- Email: [email protected]
P052 Classification of gait patterns in chronic post stroke patients
S80
Abstracts of the 17th Annual Meeting of ESMAC, Poster Presentations / Gait & Posture 28S (2008) S49–S118
groups (p < 0.098, p < 0.158) respectively. Run test results showed that 4 subjects’ signals were non-stationary while 11 subjects had stationary ergodic signals. Discussion: The dynamics of the human postural sway demonstrated both common characteristics in linear behavioral aspects among the groups with some remarkable individual differences. References [1] E. Kreyszig, Advanced Engineering Mathematics, New York: Wiley, 1999. [2] W. Cochran et al., IEEE Transactions on Audio and Electroacoustics, Vol.15, No.2, 1967, pp.45−55. [3] P. D. Welch, IEEE Transactions on Audio and Electroacoustics, Vol. 15, 1967, pp. 70−73. [4] J.S. Bendat and A.G. Piersol, Random Data: Analysis and Measurement Procedures, New York: Wiley, 1971. [5] A.H. Nayfeh and B. Balachandran, Applied Nonlinear Dynamics, New York: Wiley, 1995. [6] D.A. Winter, Biomechanics and Motor Control of Human Movement, New York: Wiley, 1990.
P051 Relation between upper limb function and neuroradiological findings in hemiplegic cerebral palsy H. Feys1 , E. Jaspers2 , M. Eyssen3 , K. Klingels4 , K. Desloovere5 , G. Molenaers6 , P. De Cock7 . 1 Rehabilitation Sciences, Katholieke Universiteit Leuven; 2 Rehabilitation Sciences, PhD fellowship of the Research Foundation Flanders, Katholieke Universiteit Leuven; 3 Pediatrics, University Hospital Leuven; 4 Rehabilitation Sciences, Katholieke Universiteit Leuven; 5 Clinical Motion Analysis Laboratory, University Hospital Pellenberg; 6 Musculoskeletal Sciences, Katholieke Universiteit Leuven; 7 Center for Developmental Disabilities, Katholieke Universiteit Leuven, Belgium Summary: The relation between MRI images and upper limb function was examined in 54 children with hemiplegic cerebral palsy (CP). The findings emphasize the importance of the timing of the brain insult with regard to prognosis of hand motor abilities. Conclusions: The type of brain lesions may distinguish children with hemiplegic CP regarding upper limb function. In particular, the timing of the brain lesion seems to play an important role. These insights are valuable for prognosis and may contribute to a better delineation of future interventions. Introduction: Although some studies have been published on the relation between lesion type and clinical findings, results were not always convergent. The objective of this study was to examine the relation between neuroradiological findings and upper limb function in children with hemiplegic CP. Patients/Materials and Methods: Fifty-four children with hemiplegic CP (age 5 to 15 years) who were referred to the Clinical Motion Analysis Laboratory (University Hospital Pellenberg) and had available MRI, were included in the study. MRI images were independently reviewed by a neuropediatrician and neuroradiologist, blind to the clinical results. Children were classified into 4 groups: congenital brain malformations (CBM), periventricular lesions (PVL), cortical-subcortical pre- or perinatal lesions (CSC) and postnatally acquired brain lesions (ACQ). This classification corresponds to different types of structural pathologies, depending mainly on the stage of brain development. The involvement of the different neuroanatomical structures was
also scored (absence or presence of a lesion). Clinical assessments included measures of upper limb function (Melbourne Assessment for Unilateral Upper Limb Function, House Classification) and motor impairments (passive range of motion, Modified Ashworth Scale, grip strength). Comparative statistics and correlation analysis were used. Results: According to the classification groups, 3 children had a CBM, 29 a PVL, 14 a CSC and 8 an ACQ lesion. Comparison between the PVL, CSC, and ACQ groups revealed highly significant differences for the Melbourne Assessment (p = 0.003) and House Classification (p = 0.001). Post-hoc tests showed significant differences between the PVL versus CSC and ACQ groups (p < 0.05), in favor of the children with PVL. Differences between the CSC and ACQ groups were not significant, although trends toward a worse performance in the ACQ group were found. Comparison of motor impairments revealed a significant lower muscle tone (p = 0.04) and higher grip strength in PVL lesions (p = 0.002). Within the PVL group, significantly better results were found in children with a pure lesion compared to mixed lesions. Correlation analysis additionally showed that middle cerebral artery infarctions and basal ganglia/thalamic lesions were significantly correlated to a worse performance (values of coefficients 0.32−0.54). Discussion: Children in the PVL group displayed a moderate to good upper limb function and use of the hemiplegic hand and performed significantly better compared to both the children in the CSC and ACQ group. Comparison of the CSC group and ACQ group showed a clear trend for a worse performance in the latter group. These discrepancies in upper limb function emphasize the importance of the timing of the brain insult with regard to prognosis of hand motor abilities. However part of the variance in upper limb function is not yet explained. Further studies using advanced brain evaluation techniques in combination with thorough clinical evaluations (e.g. three-dimensional kinematic analysis), might improve our understanding of the complex interaction between the central nervous system and the hemiplegic child’s upper limb abilities. P052 Classification of gait patterns in chronic post stroke patients C. Huenaerts1 , G. Molenaers2 , H. Beyens2 , K. Daniels1 , K. Desloovere1 . 1 Clinical motion analysis laboratory, University Hospital Pellenberg; 2 Department of musculoskeletal sciences, University Hospital Pellenberg, Belgium Summary: The results of gait analysis data showed that gait after stroke deviates strongly from the normal gait and presents a variety of patterns. In this study of 42 patients, three different patterns could be observed. Conclusions: Gait following stroke differs significantly from normal gait. Three different patterns could be observed. The knee extension thrust pattern (N = 19) and flexed pattern (N = 16) resulted in abnormal kinematics and kinetics at all levels and a lower walking speed, while for the mild group (N = 7) gait pathology was limited to significant deviated ankle kinematics and kinetics, with almost normal knee motion. Introduction: Many authors described gait abnormalities, as well as classifications of gait patterns after acute and chronic stroke [1−3]. The aim of this study was (1) to define the different
Abstracts of the 17th Annual Meeting of ESMAC, Poster Presentations / Gait & Posture 28S (2008) S49–S118 gait patterns (known from literature) in chronic stroke, based on qualitative manual template matching of gait analysis plots and (2) to document objectively the characteristic gait parameters in those patterns. Patients/Materials and Methods: 42 patients with chronic cerebrovascular accident (CVA) (23L, 19R) with a mean age of 55 years (±13.3) were included in this study, as well as 10 agerelated control subjects. Mean time since CVA was 5.2 years (±3.9 y). Lesions were classified as an infarct in 23 patients and a haemorrhage in 11 patients. In 8 patients, details of the lesion were not known. All subjects underwent a lower limb 3D gait analysis, with or without walking aids, including kinematic and kinetic data (8 camera VICON system and 2 AMTI forceplates), EMG of 8 lower extremity muscle groups and a full clinical evaluation. Three representative trials were selected bilaterally. For each subject, 58 relevant gait parameters were defined and mean data and SD of these parameters were calculated. ANOVA analysis (post hoc Kruskal Wallis) was used for comparison between pathological and normal reference data and to examine the overall differences in parameters between observed gait patterns. Results: Gait following stroke significantly deviated form normal walking at all levels; including reduced walking speed and step length, decreased hip ROM in the sagittal plane, decreased knee flexion in swing and decreased power generation at hip, knee and ankle (p < 0.0005). Based on literature and visual inspection, the knee seemed to be the most discriminating factor between patterns. Three different gait patterns could then be observed: (1) knee extension thrust pattern (N = 19), (2) flexed pattern (N = 16) and (3) mild pattern (N = 7). The latter had normal knee flexion during loading response and showed no significant differences in hip and knee kinematics and kinetics compared to normal. Group 1 was characterized by a knee extension thrust pattern leading to knee hyperextension (in 17/19 patients) and an early knee flexion moment in stance. Knee flexion during loading response was absent (p < 0.0001). The second group showed increased knee flexion at initial contact, decreased maximum knee-extension in stance and decreased knee flexion moment (p < 0.005). Significant decreased hip ROM in the sagittal plane and decreased peak knee flexion in swing was seen in group 1 and 2 (p < 0.0001). At the ankle, significant lower ROM at push-off was seen in all three groups. In the first and second pattern, a significant decreased walking speed (p < 0.0001) was observed, whereas the mild group showed only slightly lower velocity compared to normal. Circumduction, as a compensation for reduced foot clearance, was most often seen in group 1 (18/19) and group 2 (10/16). Pelvic hiking was observed for 17 out of 19 patients in group 1 and 11 out of 16 patients in group 2. In group 3, 3 out of 7 patients showed circumduction and pelvic hiking. Discussion: The differences between chronic post stroke gait and normal gait found in this study were consistent with those found in literature [1]. The three different patterns could be compared to some of the patterns described by Mulroy and Kinsella. They both found a flexed and an extended group with slow walking speed. The mild group could be compared to their ‘fast walking’ group, with a more normal gait compared to the other subgroups.
P053 Accuracy requirements in BSIP estimations for analysis of normal gait K.S. Erer, E. T¨on¨uk, S.T. T¨umer. Mechanical Engineering, Middle East Technical University, Turkey Summary: This work investigates the role played by accuracy in body segment inertial parameter (BSIP) estimations in analysis results of normal human gait. The findings suggest that accuracy in estimations is not a critical factor owing to dominance of ground reactions. Conclusions: In normal human gait, ground reactions are so dominant over inertial effects that any BSIP estimation method that produces sufficiently accurate estimates can be employed without the kinetic results getting appreciably affected. Introduction: Inverse dynamics analysis of human gait requires BSIP values for the calculation of kinetic joint variables. While it is true that estimation and solution accuracies are positively correlated, it is questionable whether it is justified to employ advanced estimation techniques when time and labour devoted to the process is considered, especially in the unavoidable presence of other error sources such as noise contaminated marker trajectories [1]. A way for evaluating accuracy requirements might be to compare kinetic results derived by solving dynamic equilibrium equations with those obtained by artificially satisfying static equilibrium conditions. Conclusions applicable to this specific case may then be generalized.
References [1] Kim et al, Gait and Posture 20(2004) 140–146. [2] Mulroy et al, Gait and Posture 18(2003) 114–125. [3] Kinsella et al, Gait and Posture 27(2008) 144–151.
S81
Figure 1. Variations of joint moments over the gait cycle.
Abstracts of the 17th Annual Meeting of ESMAC, Poster Presentations / Gait & Posture 28S (2008) S49–S118
groups (p < 0.098, p < 0.158) respectively. Run test results showed that 4 subjects’ signals were non-stationary while 11 subjects had stationary ergodic signals. Discussion: The dynamics of the human postural sway demonstrated both common characteristics in linear behavioral aspects among the groups with some remarkable individual differences. References [1] E. Kreyszig, Advanced Engineering Mathematics, New York: Wiley, 1999. [2] W. Cochran et al., IEEE Transactions on Audio and Electroacoustics, Vol.15, No.2, 1967, pp.45−55. [3] P. D. Welch, IEEE Transactions on Audio and Electroacoustics, Vol. 15, 1967, pp. 70−73. [4] J.S. Bendat and A.G. Piersol, Random Data: Analysis and Measurement Procedures, New York: Wiley, 1971. [5] A.H. Nayfeh and B. Balachandran, Applied Nonlinear Dynamics, New York: Wiley, 1995. [6] D.A. Winter, Biomechanics and Motor Control of Human Movement, New York: Wiley, 1990.
P051 Relation between upper limb function and neuroradiological findings in hemiplegic cerebral palsy H. Feys1 , E. Jaspers2 , M. Eyssen3 , K. Klingels4 , K. Desloovere5 , G. Molenaers6 , P. De Cock7 . 1 Rehabilitation Sciences, Katholieke Universiteit Leuven; 2 Rehabilitation Sciences, PhD fellowship of the Research Foundation Flanders, Katholieke Universiteit Leuven; 3 Pediatrics, University Hospital Leuven; 4 Rehabilitation Sciences, Katholieke Universiteit Leuven; 5 Clinical Motion Analysis Laboratory, University Hospital Pellenberg; 6 Musculoskeletal Sciences, Katholieke Universiteit Leuven; 7 Center for Developmental Disabilities, Katholieke Universiteit Leuven, Belgium Summary: The relation between MRI images and upper limb function was examined in 54 children with hemiplegic cerebral palsy (CP). The findings emphasize the importance of the timing of the brain insult with regard to prognosis of hand motor abilities. Conclusions: The type of brain lesions may distinguish children with hemiplegic CP regarding upper limb function. In particular, the timing of the brain lesion seems to play an important role. These insights are valuable for prognosis and may contribute to a better delineation of future interventions. Introduction: Although some studies have been published on the relation between lesion type and clinical findings, results were not always convergent. The objective of this study was to examine the relation between neuroradiological findings and upper limb function in children with hemiplegic CP. Patients/Materials and Methods: Fifty-four children with hemiplegic CP (age 5 to 15 years) who were referred to the Clinical Motion Analysis Laboratory (University Hospital Pellenberg) and had available MRI, were included in the study. MRI images were independently reviewed by a neuropediatrician and neuroradiologist, blind to the clinical results. Children were classified into 4 groups: congenital brain malformations (CBM), periventricular lesions (PVL), cortical-subcortical pre- or perinatal lesions (CSC) and postnatally acquired brain lesions (ACQ). This classification corresponds to different types of structural pathologies, depending mainly on the stage of brain development. The involvement of the different neuroanatomical structures was
also scored (absence or presence of a lesion). Clinical assessments included measures of upper limb function (Melbourne Assessment for Unilateral Upper Limb Function, House Classification) and motor impairments (passive range of motion, Modified Ashworth Scale, grip strength). Comparative statistics and correlation analysis were used. Results: According to the classification groups, 3 children had a CBM, 29 a PVL, 14 a CSC and 8 an ACQ lesion. Comparison between the PVL, CSC, and ACQ groups revealed highly significant differences for the Melbourne Assessment (p = 0.003) and House Classification (p = 0.001). Post-hoc tests showed significant differences between the PVL versus CSC and ACQ groups (p < 0.05), in favor of the children with PVL. Differences between the CSC and ACQ groups were not significant, although trends toward a worse performance in the ACQ group were found. Comparison of motor impairments revealed a significant lower muscle tone (p = 0.04) and higher grip strength in PVL lesions (p = 0.002). Within the PVL group, significantly better results were found in children with a pure lesion compared to mixed lesions. Correlation analysis additionally showed that middle cerebral artery infarctions and basal ganglia/thalamic lesions were significantly correlated to a worse performance (values of coefficients 0.32−0.54). Discussion: Children in the PVL group displayed a moderate to good upper limb function and use of the hemiplegic hand and performed significantly better compared to both the children in the CSC and ACQ group. Comparison of the CSC group and ACQ group showed a clear trend for a worse performance in the latter group. These discrepancies in upper limb function emphasize the importance of the timing of the brain insult with regard to prognosis of hand motor abilities. However part of the variance in upper limb function is not yet explained. Further studies using advanced brain evaluation techniques in combination with thorough clinical evaluations (e.g. three-dimensional kinematic analysis), might improve our understanding of the complex interaction between the central nervous system and the hemiplegic child’s upper limb abilities. P052 Classification of gait patterns in chronic post stroke patients C. Huenaerts1 , G. Molenaers2 , H. Beyens2 , K. Daniels1 , K. Desloovere1 . 1 Clinical motion analysis laboratory, University Hospital Pellenberg; 2 Department of musculoskeletal sciences, University Hospital Pellenberg, Belgium Summary: The results of gait analysis data showed that gait after stroke deviates strongly from the normal gait and presents a variety of patterns. In this study of 42 patients, three different patterns could be observed. Conclusions: Gait following stroke differs significantly from normal gait. Three different patterns could be observed. The knee extension thrust pattern (N = 19) and flexed pattern (N = 16) resulted in abnormal kinematics and kinetics at all levels and a lower walking speed, while for the mild group (N = 7) gait pathology was limited to significant deviated ankle kinematics and kinetics, with almost normal knee motion. Introduction: Many authors described gait abnormalities, as well as classifications of gait patterns after acute and chronic stroke [1−3]. The aim of this study was (1) to define the different
Abstracts of the 17th Annual Meeting of ESMAC, Poster Presentations / Gait & Posture 28S (2008) S49–S118 gait patterns (known from literature) in chronic stroke, based on qualitative manual template matching of gait analysis plots and (2) to document objectively the characteristic gait parameters in those patterns. Patients/Materials and Methods: 42 patients with chronic cerebrovascular accident (CVA) (23L, 19R) with a mean age of 55 years (±13.3) were included in this study, as well as 10 agerelated control subjects. Mean time since CVA was 5.2 years (±3.9 y). Lesions were classified as an infarct in 23 patients and a haemorrhage in 11 patients. In 8 patients, details of the lesion were not known. All subjects underwent a lower limb 3D gait analysis, with or without walking aids, including kinematic and kinetic data (8 camera VICON system and 2 AMTI forceplates), EMG of 8 lower extremity muscle groups and a full clinical evaluation. Three representative trials were selected bilaterally. For each subject, 58 relevant gait parameters were defined and mean data and SD of these parameters were calculated. ANOVA analysis (post hoc Kruskal Wallis) was used for comparison between pathological and normal reference data and to examine the overall differences in parameters between observed gait patterns. Results: Gait following stroke significantly deviated form normal walking at all levels; including reduced walking speed and step length, decreased hip ROM in the sagittal plane, decreased knee flexion in swing and decreased power generation at hip, knee and ankle (p < 0.0005). Based on literature and visual inspection, the knee seemed to be the most discriminating factor between patterns. Three different gait patterns could then be observed: (1) knee extension thrust pattern (N = 19), (2) flexed pattern (N = 16) and (3) mild pattern (N = 7). The latter had normal knee flexion during loading response and showed no significant differences in hip and knee kinematics and kinetics compared to normal. Group 1 was characterized by a knee extension thrust pattern leading to knee hyperextension (in 17/19 patients) and an early knee flexion moment in stance. Knee flexion during loading response was absent (p < 0.0001). The second group showed increased knee flexion at initial contact, decreased maximum knee-extension in stance and decreased knee flexion moment (p < 0.005). Significant decreased hip ROM in the sagittal plane and decreased peak knee flexion in swing was seen in group 1 and 2 (p < 0.0001). At the ankle, significant lower ROM at push-off was seen in all three groups. In the first and second pattern, a significant decreased walking speed (p < 0.0001) was observed, whereas the mild group showed only slightly lower velocity compared to normal. Circumduction, as a compensation for reduced foot clearance, was most often seen in group 1 (18/19) and group 2 (10/16). Pelvic hiking was observed for 17 out of 19 patients in group 1 and 11 out of 16 patients in group 2. In group 3, 3 out of 7 patients showed circumduction and pelvic hiking. Discussion: The differences between chronic post stroke gait and normal gait found in this study were consistent with those found in literature [1]. The three different patterns could be compared to some of the patterns described by Mulroy and Kinsella. They both found a flexed and an extended group with slow walking speed. The mild group could be compared to their ‘fast walking’ group, with a more normal gait compared to the other subgroups.
P053 Accuracy requirements in BSIP estimations for analysis of normal gait K.S. Erer, E. T¨on¨uk, S.T. T¨umer. Mechanical Engineering, Middle East Technical University, Turkey Summary: This work investigates the role played by accuracy in body segment inertial parameter (BSIP) estimations in analysis results of normal human gait. The findings suggest that accuracy in estimations is not a critical factor owing to dominance of ground reactions. Conclusions: In normal human gait, ground reactions are so dominant over inertial effects that any BSIP estimation method that produces sufficiently accurate estimates can be employed without the kinetic results getting appreciably affected. Introduction: Inverse dynamics analysis of human gait requires BSIP values for the calculation of kinetic joint variables. While it is true that estimation and solution accuracies are positively correlated, it is questionable whether it is justified to employ advanced estimation techniques when time and labour devoted to the process is considered, especially in the unavoidable presence of other error sources such as noise contaminated marker trajectories [1]. A way for evaluating accuracy requirements might be to compare kinetic results derived by solving dynamic equilibrium equations with those obtained by artificially satisfying static equilibrium conditions. Conclusions applicable to this specific case may then be generalized.
References [1] Kim et al, Gait and Posture 20(2004) 140–146. [2] Mulroy et al, Gait and Posture 18(2003) 114–125. [3] Kinsella et al, Gait and Posture 27(2008) 144–151.
S81
Figure 1. Variations of joint moments over the gait cycle.