Summa et al., 2016) and suggests a defective control of the head in patients with stroke: excessive ML and AP accelerations at the pelvis are coupled with compensatory strategies at the trunk and head levels aiming at maintaining locomotion stability. Both symmetry indices (iHR and HR) showed a clear increasing trend as the walking ability increased, with SG1 patients presenting the lowest values (Fig. 3). This is in agreement with the existing literature and supports the association of reduced gait symmetry with both impairment and increased fall risk (Doi et al., 2013; Menz et al., 2003b). In addition, both indices discriminated among different levels of walking ability, especially along the AP and CC directions. The iHR index, in addition to relying on a formally correct mathematical definition, allowed a more intuitive understanding and interpretation with respect to the traditional HR. Significant relationships were found between almost all gait quality indices and the administered scales, in particular with the FAC (Table 3). The strong correlation obtained between all scales and WS, SL, iHR and HR is in agreement with the existing literature (Isho and Usuda, 2016; Senden et al., 2012). In addition, the association between the TBG and the RMS values supports the integration of traditional and technology-based protocols to improve current clinical routines for gait stability and fall risk assessment. As concerns the follow-up results, the TBG correctly identified the two patients who reported a fall episode as high-risk patients (score 19), supporting the existing literature which confirms the validity of this scale to identify individuals at risk of falling (Raîche et al., 2000). Conversely, the use of the sole BBS and BI to identify subject at high fall risk could be questioned as they were not effective in classifying P1 and P2. When considering the gait quality indices, interestingly, each patient is positioned outside (or at the upper bound limit) of his/her subgroup boxplot (Figs. 2 and 3), showing a remarkably increased instability, a deficit in attenuating accelerations, and a decreased symmetry during walking. Although promising, this result is far from representing a statistical evidence, due to the small number of patients reporting a fall, and further studies are needed to verify whether the proposed indices could represent a useful tool for the prediction of fall events in stroke patients. The main limitation of this study is about the different size of the stroke subgroups, rather small for SG1 and SG3. The stroke severity was assessed only in terms of the motor deficit affecting the degree of independence in various activities of daily living (thus using the Barthel Index) and no speculation was presented about the possible relationship between the current results and the stroke type, locus, or extent. Although this issue deserves attention, it was far beyond the aims of the present study. Finally, a small number of patients reported a fall in the follow-up, thus preventing to draw certain and general conclusions about the actual added value of the proposed indices in fall risk assessment. Nevertheless, interesting considerations were presented that need to be confirmed in future prospective studies.
5. Conclusions The main contribution of the present study with respect to the current literature is the use of a multi-sensor approach, which allowed identifying patient-specific motor strategies related to the trading between gait progression and stability. Using this approach, it was possible to demonstrate the formulated hypothe-
7
sis, proving that patients with stroke lack the ability of attenuating accelerations from the pelvis to the head while walking, presenting a deficit in stabilising the head, and thus potentially exposing themselves to an increased risk of falling. In particular, this study demonstrates that technology-based assessment protocols relying on the use of inertial measurement units can proficiently be included in the clinical routine assessment of gait stability of patients with stroke. The presented gait quality indices provide valuable information about the different motor strategies implemented by each patient during walking, complementing and integrating the outcomes of traditional clinical scales. Specifically, although clinical scales are able to depict general trends, the proposed instrumented protocol allows to obtain objective patient-specific information, thus contributing both to evidence-based and personalised medicine. The information obtained may be useful both to assess the effectiveness of rehabilitation protocols aimed at improving patients’ stability and limiting the risk of fall and to design personalised treatments. The latter, according to current results, may be focused on increasing the capacity of stabilising the head and the pelvis in order to reduce high accelerations displayed by the patients at these body levels. Furthermore, considering the great potential of task-oriented biofeedback in neurorehabilitation (Huang et al., 2006), the proposed multi-sensor approach could be integrated in a system providing real-time biofeedback about patients’ efficiency in attenuating accelerations from the pelvis to the head and, thus, in maintaining upper body stability during walking. Acknowledgements This work was supported by the University of Rome ‘‘Foro Italico” under the Grant PR_15/05196-01. The authors wish to thank Drs. Paolo Varvara Casella, Alessio Bricca, Cristina Calderone, Giacomo Palchetti, Giulia Burattini, and Cinzia Salvatore for their support in patients’ recruitment and data acquisition. The contribution of Dr. Stefano Paolucci for his assistance in the clinical interpretation of the results is also warmly acknowledged. Conflict of interest statement All the authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. References Bellanca, J.L., Lowry, K.A., VanSwearingen, J.M., Brach, J.S., Redfern, M.S., 2013. Harmonic ratios: a quantification of step to step symmetry. J. Biomech. 46, 828– 831. Berg, K.O., Wood-Dauphinee, S.L., Williams, J.I., Maki, B., 1992. Measuring balance in the elderly: validation of an instrument. Can. J. Public Health, S7–S11. Bergamini, E., Ligorio, G., Summa, A., Vannozzi, G., Cappozzo, A., Sabatini, A., 2014. Estimating orientation using magnetic and inertial sensors and different sensor fusion approaches: accuracy assessment in manual and locomotion tasks. Sensors 14, 18625–18649. Berthoz, A., Pozzo, T., 1994. Head and body coordination during loco- motion and complex movements. In: Swinnen, S., Massion, J., Heuer, H., Casaer, P. (Eds.), Interlimb Coordination: Neu- Ral, Dynamical, and Cognitive Constraints. Academic Press, pp. 147–165. Bonan, I.V.F.M., Colle, J.P., Guichard, E., Vicaut, M., Eisenfisz, Tran Ba, Huy, P., Yelnik, A.P., 2004. Reliance on visual information after stroke. Part I: balance on dynamic posturography. Arch. Phys. Med. Rehabil. 85, 268–273. Buckley, C., Galna, B., Rochester, L., Mazzà, C., 2015. Attenuation of upper body accelerations during gait: piloting an innovative assessment tool for Parkinson’s disease. Biomed. Res. Int., 865873 Campbell, G.B., Matthews, J.T., 2010. An integrative review of factors associated with falls during post-stroke rehabilitation. J. Nurs. Scholarsh. 42, 395–404. Cappozzo, A., 1981. Analysis of the linear displacement of the head and trunk during walking at different speeds. J. Biomech. 14, 411–425.
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
8
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
Cappozzo, A., 1982. Low frequency self-generated vibration during ambulation in normal men. J. Biomech. 15, 599–609. Cappozzo, A., 1983. Considerations on clinical gait evaluation. J. Biomech. 16, 302. Cimolin, V., Galli, M., 2014. Summary measures for clinical gait analysis: a literature review. Gait Posture 39, 1005–1010. Collin, C., Wade, D.T., Davies, S., Horne, V., 1988. The Barthel ADL index: a reliability study. Int. Disabil. Stud. 10, 61–63. Dickstein, R., 2008. Rehabilitation of gait speed after stroke: a critical review of intervention approaches. Neurorehabil. Neural Repair 22, 649–660. Doi, T., Hirata, S., Ono, R., Tsutsumimoto, K., Misu, S., Ando, H., 2013. The harmonic ratio of trunk acceleration predicts falling among older people: results of a 1year prospective study. J. Neuroeng. Rehabil. 10, 1–6. Duncan, P., Studenski, S., Richards, L., Gollub, S., Lai, S.M., Reker, D., Perera, S., Yates, J., Koch, V., Rigler, S., Johnson, D., 2003. Randomized clinical trial of therapeutic exercise in subacute stroke. Stroke 34, 2173–2180. Fletcher, P.C., Hirdes, J.P., 2002. Risk factors for falling among community-based seniors using home care services. J. Gerontol. A. Biol. Sci. Med. Sci. 57, M504– M510. Gage, H., 1967. Accelerographic analysis of human gait. In: Biomechanics Monograph. ASME United Engineering Center, New York, US, pp. 137–152. Hamacher, D., Singh, N.B., Van Dieën, J.H., Heller, M.O., Taylor, W.R., 2011. Kinematic measures for assessing gait stability in elderly individuals: a systematic review. J. R. Soc. Interface 8, 1682–1698. Holden, M.K., Gill, K.M., Magliozzi, M.R., Nathan, J., Piehl-Baker, L., 1984. Clinical gait assessment in the neurologically impaired. Reliability and meaningfulness. Phys. Ther. 64, 35–40. Huang, H., Wolf, S.L., He, J., 2006. Recent developments in biofeedback for neuromotor rehabilitation. J. Neuroeng. Rehabil. 3, 1–12. Iosa, M., Bini, F., Marinozzi, F., Fusco, A., Morone, G., Koch, G., Martino Cinnera, A., Bonnì, S., Paolucci, S., 2016. Stability and harmony of gait in patients with subacute stroke. J. Med. Biol. Eng. 36, 635–643. Iosa, M., Fusco, A., Morone, G., Pratesi, L., Coiro, P., Venturiero, V., De Angelis, D., Bragoni, M., Paolucci, S., 2012a. Assessment of upper-body dynamic stability during walking in patients with subacute stroke. J. Rehabil. Res. Dev. 49, 439– 450. Iosa, M., Marro, T., Paolucci, S., Morelli, D., 2012b. Stability and harmony of gait in children with cerebral palsy. Res. Dev. Disabil. 33, 129–135. Iosa, M., Paradisi, F., Brunelli, S., Delussu, A.S., Pellegrini, R., Zenardi, D., Paolucci, S., Traballesi, M., 2014. Assessment of gait stability, harmony, and symmetry in subjects with lower-limb amputation evaluated by trunk accelerations. J. Rehabil. Res. Dev. 51, 623–634. Isho, T., Tashiro, H., Usuda, S., 2015. Accelerometry-based gait characteristics evaluated using a smartphone and their association with fall risk in people with chronic stroke. J. Stroke Cerebrovasc. Dis. 24, 1305–1311. Isho, T., Usuda, S., 2016. Association of trunk control with mobility performance and accelerometry-based gait characteristics in hemiparetic patients with subacute stroke. Gait Posture 44, 89–93. Lamb, S.E., Jørstad-Stein, E.C., Hauer, K., Becker, C.Prevention of Falls Network Europe and Outcomes Consensus Group, 2005. Development of a common outcome data set for fall injury prevention trials: the prevention of falls network Europe consensus. J. Am. Geriatr. Soc. 53, 1618–1622.
Langhorne, P., Stott, D.J., Robertson, L., MacDonald, J., Jones, L., McAlpine, C., Dick, F., Taylor, G.S., Murray, G., 2000. Medical complications after stroke: a multicenter study. Stroke 31, 1223–1229. Lowry, K.A., Smiley-Oyen, A.L., Carrel, A.J., Kerr, J.P., 2009. Walking stability using harmonic ratios in Parkinson’s disease. Mov. Disord. 24, 261–267. Mancini, M., Horak, F.B., 2010. The relevance of clinical balance assessment tools to differentiate balance deficits. Eur. J. Phys. Rehabil. Med. 46, 239–248. Marigold, D.S., Patla, A.E., 2008. Age-related changes in gait for multi-surface terrain. Gait Posture 27, 689–696. Masud, T., Morris, R.O., 2001. Epidemiology of falls. Age Ageing 30, 3–7. Mazzà, C., Iosa, M., Pecoraro, F., Cappozzo, A., 2008. Control of the upper body accelerations in young and elderly women during level walking. J. Neuroeng. Rehabil. 5, 1–10. Menz, H.B., Lord, S.R., Fitzpatrick, R.C., 2003a. Acceleration patterns of the head and pelvis when walking on level and irregular surfaces. Gait Posture 18, 35–46. Menz, H.B., Lord, S.R., Fitzpatrick, R.C., 2003b. Acceleration patterns of the head and pelvis when walking are associated with risk of falling in community-dwelling older people. J. Gerontol. Ser. A Biol. Sci. Med. Sci. 58, 446–452. Mizuike, C., Ohgi, S., Morita, S., 2009. Analysis of stroke patient walking dynamics using a tri-axial accelerometer. Gait Posture 30, 60–64. Morone, G., Iosa, M., Pratesi, L., Paolucci, S., 2014. Can overestimation of walking ability increase the risk of falls in people in the subacute stage after stroke on their return home? Gait Posture 39, 965–970. Paolucci, S., Bragoni, M., Coiro, P., De Angelis, D., Fusco, F.R., Morelli, D., Venturiero, V., Pratesi, L., 2008. Quantification of the probability of reaching mobility independence at discharge from a rehabilitation hospital in nonwalking early ischemic stroke patients: a multivariate study. Cerebrovasc. Dis. 26, 16–22. Pasciuto, I., Bergamini, E., Iosa, M., Vannozzi, G., Cappozzo, A., 2017. Overcoming the limitations of the Harmonic Ratio for the reliable assessment of gait symmetry. J. Biomech. 53, 84–89. Perera, S., Mody, S.H., Woodman, R.C., Studenski, S.A., 2006. Meaningful change and responsiveness in common physical performance measures in older adults. J. Am. Geriatr. Soc. 54, 743–749. Raîche, M., Hébert, R., Prince, F., Corriveau, H., 2000. Screening older adults at risk of falling with the Tinetti balance scale. Lancet 356, 1001–1002. Roche, J.L., Lowry, K.A., Vanswearingen, J.M., Brach, J.S., Redfern, M.S., 2013. Harmonic Ratios: a quantification of step to step symmetry. J. Biomech. 46, 828–831. Senden, R., Savelberg, H.H.C.M., Grimm, B., Heyligers, I.C., Meijer, K., 2012. Accelerometry-based gait analysis, an additional objective approach to screen subjects at risk for falling. Gait Posture 36, 296–300. Summa, A., Vannozzi, G., Bergamini, E., Iosa, M., Morelli, D., Cappozzo, A., 2016. Multilevel upper body movement control during gait in children with cerebral palsy. PLoS One 11, e0151792. Tinetti, M.E., Williams, T.F., Mayewski, R., 1986. Fall risk index for elderly patients based on number of chronic disabilities. Am. J. Med. 80, 429–434. WHO, 2001. World Health Organization. International Classification of Functioning, Disability and Health (ICF). World Health Organization, Geneva. Winter, D.A., 1990. Biomechanics and motor control of human movement. John Wiley & Sons, New York, US.
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke Elena Bergamini a,⇑, Marco Iosa b, Valeria Belluscio a, Giovanni Morone b,c, Marco Tramontano b, Giuseppe Vannozzi a a Interuniversity Centre of Bioengineering of the Human Neuromusculoskeletal System (BOHNES), Department of Movement, Human and Health Sciences, University of Rome ‘‘Foro Italico”, Rome, Italy b Clinical Laboratory of Experimental Neurorehabilitation, Fondazione Santa Lucia (Scientific Institute for Research Hospitalization and Health Care), Rome, Italy c Private Inpatients Unit, Fondazione Santa Lucia (Scientific Institute for Research Hospitalization and Health Care), Rome, Italy
a r t i c l e
i n f o
Article history: Accepted 22 July 2017 Available online xxxx Keywords: Stroke Gait stability Acceleration Inertial sensors Locomotion Fall risk
a b s t r a c t The capacity to maintain upright balance by minimising upper body oscillations during walking, also referred to as gait stability, has been associated with a decreased risk of fall. Although it is well known that fall is a common complication after stroke, no study considered the role of both trunk and head when assessing gait stability in this population. The primary aim of this study was to propose a multi-sensor protocol to quantify gait stability in patients with subacute stroke using gait quality indices derived from pelvis, sternum, and head accelerations. Second, the association of these indices with the level of walking ability, with traditional clinical scale scores, and with fall events occurring within the six months after patients’ dismissal was investigated. The accelerations corresponding to the three abovementioned body levels were measured using inertial sensors during a 10-Meter Walk Test performed by 45 inpatients and 25 control healthy subjects. A set of indices related to gait stability were estimated and clinical performance scales were administered to each patient. The amplitude of the accelerations, the way it is attenuated/amplified from lower to upper body levels, and the gait symmetry provide valuable information about subject-specific motor strategies, discriminate between different levels of walking ability, and correlate with clinical scales. In conclusion, the proposed multi-sensor protocol could represent a useful tool to quantify gait stability, support clinicians in the identification of patients potentially exposed to a high risk of falling, and assess the effectiveness of rehabilitation protocols in the clinical routine. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Falls are known to produce physical and psychological consequences imposing a tremendous economic burden on the health care system, bringing the efficacy of the entire rehabilitative pathway into question (Langhorne et al., 2000). Several studies focused on the identification of risk factors that could help recognising and treating patients exposed to a high risk of falling, such as persons who experienced cerebrovascular events (Campbell and Matthews, 2010; Fletcher and Hirdes, 2002). Among the huge number of risk factors that has been reported in the literature (Masud and Morris, 2001), the presence of gait instability has been
⇑ Corresponding author at: Interuniversity Centre of Bioengineering of the Human Neuromusculoskeletal System, Department of Movement, Human and Health Sciences, University of Rome ‘‘Foro Italico”, Piazza Lauro de Bosis 6, 00135 Rome, Italy. E-mail address: [email protected] (E. Bergamini).
acknowledged as one of the most important fall predictors (Campbell and Matthews, 2010; Hamacher et al., 2011). It has been shown, in fact, that individuals with impaired mobility are 1.65 times more likely to experience a fall and that up to 70% of the falls occurs during walking (Fletcher and Hirdes, 2002). As walking is one of the most frequent dynamic activities of daily living (Hamacher et al., 2011), the recovery of gait stability is one of the most important aim of neurorehabilitation (Paolucci et al., 2008). Gait stability can be referred to as the capacity to minimise oscillations, in a progressive way, from the lower to upper levels of the human body, and thus to maintain upright balance during walking (Cappozzo, 1981). In this framework, the development and validation of protocols aimed at objectively quantifying gait stability, and that can be regularly employed in the clinical routine, is crucial. Traditionally, clinical performance scales based on questionnaire checklists or patients’ qualitative observation are used. However, these scales may lack inter-rater reliability and specificity
http://dx.doi.org/10.1016/j.jbiomech.2017.07.034 0021-9290/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
2
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
(Mancini and Horak, 2010; Senden et al., 2012). For these reasons, instrumented gait analysis is used, allowing the estimation of numerous parameters based on biomechanical measures (Hamacher et al., 2011). However, the estimation of these parameters is limited to the laboratory environment and, due to the complexity of their interpretation, their clinical application is still limited (Cimolin and Galli, 2014). Therefore, technology-based protocols relying on the use of wearable Inertial Measurement Units (IMUs) have been recently proposed and flanked to clinical scale routines (Buckley et al., 2015; Iosa et al., 2012a; Senden et al., 2012). These protocols aim at collecting information directly inthe-field and obtaining concise quantitative metrics of the overall quality of gait (namely gait quality indices), that are able to assess patients’ motor ability at person level and that take into consideration features like maintenance of balance during walking and gait symmetry (Cappozzo, 1983; WHO, 2001). Several studies have focused on the assessment of gait stability using a single IMU located at the pelvis level and quantified the amount of accelerations and/or the gait symmetry in patients with Parkinson’s disease (Lowry et al., 2009), stroke (Iosa et al., 2016), cerebral palsy (Iosa et al., 2012b), lower limb amputation (Iosa et al., 2014), and in individuals at risk of falls (Isho et al., 2015; Senden et al., 2012). All these studies agree in associating to decreased gait stability higher values of accelerations and decreased gait symmetry. The use of a single IMU, however, does not allow to obtain information about the role of the whole trunk and head, which, in stroke patients, is crucial both in movement control and postural balance (Isho and Usuda, 2016). It has been reported, in fact, that the control of the head movements during walking allows for the stabilisation of the optic flow, for a more effective processing of the vestibular system signals, and for the consequent control of equilibrium (Berthoz and Pozzo, 1994; Cappozzo, 1981; Mazzà et al., 2008). In addition, the literature suggests that difficulties in controlling the upper body accelerations are also associated with a higher risk of fall (Marigold and Patla, 2008). In this respect, the use of a multi-sensor approach to gain insight on the way accelerations are attenuated from the pelvis to the head, in patients with stroke, may represent an added value for clinicians, supporting them in the definition of patient-specific treatments and in the assessment of rehabilitation programs’ efficacy. The literature provides examples of multi-sensor assessment of gait stability in the elderly (Doi et al., 2013; Mazzà et al., 2008; Menz et al., 2003a) and in patients with Parkinson’s disease (Buckley et al., 2015) or cerebral palsy (Summa et al., 2016), but lacks of studies on patients with stroke in the subacute phase (1-to-6 months post-stroke event). Furthermore, whereas a large number of studies focused on the ability of the abovementioned quantities to discriminate between healthy and pathological populations (Buckley et al., 2015; Iosa et al., 2014, 2012a, 2012b; Summa et al., 2016), no information is available about their capability to discriminate among different levels of walking ability, as defined by currently administered clinical scales, like the
Functional Ambulation Classification scale (Holden et al., 1984). Finally, it is still unclear whether an association exists between these quantities and clinical scale scores, as well as whether the former could be of additional value in current fall risk screening. The primary aim of the present study is, thus, to propose a multi-sensor protocol to quantify the stability of patients with subacute stroke during level walking, using indices based on accelerations measured at head, trunk, and pelvis levels. Second, the association of the estimated gait quality indices with the following aspects was investigated: (i) the level of walking ability; (ii) the scores of commonly administered clinical scales; (iii) the occurrence of fall episodes within the six months following patients’ dismissal. The outcome of the proposed multi-sensor protocol is expected to corroborate the following hypothesis: patients with stroke in the subacute phase present a lack of ability in attenuating accelerations from the pelvis to the head while walking, and thus a deficit in maintaining the head stable. This deficit is assumed to endanger the stabilisation of the optic flow and the physiological processing of the vestibular system signals, exposing the patients to an increased risk of fall. 2. Methods 2.1. Participants Two groups of subjects participated in this study, which was conducted according to the World Medical Association Declaration of Helsinki and was approved by the S. Lucia Foundation Ethics Committee (CE/AG4/PROG.383-11 and successive integrations). The first group was composed of 45 inpatients with subacute stroke (SG, age: 63 ± 13 years, 18 males) (Table 1) complying with the following inclusion criteria: first ever stroke with unilateral hemiplegia, stroke event occurred within the previous six months, and ability to walk without any device or need of continuous physical assistance (Functional Ambulation Classification 3). Exclusion criteria were: cognitive deficits affecting the capacity of patients to understand the task instructions (Mini Mental State Examination >24), severe unilateral spatial neglect, severe aphasia, and presence of neurological, orthopaedic or cardiac comorbidities. The second group was composed of 25 adults without neurological, orthopaedic, or cardiothoracic conditions that may have affected their walking (CG, age: 54 ± 8 years, 19 males). Each participant gave written informed consent. 2.2. Experimental protocol All acquisitions were performed in the rehabilitation gym of the S. Lucia Foundation hospital before patients’ dismissal. First, for each patient, the walking ability, the risk of fall, the presence of impairment in balance function, and the degree of independence in various activities of daily living were assessed, by an expert physiotherapist, using the following clinical scales:
Table 1 Demographic characteristic of the three groups of stroke patients (SG1, SG2, SG3).
Age [years]
Mean Standard deviation
Male sex [%] Time since stroke [days]
Mean Standard deviation
SG1
SG2
SG3
58.0 12.7
65.8 14.0
63.2 9.2
25.0
34.6
63.6
30.3 13.6
46.4 17.7
71.0 29.0
Stroke type, ischemic [%]
Ischemic
85.2
84.6
83.4
Stroke location [%]
Right
45.8
51.6
44.0
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
3
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
Functional Ambulation Classification (FAC) (Holden et al., 1984), Tinetti Balance and Gait (TBG) (Tinetti et al., 1986), Berg Balance Scale (BBS) (Berg et al., 1992), and Barthel Index (BI) (Collin et al., 1988). To codify for the different levels of walking ability, patients were further divided into three sub-groups according to their FAC score: SG1 (8 patients), SG2 (26 patients) and SG3 (11 patients) characterised by a score of 3, 4, and 5, respectively. The demographic characteristics of each subgroup are reported in Table 1. Second, participants were asked to stand still for 5 s and to perform a 10-Meter Walk Test (10-MWT), for three times consecutively, on a straight pathway at their self-selected walking speed, according to the criteria described by Duncan et al. (2003) and Perera et al. (2006). Five IMUs (Opal, APDM Inc., Portland, Oregon, USA) were used to collect 3D linear accelerations and angular velocities during the 10-MWT trials. Each unit embedded threeaxial accelerometers and gyroscopes (±6 g with g = 9.81 ms2, and ±1500 °/s of full-range scale, respectively) and provided the measured quantities with respect to a unit-embedded system of reference. To assess gait stability, three IMUs were secured to the participants’ upper body: one on the occipital cranium bone close to the lambdoid suture of the head (H), one on the centre of the sternum body (S), and one at L4-L5 level, slightly above the pelvis (P) (Fig. 1). The other two units were located on both distal tibiae (lateral malleoli) and were used to perform stride segmentation. To limit the relative movement between the units and the skin, IMUs were secured to the relevant body segment using ad hoc supports (a swim cap with a tailored pocket for the head IMU and elastic straps for the other units).
In order to investigate if any fall event occurred after patients’ dismissal, they were followed-up by phone interviews, made by the same physiotherapist, at two, four and six months after their dismissal (Morone et al., 2014). They were asked if any fall had occurred to them and, if yes, to describe how and why it happened. A fall event was defined as an unexpected event in which the subject come to rest on the ground, floor, or lower level (Lamb et al., 2005). 2.3. Data processing Data were processed using custom algorithms implemented in the MatlabÒ software (The MathWorks Inc., MA, US). To guarantee a repeatable system of reference for all participants, a verticalized local reference frame was used for the three IMUs located on the upper body. To this aim, a rigid transformation was determined, during the initial static window, that rotated the IMU frame so as to have an axis aligned to the gravity vector (Bergamini et al., 2014). The resulting axes could be considered to approximate antero-posterior (AP), medio-lateral (ML), and cranio-caudal (CC) anatomical axes. This time-invariant transformation was applied to the measured accelerations in each instant of time. The beginning and the end of each 10-MWT were identified on the tibiae angular velocity signals and, for each acceleration component, the average of the signal over the 10-MWT was subtracted from the whole data series. Finally, acceleration signals were lowpass filtered using a 4th-order Butterworth filter at 20 Hz, according to the result of a residual analysis (Winter, 1990) and to the guidelines reported in Pasciuto et al. (2017). Stride segmentation was then performed through a peak-detection algorithm applied to the angular velocity signals measured around the leg ML axis. For each 10-MWT, the following spatiotemporal parameters were obtained: average walking speed (WS = 10 m/time to complete the test), average stride length (SL = 10 m/total number of strides), and stride frequency (SF = total number of strides/time to complete the test). For the indices related to the gait stability, only steady-state strides were analysed, excluding the first and the last two strides. The following parameters were then estimated: Root Mean Square (RMS) values of each acceleration component (j) measured by the IMUs located on the upper body (P, S, H). To account for the influence of the participants’ walking speed on this parameter, RMS values were normalised according to Mizuike et al. (2009). Attenuation Coefficients (Mazzà et al., 2008) between each level pair of the upper body, for each acceleration component (j), defined as:
RMSj S 1 RMSj P RMSj H ACPHj ¼ 1 RMSj P RMSj H ACSHj ¼ 1 RMSj S
ACPSj ¼
Fig. 1. Location of the Inertial Measurement Units (IMUs) attached on the participants’ body segments. The axes orientation of the pelvis (P), sternum (S), and head (H) IMUs was the same during the static phase at the beginning of each trial. For the sake of clarity only the orientation of the pelvis unit is depicted (AP, antero-posterior; ML, medio-lateral; CC, cranio-caudal).
ð1Þ
A positive coefficient represents an attenuation of the accelerations from lower to upper body levels, whereas a negative coefficient indicates an amplification. Symmetry indices: improved Harmonic Ratio (iHR) (Pasciuto et al., 2017) for each acceleration component measured at the pelvis level. This index is based on a spectral analysis of the acceleration signals and was calculated as follows: P
iHR ¼ P
Power of intrinsic harmonics P 100 Power of intrinsic harmonics þ Power of extrinsic harmonics ð2Þ
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
4
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
where harmonics characterising a perfectly symmetrical gait are named intrinsic and harmonics leading to deviations from the ideal gait are named extrinsic (Cappozzo, 1982). This index has been recently proposed as an alternative to the commonly used Harmonic Ratio (HR) (Gage, 1967), in order to overcome a number of limitations of the latter (Bellanca et al., 2013; Pasciuto et al., 2017; Roche et al., 2013). It allows indeed a more intuitive interpretation (being expressed in percentage: 0% = total asymmetry, 100% = total symmetry) and relies on a more rigorous mathematical definition. To allow for a direct comparison with the existing literature, however, the HR was also calculated as follows, for each acceleration component measured at the pelvis level:
P Amplitude of intrinsinc harmonics HR ¼ P Amplitude of extrinsic harmonics
ð3Þ
2.4. Statistical analysis For the spatiotemporal parameters, the values corresponding to the 10-MWT trial characterised by the median WS were considered for each participant. For the RMS values, the attenuation coefficients and the symmetry indices, which were obtained over all the steady-state strides performed during the three 10-MWT trials, ~) and inter-quartile range (IQR) values were comthe median (x puted and further considered. Inferential statistical analysis was performed using the IBM SPSS Statistics software (v23, IBM Corp., Armonk, NY, U.S.A.; alpha level of significance = 0.05). The normal distribution of each parameter was verified using the Shapiro-Wilk test. As most of the parameters were not normally distributed, the following nonparametric tests were performed: Kruskal-Wallis H-test on all estimated parameters, to investigate if significant differences existed among the different levels of walking ability (‘‘group” factor). When a significant ‘‘group” effect was found, pairwise comparisons were performed using the independent Mann-Whitney U test with the HolmBonferroni correction. Spearman’s rank correlation coefficient (q) between gait quality indices and clinical scale scores, considering the whole SG. 3. Results 3.1. Clinical scale scores and spatiotemporal parameters The scores of the administered clinical scales and the values of the spatiotemporal parameters are reported in Table 2. A consis-
tent trend was observed for all scale scores and spatiotemporal parameters, which increased with the level of walking ability, i.e. from SG1 (FAC = 3) to SG3 (FAC = 5). Specifically, the Kruskal-Wallis H-test showed a significant effect of the ‘‘group” factor for the spatiotemporal parameters (v2(3) > 39.3, p < 0.001). Pairwise comparisons revealed that significant differences existed between each group pair for WS (p < 0.02, U < 74.5). Similar results were obtained for SL (p < 0.002, U < 64.0), for which the only non-significant difference was between SG3 and CG, whereas for SF, significant differences were found between CG and each stroke subgroup (p < 0.001, U < 41.0). 3.2. RMS values, attenuation coefficients and symmetry indices For what concerns the RMS values, attenuation coefficients and symmetry indices, the Kruskal-Wallis H-test showed a significant effect of the ‘‘group” factor for all parameters except for ACPHML and ACSHML (v2(3) > 11.5, p < 0.001 for the RMS values; v2(3) > 8.3, p < 0.05 for the attenuation coefficients, and v2(3) > 11.9, p < 0.001 for the symmetry indices). The box-plots and the detailed results of the Mann-Whitney U pairwise comparisons for RMS values and attenuation coefficients as well as both symmetry indices are reported in Figs. 2 and 3, respectively (p < 0.04, U < 77.0 for the RMS values; p < 0.05, U < 52.0 for the attenuation coefficients, and p < 0.03, U < 80.0 for the symmetry indices). 3.3. Association of the gait quality indices with the clinical scale scores The results of the correlation analysis (Table 3) show as several indices displayed a significant correlation with the clinical scales scores, particularly with the FAC scale. 3.4. Follow-up interviews and fall episodes According to the follow-up interviews, eight patients reported a fall episode within the six months after dismissal. Two of them reported that the fall was due to postural or gait instability, whereas the remaining falls were due to dizziness and orthostatic hypotension. Specifically, the first patient (P1) was part of SG1, whereas the second patient (P2) was part of SG2. Their scores in the clinical scales and the values of their spatiotemporal parameters are reported in Table 4, whereas their RMS values, attenuation coefficients, and symmetry indices are depicted in Figs. 2 and 3. It can be noted that, for almost all parameters, both patients are characterised by extremely different values (higher RMS values, or lower attenuation coefficients and symmetry indices) with respect
Table 2 Median (~ x) and interquartile range (IQR) values of the clinical scale scores and spatiotemporal parameters for the three groups of stroke patients (SG1, SG2, SG3) and for the control group (CG). Abbreviations: FAC, Functional Ambulation Classification; TBG, Tinetti Balance and Gait scale; BBS, Berg Balance Scale; BI, Barthel Index; WS, walking speed; SL, stride length; SF, stride frequency. SG1
SG2
SG3
CG
TBG
~ x IQR
17.5 3.5
21.0 4.5
25.0 2.0
– –
BBS
~ x IQR
37.5 10.2
45.0 7.0
50.0 4.0
– –
BI
~ x IQR
84.5 18.7
91.5 12.5
100.0 3.0
– –
WS [ms1]
~ x IQR
0.41 0.25
0.66 0.29
0.83 0.28
1.19 0.21
SL [m]
~ x IQR
0.69 0.21
1.00 0.28
1.25 0.32
1.29 0.25
SF [stridess1]
~ x IQR
0.67 0.09
0.67 0.15
0.72 0.10
0.85 0.16
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
5
Fig. 2. (A) Normalized RMS values and (B) attenuation coefficients (AC) for the three groups of stroke patients (SG1, SG2, SG3) and for CG. Medians and interquartile ranges are reported. AP, antero-posterior; ML, medio-lateral; CC, cranio-caudal; P, pelvis; S, sternum; H, head. The horizontal lines indicate statistically significant between-group differences. The values of each parameter for the two patients who experienced a fall within the six months after dismissal are also indicated (triangle: P1 and circle: P2).
Fig. 3. Symmetry indices (iHR, improved Harmonic Ratio; HR, Harmonic Ratio) for the three groups of stroke patients (SG1, SG2, SG3) and for CG. Medians and interquartile ranges are reported. AP, antero-posterior; ML, medio-lateral; CC, cranio-caudal; P, pelvis; S, sternum; H, head. The horizontal lines indicate statistically significant betweengroup differences. The values of each parameter for the two patients who experienced a fall within the six months after dismissal are also indicated (triangle: P1 and circle: P2).
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
6
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
Table 3 Spearman’s correlation coefficients (q) between each estimated parameter and each clinical scale. Statistical significance is indicated by asterisks (*p < 0.05; **p < 0.001). Abbreviations: FAC, Functional Ambulation Classification scale; TBG, Tinetti Balance and Gait scale; BBS, Berg Balance Scale; BI, Barthel Index; WS, walking speed; SL, stride length; SF, stride frequency; RMS, root mean square; AC, attenuation coefficient; HR, Harmonic Ratio; iHR, improved Harmonic Ratio; AP, antero-posterior; ML, medio-lateral; CC, craniocaudal; P, pelvis; S, sternum; H, head.
WS SL SF
FAC
TBG
BBS
BI
0.84** 0.77** 0.72**
0.54** 0.58** 0.20
0.71** 0.75** 0.11
0.44** 0.59** 0.06
RMS P
AP ML CC
0.66** 0.57** 0.44**
0.07 0.35* 0.09
0.16 0.38* 0.11
0.05 0.18 0.11
RMS S
AP ML CC
0.33** 0.68** 0.29*
0.38* 0.56** 0.24
0.45** 0.60** 0.24
0.15 0.40** 0.09
RMS H
AP ML CC
0.65** 0.62** 0.39**
0.37* 0.45** 0.24
0.42** 0.53** 0.26
0.21 0.22 0.07
AC PH
AP ML CC
0.38** 0.27* 0.25*
0.37* 0.29 0.20
0.41** 0.32* 0.13
0.25 0.15 0.19
AC PS
AP ML CC
0.32** 0.32** 0.49**
0.46** 0.39* 0.08
0.44** 0.42** 0.15
0.30 0.31* 0.10
AC SH
AP ML CC
0.66** 0.01 0.31**
0.28 0.01 0.31
0.33* 0.08 0.27
0.18 0.17 0.07
iHR
AP ML CC
0.68** 0.59** 0.69**
0.56** 0.00 0.57**
0.63** 0.10 0.72**
0.39** 0.17 0.49**
HR
AP ML CC
0.61** 0.39** 0.71**
0.60** 0.17 0.54**
0.67** 0.30 0.70**
0.45** 0.01 0.49**
Table 4 Results of the four administered clinical scales and of the spatiotemporal parameters for the two patients (P1 and P2) who reported a fall episode within the six months of followup. Abbreviations: FAC, Functional Ambulation Classification scale; TBG, Tinetti Balance and Gait scale; BBS, Berg Balance Scale; BI, Barthel Index; WS, walking speed; SL, stride length; SF, stride frequency.
P1 P2
FAC
TBG
BBS
BI
WS [ms1]
SL [m]
SF [stridess1]
3 4
15 18
38 45
93 65
0.30 0.44
0.71 0.77
0.42 0.57
to the other patients of the same subgroup (SG1 for P1 and SG2 for P2). 4. Discussion The present study proposed a multi-sensor approach to quantify upright balance during gait in patients with subacute stroke using gait quality indices derived from pelvis, sternum and head accelerations. In addition, the ability of these indices to discriminate between different levels of walking ability and their association with the scores of four commonly administered clinical scales, as well as with fall episodes occurred within the six months after patients’ dismissal, were investigated. The multi-sensor approach allowed to obtain information about trunk and head movements and, thus, about the different strategies patients adopt to control dynamic balance during walking. As expected, the clinical scale scores displayed a consistent increasing trend from low to high walking ability levels (Table 2), proving an interesting consistency among the different abilities/ impairments assessed by the four scales. A similar consistent trend was displayed by the spatiotemporal indices (Table 2), which increased with the level of walking ability. In addition, both WS and SL proved to significantly discriminate between each FAC level and between SG and CG, supporting their important role as infor-
mative and concise parameters to evaluate overall walking ability (Dickstein, 2008). RMS values increased as walking ability decreased in all three directions (AP, ML, CC) (Fig. 2A). This result supports the existing literature which associates high acceleration values with a decreased gait stability and an increased fall risk (Iosa et al., 2012a; Mazzà et al., 2008; Menz et al., 2003b; Summa et al., 2016). In addition, the RMS parameter proved to be a good indicator of the level of walking ability, discriminating not only between SG and CG, but also between most of the walking ability level pairs. Specifically, patients characterised by the most severe impairment (SG1, FAC = 3) displayed remarkably high accelerations, particularly at the pelvis and head. The main contribution of the present work is represented by the analysis of the way upper body accelerations are controlled by patients with stroke. This analysis, which was made possible by the devised multi-sensor protocol, indicates that patients are characterised by a general lack of ability in attenuating accelerations from lower to upper body levels, particularly in the AP direction, from both the pelvis and sternum to the head (Fig. 2B). This behaviour presumably leads to a decreased stabilisation of the head which, associated to the difficulty to properly select the pertinent sensory input when visual, vestibular and somatosensory systems send information simultaneously (Bonan et al., 2004), may cause
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
postural instability and expose patients to a high risk of falling (Marigold and Patla, 2008). On the other hand, when considering the AP component of the pelvis-to-sternum coefficient, SG displayed a higher attenuation with respect to CG. This result is in agreement with a previous study considering individuals with cerebral palsy (Summa et al., 2016) and suggests a defective control of the head in patients with stroke: excessive ML and AP accelerations at the pelvis are coupled with compensatory strategies at the trunk and head levels aiming at maintaining locomotion stability. Both symmetry indices (iHR and HR) showed a clear increasing trend as the walking ability increased, with SG1 patients presenting the lowest values (Fig. 3). This is in agreement with the existing literature and supports the association of reduced gait symmetry with both impairment and increased fall risk (Doi et al., 2013; Menz et al., 2003b). In addition, both indices discriminated among different levels of walking ability, especially along the AP and CC directions. The iHR index, in addition to relying on a formally correct mathematical definition, allowed a more intuitive understanding and interpretation with respect to the traditional HR. Significant relationships were found between almost all gait quality indices and the administered scales, in particular with the FAC (Table 3). The strong correlation obtained between all scales and WS, SL, iHR and HR is in agreement with the existing literature (Isho and Usuda, 2016; Senden et al., 2012). In addition, the association between the TBG and the RMS values supports the integration of traditional and technology-based protocols to improve current clinical routines for gait stability and fall risk assessment. As concerns the follow-up results, the TBG correctly identified the two patients who reported a fall episode as high-risk patients (score 19), supporting the existing literature which confirms the validity of this scale to identify individuals at risk of falling (Raîche et al., 2000). Conversely, the use of the sole BBS and BI to identify subject at high fall risk could be questioned as they were not effective in classifying P1 and P2. When considering the gait quality indices, interestingly, each patient is positioned outside (or at the upper bound limit) of his/her subgroup boxplot (Figs. 2 and 3), showing a remarkably increased instability, a deficit in attenuating accelerations, and a decreased symmetry during walking. Although promising, this result is far from representing a statistical evidence, due to the small number of patients reporting a fall, and further studies are needed to verify whether the proposed indices could represent a useful tool for the prediction of fall events in stroke patients. The main limitation of this study is about the different size of the stroke subgroups, rather small for SG1 and SG3. The stroke severity was assessed only in terms of the motor deficit affecting the degree of independence in various activities of daily living (thus using the Barthel Index) and no speculation was presented about the possible relationship between the current results and the stroke type, locus, or extent. Although this issue deserves attention, it was far beyond the aims of the present study. Finally, a small number of patients reported a fall in the follow-up, thus preventing to draw certain and general conclusions about the actual added value of the proposed indices in fall risk assessment. Nevertheless, interesting considerations were presented that need to be confirmed in future prospective studies.
5. Conclusions The main contribution of the present study with respect to the current literature is the use of a multi-sensor approach, which allowed identifying patient-specific motor strategies related to the trading between gait progression and stability. Using this approach, it was possible to demonstrate the formulated hypothe-
7
sis, proving that patients with stroke lack the ability of attenuating accelerations from the pelvis to the head while walking, presenting a deficit in stabilising the head, and thus potentially exposing themselves to an increased risk of falling. In particular, this study demonstrates that technology-based assessment protocols relying on the use of inertial measurement units can proficiently be included in the clinical routine assessment of gait stability of patients with stroke. The presented gait quality indices provide valuable information about the different motor strategies implemented by each patient during walking, complementing and integrating the outcomes of traditional clinical scales. Specifically, although clinical scales are able to depict general trends, the proposed instrumented protocol allows to obtain objective patient-specific information, thus contributing both to evidence-based and personalised medicine. The information obtained may be useful both to assess the effectiveness of rehabilitation protocols aimed at improving patients’ stability and limiting the risk of fall and to design personalised treatments. The latter, according to current results, may be focused on increasing the capacity of stabilising the head and the pelvis in order to reduce high accelerations displayed by the patients at these body levels. Furthermore, considering the great potential of task-oriented biofeedback in neurorehabilitation (Huang et al., 2006), the proposed multi-sensor approach could be integrated in a system providing real-time biofeedback about patients’ efficiency in attenuating accelerations from the pelvis to the head and, thus, in maintaining upper body stability during walking. Acknowledgements This work was supported by the University of Rome ‘‘Foro Italico” under the Grant PR_15/05196-01. The authors wish to thank Drs. Paolo Varvara Casella, Alessio Bricca, Cristina Calderone, Giacomo Palchetti, Giulia Burattini, and Cinzia Salvatore for their support in patients’ recruitment and data acquisition. The contribution of Dr. Stefano Paolucci for his assistance in the clinical interpretation of the results is also warmly acknowledged. Conflict of interest statement All the authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. References Bellanca, J.L., Lowry, K.A., VanSwearingen, J.M., Brach, J.S., Redfern, M.S., 2013. Harmonic ratios: a quantification of step to step symmetry. J. Biomech. 46, 828– 831. Berg, K.O., Wood-Dauphinee, S.L., Williams, J.I., Maki, B., 1992. Measuring balance in the elderly: validation of an instrument. Can. J. Public Health, S7–S11. Bergamini, E., Ligorio, G., Summa, A., Vannozzi, G., Cappozzo, A., Sabatini, A., 2014. Estimating orientation using magnetic and inertial sensors and different sensor fusion approaches: accuracy assessment in manual and locomotion tasks. Sensors 14, 18625–18649. Berthoz, A., Pozzo, T., 1994. Head and body coordination during loco- motion and complex movements. In: Swinnen, S., Massion, J., Heuer, H., Casaer, P. (Eds.), Interlimb Coordination: Neu- Ral, Dynamical, and Cognitive Constraints. Academic Press, pp. 147–165. Bonan, I.V.F.M., Colle, J.P., Guichard, E., Vicaut, M., Eisenfisz, Tran Ba, Huy, P., Yelnik, A.P., 2004. Reliance on visual information after stroke. Part I: balance on dynamic posturography. Arch. Phys. Med. Rehabil. 85, 268–273. Buckley, C., Galna, B., Rochester, L., Mazzà, C., 2015. Attenuation of upper body accelerations during gait: piloting an innovative assessment tool for Parkinson’s disease. Biomed. Res. Int., 865873 Campbell, G.B., Matthews, J.T., 2010. An integrative review of factors associated with falls during post-stroke rehabilitation. J. Nurs. Scholarsh. 42, 395–404. Cappozzo, A., 1981. Analysis of the linear displacement of the head and trunk during walking at different speeds. J. Biomech. 14, 411–425.
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
8
E. Bergamini et al. / Journal of Biomechanics xxx (2017) xxx–xxx
Cappozzo, A., 1982. Low frequency self-generated vibration during ambulation in normal men. J. Biomech. 15, 599–609. Cappozzo, A., 1983. Considerations on clinical gait evaluation. J. Biomech. 16, 302. Cimolin, V., Galli, M., 2014. Summary measures for clinical gait analysis: a literature review. Gait Posture 39, 1005–1010. Collin, C., Wade, D.T., Davies, S., Horne, V., 1988. The Barthel ADL index: a reliability study. Int. Disabil. Stud. 10, 61–63. Dickstein, R., 2008. Rehabilitation of gait speed after stroke: a critical review of intervention approaches. Neurorehabil. Neural Repair 22, 649–660. Doi, T., Hirata, S., Ono, R., Tsutsumimoto, K., Misu, S., Ando, H., 2013. The harmonic ratio of trunk acceleration predicts falling among older people: results of a 1year prospective study. J. Neuroeng. Rehabil. 10, 1–6. Duncan, P., Studenski, S., Richards, L., Gollub, S., Lai, S.M., Reker, D., Perera, S., Yates, J., Koch, V., Rigler, S., Johnson, D., 2003. Randomized clinical trial of therapeutic exercise in subacute stroke. Stroke 34, 2173–2180. Fletcher, P.C., Hirdes, J.P., 2002. Risk factors for falling among community-based seniors using home care services. J. Gerontol. A. Biol. Sci. Med. Sci. 57, M504– M510. Gage, H., 1967. Accelerographic analysis of human gait. In: Biomechanics Monograph. ASME United Engineering Center, New York, US, pp. 137–152. Hamacher, D., Singh, N.B., Van Dieën, J.H., Heller, M.O., Taylor, W.R., 2011. Kinematic measures for assessing gait stability in elderly individuals: a systematic review. J. R. Soc. Interface 8, 1682–1698. Holden, M.K., Gill, K.M., Magliozzi, M.R., Nathan, J., Piehl-Baker, L., 1984. Clinical gait assessment in the neurologically impaired. Reliability and meaningfulness. Phys. Ther. 64, 35–40. Huang, H., Wolf, S.L., He, J., 2006. Recent developments in biofeedback for neuromotor rehabilitation. J. Neuroeng. Rehabil. 3, 1–12. Iosa, M., Bini, F., Marinozzi, F., Fusco, A., Morone, G., Koch, G., Martino Cinnera, A., Bonnì, S., Paolucci, S., 2016. Stability and harmony of gait in patients with subacute stroke. J. Med. Biol. Eng. 36, 635–643. Iosa, M., Fusco, A., Morone, G., Pratesi, L., Coiro, P., Venturiero, V., De Angelis, D., Bragoni, M., Paolucci, S., 2012a. Assessment of upper-body dynamic stability during walking in patients with subacute stroke. J. Rehabil. Res. Dev. 49, 439– 450. Iosa, M., Marro, T., Paolucci, S., Morelli, D., 2012b. Stability and harmony of gait in children with cerebral palsy. Res. Dev. Disabil. 33, 129–135. Iosa, M., Paradisi, F., Brunelli, S., Delussu, A.S., Pellegrini, R., Zenardi, D., Paolucci, S., Traballesi, M., 2014. Assessment of gait stability, harmony, and symmetry in subjects with lower-limb amputation evaluated by trunk accelerations. J. Rehabil. Res. Dev. 51, 623–634. Isho, T., Tashiro, H., Usuda, S., 2015. Accelerometry-based gait characteristics evaluated using a smartphone and their association with fall risk in people with chronic stroke. J. Stroke Cerebrovasc. Dis. 24, 1305–1311. Isho, T., Usuda, S., 2016. Association of trunk control with mobility performance and accelerometry-based gait characteristics in hemiparetic patients with subacute stroke. Gait Posture 44, 89–93. Lamb, S.E., Jørstad-Stein, E.C., Hauer, K., Becker, C.Prevention of Falls Network Europe and Outcomes Consensus Group, 2005. Development of a common outcome data set for fall injury prevention trials: the prevention of falls network Europe consensus. J. Am. Geriatr. Soc. 53, 1618–1622.
Langhorne, P., Stott, D.J., Robertson, L., MacDonald, J., Jones, L., McAlpine, C., Dick, F., Taylor, G.S., Murray, G., 2000. Medical complications after stroke: a multicenter study. Stroke 31, 1223–1229. Lowry, K.A., Smiley-Oyen, A.L., Carrel, A.J., Kerr, J.P., 2009. Walking stability using harmonic ratios in Parkinson’s disease. Mov. Disord. 24, 261–267. Mancini, M., Horak, F.B., 2010. The relevance of clinical balance assessment tools to differentiate balance deficits. Eur. J. Phys. Rehabil. Med. 46, 239–248. Marigold, D.S., Patla, A.E., 2008. Age-related changes in gait for multi-surface terrain. Gait Posture 27, 689–696. Masud, T., Morris, R.O., 2001. Epidemiology of falls. Age Ageing 30, 3–7. Mazzà, C., Iosa, M., Pecoraro, F., Cappozzo, A., 2008. Control of the upper body accelerations in young and elderly women during level walking. J. Neuroeng. Rehabil. 5, 1–10. Menz, H.B., Lord, S.R., Fitzpatrick, R.C., 2003a. Acceleration patterns of the head and pelvis when walking on level and irregular surfaces. Gait Posture 18, 35–46. Menz, H.B., Lord, S.R., Fitzpatrick, R.C., 2003b. Acceleration patterns of the head and pelvis when walking are associated with risk of falling in community-dwelling older people. J. Gerontol. Ser. A Biol. Sci. Med. Sci. 58, 446–452. Mizuike, C., Ohgi, S., Morita, S., 2009. Analysis of stroke patient walking dynamics using a tri-axial accelerometer. Gait Posture 30, 60–64. Morone, G., Iosa, M., Pratesi, L., Paolucci, S., 2014. Can overestimation of walking ability increase the risk of falls in people in the subacute stage after stroke on their return home? Gait Posture 39, 965–970. Paolucci, S., Bragoni, M., Coiro, P., De Angelis, D., Fusco, F.R., Morelli, D., Venturiero, V., Pratesi, L., 2008. Quantification of the probability of reaching mobility independence at discharge from a rehabilitation hospital in nonwalking early ischemic stroke patients: a multivariate study. Cerebrovasc. Dis. 26, 16–22. Pasciuto, I., Bergamini, E., Iosa, M., Vannozzi, G., Cappozzo, A., 2017. Overcoming the limitations of the Harmonic Ratio for the reliable assessment of gait symmetry. J. Biomech. 53, 84–89. Perera, S., Mody, S.H., Woodman, R.C., Studenski, S.A., 2006. Meaningful change and responsiveness in common physical performance measures in older adults. J. Am. Geriatr. Soc. 54, 743–749. Raîche, M., Hébert, R., Prince, F., Corriveau, H., 2000. Screening older adults at risk of falling with the Tinetti balance scale. Lancet 356, 1001–1002. Roche, J.L., Lowry, K.A., Vanswearingen, J.M., Brach, J.S., Redfern, M.S., 2013. Harmonic Ratios: a quantification of step to step symmetry. J. Biomech. 46, 828–831. Senden, R., Savelberg, H.H.C.M., Grimm, B., Heyligers, I.C., Meijer, K., 2012. Accelerometry-based gait analysis, an additional objective approach to screen subjects at risk for falling. Gait Posture 36, 296–300. Summa, A., Vannozzi, G., Bergamini, E., Iosa, M., Morelli, D., Cappozzo, A., 2016. Multilevel upper body movement control during gait in children with cerebral palsy. PLoS One 11, e0151792. Tinetti, M.E., Williams, T.F., Mayewski, R., 1986. Fall risk index for elderly patients based on number of chronic disabilities. Am. J. Med. 80, 429–434. WHO, 2001. World Health Organization. International Classification of Functioning, Disability and Health (ICF). World Health Organization, Geneva. Winter, D.A., 1990. Biomechanics and motor control of human movement. John Wiley & Sons, New York, US.
Please cite this article in press as: Bergamini, E., et al. Multi-sensor assessment of dynamic balance during gait in patients with subacute stroke. J. Biomech. (2017), http://dx.doi.org/10.1016/j.jbiomech.2017.07.034
