Identification of individual walking patterns using time discrete and time continuous data sets

doi:10.1016/S0966-6362(01)00193-X

Gait & Posture

Volume 15, Issue 2, April 2002, Pages 180-186

https://doi.org/10.1016/S0966-6362(01)00193-X Get rights and content

Abstract

Scientific studies typically treat data by studying effects of groups. Clinical therapy typically treats patients on a subject specific basis. Consequently, scientific and clinical attempts to help patients are often not coordinated. The purposes of this study were (a) to identify subject and group specific locomotion characteristics quantitatively, using time discrete and time continuous data and (b) to assess the advantages and disadvantages of the two approaches. Kinematic and kinetic gait pattern of 13 female subjects walking in dress shoes with different heel heights (14, 37, 54 and 85 mm) were analysed. The results of this study showed that subject specific gait characteristics could be better identified with the time continuous than with the time discrete approach. Thus, the time continuous approach using artificial networks is an effective tool for identifying subject and group specific locomotion characteristics.

Introduction

During the 20th century, quantitative gait analysis has been developed and used to support clinicians’ and therapists’ decisions when assessing gait abnormalities and/or identifying changes due to orthopaedic or physiotherapeutic interventions. Typically, sets of time discrete variables have been used to predict prevention of injuries and/or assessment of rehabilitation procedures. Furthermore, artificial neural network applications have been developed for gait analysis to distinguish between healthy and pathological walking patterns [1], [2] using an external teacher to ‘train’ the networks.

Conventional gait analysis for clinical or general applications is associated with some inherent problems. First, conventional gait analysis uses typically time discrete variables. However, everyday experiences indicate that it is difficult to recognize a known person from a distance as long as he/she is standing. However, as soon as the person starts moving, the recognition is vastly improved, probably because the observer assesses the movement pattern of the observed person [3], [4]. In general, time discrete and time continuous data sets provide different information to the observer.

Second, conventional gait analysis reproduces only classifications or behaviours of groups that are expected by the gait analysists. Subjects are divided into normal and excessive pronators, for instance. However, groupings that are not pre-programmed by the project manager are typically not detected using this approach.

Third, conceptual differences between the scientific and the clinical approach influence conventional gait analysis. The scientific approach tends to be more group-oriented. The clinical approach tends to be more subject-oriented. However, the ability of recognizing and/or modifying subject specific movement patterns is of particular interest for successful and effective therapeutic interventions.

A novel approach of movement analysis using a time course oriented analysis with non-linear classifying self organized neural networks has been presented recently [5]. It addressed the outlined shortcomings of the conventional movement analysis and was able to identify individual movement patterns for various types of sports movements [6], [7]. Thus, it seems reasonable to assume that this movement analysis strategy could also be used in clinical situations where group and subject specific aspects are of importance.

Thus, the purposes of this study were

1.
to determine individual movement characteristics using time discrete data,
2.
to determine individual movement characteristics using time course oriented data,
3.
to identify groups with similar movement characteristics using the time discrete and time course oriented data and
4.
to compare the two approaches in their ability in identifying individual and/or group movement characteristics.

Section snippets

Data acquisition

The data collection for this study has been described in detail elsewhere [8]. Shortly, 13 female subjects (164.1±5.6 cm, 67.7±12.3 kg, 40.6±8.3 years), which were familiar with walking in high heels and did not suffer from recent lower extremity pain, were recruited for this investigation and signed informed consent forms. Four types of commercially available dress shoes were chosen for this study. They were of similar construction with the main difference in the height of the heel (14, 37, 54

Results

Both approaches, the time discrete as well as the time course oriented approach, identified individual gait patterns when pattern recognition was applied separately for each heel height. The ARs were in both cases close to 100% (Table 1, Table 2).

For the time discrete approach, the identification of individual gait characteristics was highest when using all variables together (Table 1, column ‘All’) but almost as high when using the minimum parameter (column ‘Minima’). A decrease of the AR was

Discussion

In the present investigation subject specific gait characteristics could be identified by means of time discrete and time continuous approaches. The recognition rates of 93.1–100% for individuals with respect to each heel height condition confirm the previous results for the individuality of specific movement patterns found for running, discus and javelin throwing [5], [6], [7]. The probability of finding all trials of each subject in separate clusters by random is extremely small. Therefore,

Conclusion

The results of this study have demonstrated that gait pattern contain general and individual information. Furthermore, the results of this study have shown that a time continuous data analysis approach using artificial neural networks provides an effective tool for identifying the subject specific and general characteristics for individuals and groups. The results of this study documented that the continuous data approach is more powerful in predicting subject specific and group specific

References (24)

W. Berger et al.
Pathophysiology of gait in children with cerebral palsy
EEG Clin. Neurophys.
(1982)
E. Schneider et al.
Fourier analysis of ground reaction forces in normals and patients with knee joint disease
J. Biomech.
(1983)
S.H. Holzreiter et al.
Assessment of gait patterns using neural networks
J. Biomech.
(1995)
K. Aminian et al.
Level, downhill and uphill walking identification using neural networks
Electr. Lett.
(1993)
G. Johansson
Visual perception of biological motion and a model for its analysis
Percept. Psychophys.
(1973)
J.E. Cutting et al.
Recognizing friends by their walk: gait perception without familiarity cues
Bull. Psychonomic. Soc.
(1977)
H.-U. Bauer et al.
Self-organizing maps for the analysis of complex movement patterns
Neur. Proc. Lett.
(1997)
W.I. Schöllhorn et al.
Identifying individual movement styles in high performance sports by means of self-organizing Kohonen maps
W.I. Schöllhorn et al.
Erkennung von individuellen Laufmustern mit Hilfe von neuronalen Netzen (recognition of individual running patterns by means of artificial neural nets)
Stefanyshyn D, Nigg B, Fisher V, O'Flynn B, Liu W. The influence of high heeled shoes on kinematics, kinetics and...

K.G. Holt et al.

Energetic cost and stability during human walking at the preferred stride frequency

J. Motor Behav.

(1995)

Ch Fink et al.

A more pragmatic critria for filtering data

Cited by (109)

Improved weight bearing during gait at 6 weeks post-surgery with an angle stable locking system after distal tibial fracture
2024, Gait and Posture
Functional recovery after intramedullary nailing of distal tibial fractures can be monitored using ipsilateral vertical ground reaction forces (vGRF), giving insight into recovery of patients’ gait symmetry. Previous work compared patient cohorts to healthy controls, but it remains unclear if these metrics can identify treatment-based differences in return to function post-surgery.
Is treatment of a distal tibial fracture with intramedullary nailing with an angle stable locking system (ASLS) associated with higher ipsilateral vGRF and improved symmetry compared to conventional intramedullary nailing at an early time point?
Thirty-nine patients treated with ASLS intramedullary nailing were retrospectively compared to thirty-nine patients with conventional locking. vGRFs were collected at 1, 6, 12, 26, and 52 weeks post-surgery during standing and gait. Discrete metrics of ipsilateral vGRF (maximal force, impulse) and asymmetry were compared between treatments at each time point. Time-scale comparisons of ipsilateral vGRF and lower limb asymmetry were additionally performed for gait trials. Mann-Whitney Test or a two-way analysis of variance tested discrete comparisons; statistical non-parametric mapping tested time-scale data between treatment groups.
During gait, ASLS-treated patients applied more load on the operated limb (17–38% stance, p = 0.015) and consequently loaded limbs more symmetrically (8–37% stance, p = 0.008) during the loading response at 6 weeks post-surgery compared to conventional IM treatment. Discrete measures of symmetry at the same time point identified treatment-based differences in maximal force (p = 0.039) and impulse (p = 0.012), with ASLS-treated patients exhibiting more symmetry. No differences were identified in gait trials at later time points nor from all standing trials.
During the initial loading response of gait, increased ipsilateral vGRF and improved weightbearing symmetry were identified in ASLS patients at 6 weeks post-surgery compared to conventional IM nailing. Early and objective metrics of dynamic movement are suggested to identify treatment-based differences in functional recovery.
Reference values and functional descriptions of transverse plane spinal dynamics during gait based on surface topography
2023, Human Movement Science
Spinal dynamics during gait have been of interest in research for many decades. Based on respective previous investigations, the pelvis is generally expected to be maximally forward rotated on the side of the reference leg at the beginning of each gait cycle and to reach its maximum counterrotation approximately at the end of the reference leg’s stance phase. The pelvic–upper-thoracic-spine coordination converges towards an anti-phase movement pattern in high velocities during ambulation. The vertebral bodies around the seventh thoracic vertebra are considered to be an area of transition during human ambulation where no or at least little rotary motion can be observed. The respective cranial and caudal vertebrae meanwhile are expected to rotate conversely around this spinal point of intersection. However, these previous assumptions are based on scarce existing research, whereby only isolated vertebrae have been analyzed contemporaneously. Due to huge methodological differences in data capturing approaches, the results are additionally hardly comparable to each other and involved measurement procedures are often not implementable in clinical routines. Furthermore, none of the above-mentioned methods provided reference data for spinal motion during gait based on an appropriate number of healthy participants. Hence, the aim of this study was to present such reference data for spinal rotary motion of every vertebral body from C7 down to L4 and the pelvis derived from surface topographic back shape analyses in a cohort of 201 healthy participants walking on a treadmill at a given walking speed of 5 km/h. Additionally, the spine‘s functional movement behavior during gait should be described in the transverse plane based on data derived from this noninvasive, clinically suitable measurement approach and, in conclusion, the results shall be compared against those of previous research findings derived from other measurement techniques. Contrary to the previous functional understanding, the area of the mid-thoracic spine was found to demonstrate the largest amplitude of rotary motion of all investigated vertebrae and revealed an approximately counterrotated movement behavior compared to the rotary motion of the pelvis. In both directions, spinal rotation during gait seemed to be initiated by the pelvis. The overlying vertebrae followed in succession in the sense of an ongoing movement. Therefore, the point of intersection was not statically located in a specific anatomical section of the spine. Instead, it was found to be dynamic, ascending from one vertebra to the next from caudal to cranial in dependence of the pelvis’s rotation initiation.
Modeling biological individuality using machine learning: A study on human gait
2023, Computational and Structural Biotechnology Journal
Human gait is a complex and unique biological process that can offer valuable insights into an individual’s health and well-being. In this work, we leverage a machine learning-based approach to model individual gait signatures and identify factors contributing to inter-individual variability in gait patterns. We provide a comprehensive analysis of gait individuality by (1) demonstrating the uniqueness of gait signatures in a large-scale dataset and (2) highlighting the gait characteristics that are most distinctive to each individual. We utilized the data from three publicly available datasets comprising 5368 bilateral ground reaction force recordings during level overground walking from 671 distinct healthy individuals. Our results show that individuals can be identified with a prediction accuracy of 99.3% by using the bilateral signals of all three ground reaction force components, with only 10 out of 1342 recordings in our test data being misclassified. This indicates that the combination of bilateral ground reaction force signals with all three components provides a more comprehensive and accurate representation of an individual’s gait signature. The highest accuracy was achieved by (linear) Support Vector Machines (99.3%), followed by Random Forests (98.7%), Convolutional Neural Networks (95.8%), and Decision Trees (82.8%). The proposed approach provides a powerful tool to better understand biological individuality and has potential applications in personalized healthcare, clinical diagnosis, and therapeutic interventions.
Can different training methods reduce the kinematic risk factors of ACL injuries in children?
2023, Journal of Biomechanics
According to the research, a lack of neuromuscular control is a major cause of non-contact anterior cruciate ligament (ACL) injury during locomotion. This study aimed to determine the influence of various prescriptive and Constrained Led Approach (CLA) training approaches on lower extremity kinematics and stride length in children aged 3–5 years old while walking and running. Thirty-six children with a mean age of 4.79 years were separated into three groups: 1- prescriptive training group (n = 10), 2- CLA training group (n = 11), and 3- Control group (n = 10). The kinematics of the hip, knee and ankle joints in the sagittal plane at the moment of heel contact and toe-off were recorded before and after six weeks of intervention. According to the MANOVA, there was no statistically significant difference between the two training techniques in the joint angles at heel contact and toe-off during walking and running after intervention (p ≥ 0.05). However, there was a significant difference in the kinematic characteristics of walking and running between the training and the control groups (p ≤ 0.05). The two training techniques showed a statistically significant difference in stride length during running (p ≤ 0.05). The results indicated that prescriptive and CLA training are effective at altering the kinematics and distance factors underlying children's walking and running abilities.
Functional data analysis reveals asymmetrical crank torque during cycling performed at different exercise intensities
2021, Journal of Biomechanics
Pedaling asymmetry is claimed as a factor of influence on injury and performance. However, the evidence is still controversial. Most previous studies determined peak torque asymmetries, which in our understanding does not consider the pattern of movement like torque profiles. Here we demonstrate that asymmetries in pedaling torque at different exercise intensities can be better described when the torque profiles are considered using functional analysis of variance than when only the peak values are analyzed. We compared peak torques and torque curves recorded while cyclists pedaled at submaximal intensities of 60%, 80%, and 95% of the maximal power output and compared data between the preferred and non-preferred legs. ANOVA showed symmetry or rather no difference in the amount of peak torque between legs, regardless of pedaling intensity. FANOVA, on the other hand, revealed significant asymmetries between legs, regardless of cycling intensity, apparently for different sections of the cycle, however, not for peak torque, either. We conclude that pedaling asymmetry cannot be quantified solely by peak torques and considering the analysis of the entire movement cycle can more accurately reflect the biomechanical movement pattern. Therefore, FANOVA data analysis could be an alternative to identify asymmetries. A novel approach as described here might be useful when combining kinetics assessment with other approaches like EMG and kinematics and help to better understand the role of pedaling asymmetries for performance and injury risks.
Unveiling the acute neurophysiological responses to strength training: An exploratory study on novices performing weightlifting bouts with different motor learning models
2024, Biology of Sport

View all citing articles on Scopus

View full text

Identification of individual walking patterns using time discrete and time continuous data sets

Abstract

Introduction

Section snippets

Data acquisition

Results

Discussion

Conclusion

EEG Clin. Neurophys.

J. Biomech.

Assessment of gait patterns using neural networks

J. Biomech.

Level, downhill and uphill walking identification using neural networks

Electr. Lett.

Visual perception of biological motion and a model for its analysis

Percept. Psychophys.

Recognizing friends by their walk: gait perception without familiarity cues

Bull. Psychonomic. Soc.

Self-organizing maps for the analysis of complex movement patterns

Neur. Proc. Lett.

Identifying individual movement styles in high performance sports by means of self-organizing Kohonen maps

Erkennung von individuellen Laufmustern mit Hilfe von neuronalen Netzen (recognition of individual running patterns by means of artificial neural nets)

Energetic cost and stability during human walking at the preferred stride frequency

J. Motor Behav.

A more pragmatic critria for filtering data