Systematic review of automatic post-stroke gait classification systems.

BACKGROUND: Gait classification is a clinically helpful task performed after a stroke in order to guide rehabilitation therapy. Gait disorders are commonly identified using observational gait analysis in clinical settings, but this approach is limited due to low reliability and accuracy. Data-driven gait classification can quantify gait deviations and categorise gait patterns automatically possibly improving reliability and accuracy; however, the development and clinical utility of current data driven systems has not been reviewed previously. RESEARCH QUESTION: The purpose of this systematic review is to evaluate the literature surrounding the methodology used to develop automatic gait classification systems, and their potential effectiveness in the clinical management of stroke-affected gait. METHOD: The database search included PubMed, IEEE Xplore, and Scopus. Twenty-one studies were identified through inclusion and exclusion criteria from 407 available studies published between 2015 and 2022. Development methodology, classification performance, and clinical utility information were extracted for review. RESULTS AND SIGNIFICANCE: Most of gait classification systems reported a classification accuracy between 80%-100%. However, collated studies presented methodological errors in machine learning (ML) model development. Further, many studies neglected model components such as clinical utility (e.g., predictions don't assist clinicians or therapists in making decisions, interpretability, and generalisability). We provided recommendations to guide development of future post-stroke automatic gait classification systems to better assist clinicians and therapists. Future automatic gait classification systems should emphasise the clinical significance and adopt a standardised development methodology of ML model.

Evaluación funcional y de factores limitantes del tratamiento de los trastornos de la marcha en la parálisis cerebral infantil: desarrollo del sistema de clasificación de niveles de deambulación funcional / Evaluation of function and limiting factors of gait disorder treatment in cerebral palsy: development of the Walking Abilities Levels Classification System

INTRODUCCIÓN: El Gross Motor Function Classification System ha permitido estratificar, según su habilidad para caminar, a los pacientes que padecen parálisis cerebral infantil. La falta de sensibilidad en la detección de cambios y la ausencia de una evaluación del paciente en el contexto en el que se encuentra justifican la búsqueda de alternativas de evaluación pretratamiento. OBJETIVOS. Presentar y mostrar la concordancia interobservador inicial del sistema de clasificación de niveles de deambulación funcional. Con él se evalúa la destreza para caminar y la necesidad de asistencia para realizar transferencias desde la silla de ruedas, y, posteriormente se analiza el escenario que la salud y el entorno del paciente ofrecen como condicionantes en la corrección de la marcha o la bipedestación asistida. SUJETOS Y MÉTODOS: Se describe un nuevo marco de evaluación, elaborado por un grupo interdisciplinar con más de 15 años de experiencia media, enfocado inicialmente a la toma de decisiones antes de un tratamiento quirúrgico. Como control interno, 14 participantes evaluaron la historia clínica y los vídeos de marcha de 10 casos. RESULTADOS: Se alcanzó un índice kappa de acuerdo de 0,76 en niveles funcionales y de 0,79 en el tipo de escenario biológico, de 0,69 en el psicológico y de 0,64 en el social. CONCLUSIONES: El sistema de clasificación de niveles de deambulación funcional ofrece un marco para la evaluación conjunta de la deambulación y de los factores limitantes en la eficacia de un tratamiento. La concordancia interobservador avala iniciar su validación

INTRODUCTION: The Gross Motor Function Classification System has allowed us to stratificate cerebral palsy patients, according to their walking abilities. The lack of sensitivity about detecting changes and the absence of a global patient evaluation, justify the search of new pre-operative evaluation tools. AIMS. To present the Walking Abilities Levels Classification System (WALCS) and to show the first inter-observer agreement study that has been carried out. This system uses first a different pattern for ordering gait functional skills, and after that, evaluates the reversibility of the contextual factors that may limit the result of a gait disorder treatment. SUBJECTS AND METHODS: A new evaluation frame was built by an interdisciplinary team with an average professional experience of more than 15 years, initially focused as part of the pre-surgical patient evaluation. An inter-observer agreement study was held to gain the first insight of it. 14 participants studied the medical reports and gait lab video images of 10 cases. RESULTS: The kappa index was 0.76 for the walking ability level, 0.79 for the biological type, 0.69 psychological type and 0.64 social type of limiting factors. CONCLUSIONS: The WALCS offers a new evaluation frame gathering patient walking skills and limiting factors treatment. The initial inter-observer agreement rate endorsed more intra- and inter-studies in order to achieve a more robust validation

Classification of Parkinson's disease and essential tremor based on balance and gait characteristics from wearable motion sensors via machine learning techniques: a data-driven approach.

BACKGROUND: Parkinson's disease (PD) and essential tremor (ET) are movement disorders that can have similar clinical characteristics including tremor and gait difficulty. These disorders can be misdiagnosed leading to delay in appropriate treatment. The aim of the study was to determine whether balance and gait variables obtained with wearable inertial motion sensors can be utilized to differentiate between PD and ET using machine learning. Additionally, we compared classification performances of several machine learning models. METHODS: This retrospective study included balance and gait variables collected during the instrumented stand and walk test from people with PD (n = 524) and with ET (n = 43). Performance of several machine learning techniques including neural networks, support vector machine, k-nearest neighbor, decision tree, random forest, and gradient boosting, were compared with a dummy model or logistic regression using F1-scores. RESULTS: Machine learning models classified PD and ET based on balance and gait characteristics better than the dummy model (F1-score = 0.48) or logistic regression (F1-score = 0.53). The highest F1-score was 0.61 of neural network, followed by 0.59 of gradient boosting, 0.56 of random forest, 0.55 of support vector machine, 0.53 of decision tree, and 0.49 of k-nearest neighbor. CONCLUSIONS: This study demonstrated the utility of machine learning models to classify different movement disorders based on balance and gait characteristics collected from wearable sensors. Future studies using a well-balanced data set are needed to confirm the potential clinical utility of machine learning models to discern between PD and ET.

Therapy of Parkinson's Disease Subtypes.

Early descriptions of subtypes of Parkinson's disease (PD) are dominated by the approach of predetermined groups. Experts defined, from clinical observation, groups based on clinical or demographic features that appeared to divide PD into clinically distinct subsets. Common bases on which to define subtypes have been motor phenotype (tremor dominant vs akinetic-rigid or postural instability gait disorder types), age, nonmotor dominant symptoms, and genetic forms. Recently, data-driven approaches have been used to define PD subtypes, taking an unbiased statistical approach to the identification of PD subgroups. The vast majority of data-driven subtyping has been done based on clinical features. Biomarker-based subtyping is an emerging but still quite undeveloped field. Not all of the subtyping methods have established therapeutic implications. This may not be surprising given that they were born largely from clinical observations of phenotype and not in observations regarding treatment response or biological hypotheses. The next frontier for subtypes research as it applies to personalized medicine in PD is the development of genotype-specific therapies. Therapies for GBA-PD and LRRK2-PD are already under development. This review discusses each of the major subtyping systems/methods in terms of its applicability to therapy in PD, and the opportunities and challenges designing clinical trials to develop the evidence base for personalized medicine based on subtypes.

[Effectiveness of robotic assistance for gait training in children with cerebral palsy. a systematic review]. / Efectividad de la asistencia robótica para el entrenamiento de la marcha en niños con parálisis cerebral. Una revisión sistemática.

INTRODUCTION: In recent years, the use of gait training using robotic assistance systems has progressively increased in the paediatric population with cerebral palsy. OBJECTIVE: To systematically assess the effects of robotic assistance for gait training compared with physical rehabilitation therapy in children with cerebral palsy (CP), based on the International Classification of Functioning, Health and Disability (ICF). MATERIALS AND METHODS: A systematic review was carried out according to the recommendations of the Cochrane Collaboration. We included randomised or quasi-randomised clinical trials that analysed children with CP classified according to The Gross Motor Function Classification System (GMFCS) I-III. The search was carried out in PubMed, PEDro, CENTRAL, CINALH, Cochrane, Embase, Europe PMC, LILACS and Science Direct. The selection and extraction of data from the studies was carried out by two independent researchers. Disagreements were resolved by consensus. A descriptive analysis of the selected studies was performed. Assessment of risk of bias was performed with the Cochrane Collaboration tool. RESULTS: Four studies met the eligibility criteria. Most of the temporal-spatial, kinetic and kinematic parameters of gait were evaluated, all corresponding to the activity component of the ICF. CONCLUSIONS: Due to the methodological variability of the studies, it is not possible to determine whether robot-assisted gait training is effective for treatment in children with CP.

Functional gait disorders, clinical phenomenology, and classification.

BACKGROUND: Functional gait disorders (FGDs) are relatively common in patients presenting for evaluation of a functional movement disorder (FMD). The diagnosis and classification of FGDs is complex because patients may have a primary FGD or a FMD interfering with gait. METHODS: We performed a detailed evaluation of clinical information and video recordings of gait in patients diagnosed with FMDs. RESULTS: We studied a total of 153 patients with FMDs, 68% females, with a mean age at onset of 36.4 years. A primary FGD was observed in 39.2% of patients; among these patients, 13 (8.5%) had an isolated FGD (a gait disorder without other FMDs). FMDs presented in 34% of patients with otherwise normal gait. Tremor was the most common FMD appearing during gait, but dystonia was the most common FMD interfering with gait. Patients with FGD had a higher frequency of slow-hesitant gait, astasia-abasia, bouncing, wide-based gait and scissoring compared with patients with FMDs occurring during gait. Bouncing gait with knee buckling was more frequently observed in patients with isolated FGD (P = 0.017). Patients with FGDs had a trend for higher frequency of wheelchair dependency (P = 0.073) than those with FMDs interfering with gait. CONCLUSIONS: Abnormal gait may be observed as a primary FGD or in patients with other FMDs appearing during gait; both conditions are common and may cause disability.

Development and clinical validation of inertial sensor-based gait-clustering methods in Parkinson's disease.

BACKGROUND: Gait symptoms and balance impairment are characteristic indicators for the progression in Parkinson's disease (PD). Current gait assessments mostly focus on straight strides with assumed constant velocity, while acceleration/deceleration and turning strides are often ignored. This is either due to the set up of typical clinical assessments or technical limitations in capture volume. Wearable inertial measurement units are a promising and unobtrusive technology to overcome these limitations. Other gait phases such as initiation, termination, transitioning (between straight walking and turning) and turning might be relevant as well for the evaluation of gait and balance impairments in PD. METHOD: In a cohort of 119 PD patients, we applied unsupervised algorithms to find different gait clusters which potentially include the clinically relevant information from distinct gait phases in the standardized 4x10 m gait test. To clinically validate our approach, we determined the discriminative power in each gait cluster to classify between impaired and unimpaired PD patients and compared it to baseline (analyzing all straight strides). RESULTS: As a main result, analyzing only one of the gait clusters constant, non-constant or turning led in each case to a better classification performance in comparison to the baseline (increase of area under the curve (AUC) up to 19% relative to baseline). Furthermore, gait parameters (for turning, constant and non-constant gait) that best predict motor impairment in PD were identified. CONCLUSIONS: We conclude that a more detailed analysis in terms of different gait clusters of standardized gait tests such as the 4x10 m walk may give more insights about the clinically relevant motor impairment in PD patients.

Gait-Based Diplegia Classification Using LSMT Networks.

Diplegia is a specific subcategory of the wide spectrum of motion disorders gathered under the name of cerebral palsy. Recent works proposed to use gait analysis for diplegia classification paving the way for automated analysis. A clinically established gait-based classification system divides diplegic patients into 4 main forms, each one associated with a peculiar walking pattern. In this work, we apply two different deep learning techniques, namely, multilayer perceptron and recurrent neural networks, to automatically classify children into the 4 clinical forms. For the analysis, we used a dataset comprising gait data of 174 patients collected by means of an optoelectronic system. The measurements describing walking patterns have been processed to extract 27 angular parameters and then used to train both kinds of neural networks. Classification results are comparable with those provided by experts in 3 out of 4 forms.

Unusual gait disorders: a phenomenological approach and classification.

INTRODUCTION: Gait impairment is a very common problem in clinical practice. Multiple classifications of gait disorders are available based on anatomy, etiology, pathology and phenomenology. These classifications provide a diagnostic guide but do not clearly explain the pathophysiology of some gait disorders, which can sometimes hinder the diagnostic process. In this context, unusual gait disorders become an even more difficult clinical challenge. Areas covered: The scientific and non-scientific literature contains illustrative descriptions of unusual gait disorders based on their predominant signs and/or comparisons with normal and abnormal zoological and folkloric patterns. Unusual gait disorder phenomenology can be carefully deconstructed in order to achieve an integral approach. We present a pragmatic, phenomenological approach to various unusual gait disorders and highlight key features underlying their phenotypes. We also propose unifying terminology to facilitate diagnosis and academic communication. Expert commentary: Advanced gait analysis, neurophysiological and neuroimaging techniques have allowed for us to recognize that locomotion is a complex motor behavior that requires simultaneous integration of multiple neurological and non-neurological systems. A phenomenological approach such as the one proposed in this review could be useful while those objective techniques become more widely available in clinical practice.

Gait festination in parkinsonism: introduction of two phenotypes.

Gait festination is one of the most characteristic gait disturbances in patients with Parkinson's disease or atypical parkinsonism. Although festination is common and disabling, it has received little attention in the literature, and different definitions exist. Here, we argue that there are actually two phenotypes of festination. The first phenotype entails a primary locomotion disturbance, due to the so-called sequence effect: a progressive shortening of step length, accompanied by a compensatory increase in cadence. This phenotype strongly relates to freezing of gait with alternating trembling of the leg. The second phenotype results from a postural control problem (forward leaning of the trunk) combined with a balance control deficit (inappropriately small balance-correcting steps). In this viewpoint, we elaborate on the possible pathophysiological substrate of these two phenotypes of festination and discuss their management in daily clinical practice.

The German version of the Functional Walking Categories (FWC): translation and initial validation.

BACKGROUND: Community ambulation is often affected after a stroke. However, no validated assessment in German to measure community ambulation on a participation level exists. OBJECTIVES: The purpose was to translate and cross-culturally adapt the Functional Walking Categories (FWC) into German and to assess its validity and reliability in patients with stroke. METHODS: Cross-cultural adaptation guidelines were used for translation. Face and content validity were established with the aid of an expert committee. A pilot study with patients after stroke in a neurological rehabilitation setting checked for concurrent validity using Kendall's tau and reliability using intraclass correlation coefficients. RESULTS: The results indicated that the German version of the FWC has adequate face and content validity. A total of 30 patients (mean age 62 ± 12.315 years, 56.7% female) participated in the study. The FWC correlated well with the Functional Ambulation Categories (tau-b = 0.783), cadence (tau-b = 0.640), gait velocity (tau-b = 0.628), the comfortable 10-m timed walk (tau-b = -0.629), and the fast 10-m timed walk (tau-b = -0.634). Moderate correlations were found between the FWC and step length (tau-b = 0.483) and the Timed Up and Go (tau-b = -0.520), respectively. Intrarater reliability was moderate (ICC = 0.651) while interrater reliability was excellent (ICC = 0.751) (all correlations p < 0.001). However, the study was designed as pilot study, thus, full psychometric property testing was not possible. CONCLUSIONS: The German FWC offers a reasonable tool for measuring community ambulation on participation level. However, a user manual seems to be helpful.

Automatic Musculoskeletal and Neurological Disorder Diagnosis With Relative Joint Displacement From Human Gait.

Musculoskeletal and neurological disorders are common devastating companions of ageing, leading to a reduction in quality of life and increased mortality. Gait analysis is a popular method for diagnosing these disorders. However, manually analyzing the motion data is a labor-intensive task, and the quality of the results depends on the experience of the doctors. In this paper, we propose an automatic framework for classifying musculoskeletal and neurological disorders among older people based on 3D motion data. We also propose two new features to capture the relationship between joints across frames, known as 3D Relative Joint Displacement (3DRJDP) and 6D Symmetric Relative Joint Displacement (6DSymRJDP), such that the relative movement between joints can be analyzed. To optimize the classification performance, we adapt feature selection methods to choose an optimal feature set from the raw feature input. Experimental results show that we achieve a classification accuracy of 84.29% using the proposed relative joint features, outperforming existing features that focus on the movement of individual joints. Considering the limited open motion database for gait analysis focusing on such disorders, we construct a comprehensive, openly accessible 3D full-body motion database from 45 subjects.

Evidence for subtypes of freezing of gait in Parkinson's disease.

BACKGROUND: The purpose of this study is to identify and characterize subtypes of freezing of gait by using a novel questionnaire designed to delineate freezing patterns based on self-reported and behavioral gait assessment. METHODS: A total of 41 Parkinson's patients with freezing completed the Characterizing Freezing of Gait questionnaire that identifies situations that exacerbate freezing. This instrument underwent examination for construct validity and internal consistency, after which a data-driven clustering approach was employed to identify distinct patterns amongst individual responses. Behavioral freezing assessments in both dopaminergic states were compared across 3 identified subgroups. RESULTS: This novel questionnaire demonstrated construct validity (severity scores correlated with percentage of time frozen; r = 0.54) and internal consistency (Cronbach's α = .937), and thus demonstrated promising utility for identifying patterns of freezing that are independently related to motor, anxiety, and attentional impairments. CONCLUSIONS: Patients with freezing may be dissociable based on underlying neurobiological underpinnings that would have significant implications for targeting future treatments. © 2018 International Parkinson and Movement Disorder Society.

Pelvic excursion during walking post-stroke: A novel classification system.

BACKGROUND: Researchers and clinicians often use gait speed to classify hemiparetic gait dysfunction because it offers clinical predictive capacity. However, gait speed fails to distinguish unique biomechanical characteristics that differentiate aspects of gait dysfunction. RESEARCH QUESTION: Here we describe a novel classification of hemiparetic gait dysfunction based on biomechanical traits of pelvic excursion. We hypothesize that individuals with greater deviation of pelvic excursion, relative to controls, demonstrate greater impairment in key gait characteristics. METHODS: We compared 41 participants (61.0 ±â€¯11.2yrs) with chronic post-stroke hemiparesis to 21 non-disabled controls (55.8 ±â€¯9.0yrs). Participants walked on an instrumented split-belt treadmill at self-selected walking speed. Pelvic excursion was quantified as the peak-to-peak magnitude of pelvic motion in three orthogonal planes (i.e., tilt, rotation, and obliquity). Raw values of pelvic excursion were compared against the distribution of control data to establish deviation scores which were assigned bilaterally for the three planes producing six values per individual. Deviation scores were then summed to produce a composite pelvic deviation score. Based on composite scores, participants were allocated to one of three categories of hemiparetic gait dysfunction with progressively increasing pelvic excursion deviation relative to controls: Type I (n = 15) - minimal pelvic excursion deviation; Type II (n = 20) - moderate pelvic excursion deviation; and Type III (n = 6) - marked pelvic excursion deviation. We assessed resulting groups for asymmetry in key gait parameters including: kinematics, joint powers temporally linked to the stance-to-swing transition, and timing of lower extremity muscle activity. RESULTS: All groups post-stroke walked at similar self-selected speeds; however, classification based on pelvic excursion deviation revealed progressive asymmetry in gait kinematics, kinetics and temporal patterns of muscle activity. SIGNIFICANCE: The progressive asymmetry revealed in multiple gait characteristics suggests exaggerated pelvic motion contributes to gait dysfunction post-stroke.

Profiling walking dysfunction in multiple sclerosis: characterisation, classification and progression over time.

Gait dysfunction is a common and relevant symptom in multiple sclerosis (MS). This study aimed to profile gait pathology in gait-impaired patients with MS using comprehensive 3D gait analysis and clinical walking tests. Thirty-seven patients with MS walked on the treadmill at their individual, sustainable speed while 20 healthy control subjects walked at all the different patient's paces, allowing for comparisons independent of walking velocity. Kinematic analysis revealed pronounced restrictions in knee and ankle joint excursion, increased gait variability and asymmetry along with impaired dynamic stability in patients. The most discriminative single gait parameter, differentiating patients from controls with an accuracy of 83.3% (χ2 test; p = 0.0001), was reduced knee range of motion. Based on hierarchical cluster and principal component analysis, three principal pathological gait patterns were identified: a spastic-paretic, an ataxia-like, and an unstable gait. Follow-up assessments after 1 year indicated deterioration of walking function, particularly in patients with spastic-paretic gait patterns. Our findings suggest that impaired knee/ankle control is common in patients with MS. Personalised gait profiles and clustering algorithms may be promising tools for stratifying patients and to inform patient-tailored exercise programs. Responsive, objective outcome measures are important for monitoring disease progression and treatment effects in MS trials.

Quantitative Assessment of Knee Progression Angle During Gait in Children With Cerebral Palsy.

BACKGROUND: Abnormal hip rotation is a common deviation in children with cerebral palsy (CP). Clinicians typically assess hip rotation during gait by observing the direction that the patella points relative to the path of walking, which is referred to as the knee progression angle (KPA). Two kinematic methods for calculating the KPA are compared with each other. Video-based qualitative assessment of KPA is compared with the quantitative methods to determine reliability and validity. METHODS: The KPA was calculated by both direct and indirect methods for 32 typically developing (TD) children and a convenience cohort of 43 children with hemiplegic type CP. An additional convenience cohort of 26 children with hemiplegic type CP was selected for qualitative assessment of KPA, performed by 3 experienced clinicians, using 3 categories (internal, >10 degrees; neutral, -10 to 10 degrees; and external, >-10 degrees). RESULTS: Root mean square (RMS) analysis comparing the direct and indirect KPAs was 1.14+0.43 degrees for TD children, and 1.75+1.54 degrees for the affected side of children with CP. The difference in RMS among the 2 groups was statistically, but not clinically, significant (P=0.019). Intraclass correlation coefficient revealed excellent agreement between the direct and indirect methods of KPA for TD and CP children (0.996 and 0.992, respectively; P<0.001).For the qualitative assessment of KPA there was complete agreement among all examiners for 17 of 26 cases (65%). Direct KPA matched for 49 of 78 observations (63%) and indirect KPA matched for 52 of 78 observations (67%). CONCLUSIONS: The RMS analysis of direct and indirect methods for KPA was statistically but not clinically significant, which supports the use of either method based upon availability. Video-based qualitative assessment of KPA showed moderate reliability and validity. The differences between observed and calculated KPA indicate the need for caution when relying on visual assessments for clinical interpretation, and demonstrate the value of adding KPA calculation to standard kinematic analysis. LEVEL OF EVIDENCE: Level II-diagnostic test.

An updated diagnostic approach to subtype definition of vascular parkinsonism - Recommendations from an expert working group.

This expert working group report proposes an updated approach to subtype definition of vascular parkinsonism (VaP) based on a review of the existing literature. The persistent lack of consensus on clear terminology and inconsistent conceptual definition of VaP formed the impetus for the current expert recommendation report. The updated diagnostic approach intends to provide a comprehensive tool for clinical practice. The preamble for this initiative is that VaP can be diagnosed in individual patients with possible prognostic and therapeutic consequences and therefore should be recognized as a clinical entity. The diagnosis of VaP is based on the presence of clinical parkinsonism, with variable motor and non-motor signs that are corroborated by clinical, anatomic or imaging findings of cerebrovascular disease. Three VaP subtypes are presented: (1) The acute or subacute post-stroke VaP subtype presents with acute or subacute onset of parkinsonism, which is typically asymmetric and responds to dopaminergic drugs; (2) The more frequent insidious onset VaP subtype presents with progressive parkinsonism with prominent postural instability, gait impairment, corticospinal, cerebellar, pseudobulbar, cognitive and urinary symptoms and poor responsiveness to dopaminergic drugs. A higher-level gait disorder occurs frequently as a dominant manifestation in the clinical spectrum of insidious onset VaP, and (3) With the emergence of molecular imaging biomarkers in clinical practice, our diagnostic approach also allows for the recognition of mixed or overlapping syndromes of VaP with Parkinson's disease or other neurodegenerative parkinsonisms. Directions for future research are also discussed.

Texture Classification and Visualization of Time Series of Gait Dynamics in Patients With Neuro-Degenerative Diseases.

The analysis of gait dynamics is helpful for predicting and improving the quality of life, morbidity, and mortality in neuro-degenerative patients. Feature extraction of physiological time series and classification between gait patterns of healthy control subjects and patients are usually carried out on the basis of 1-D signal analysis. The proposed approach presented in this paper departs itself from conventional methods for gait analysis by transforming time series into images, of which texture features can be extracted from methods of texture analysis. Here, the fuzzy recurrence plot algorithm is applied to transform gait time series into texture images, which can be visualized to gain insight into disease patterns. Several texture features are then extracted from fuzzy recurrence plots using the gray-level co-occurrence matrix for pattern analysis and machine classification to differentiate healthy control subjects from patients with Parkinson's disease, Huntington's disease, and amyotrophic lateral sclerosis. Experimental results using only the right stride-intervals of the four groups show the effectiveness of the application of the proposed approach.

An Automated Classification of Pathological Gait Using Unobtrusive Sensing Technology.

This paper integrates an unobtrusive and affordable sensing technology with machine learning methods to discriminate between healthy and pathological gait patterns as a result of stroke or acquired brain injury. A feature analysis is used to identify the role of each body part in separating pathological patterns from healthy patterns. Gait features, including the orientations of the hips and spine (trunk), shoulders and neck (upper limb), knees and ankles (lower limb), are calculated during walking based on Kinect skeletal tracking sequences. Sequences of these features during three types of walking conditions were examined: 1) walking at self-pace (WSP); 2) walking at distracted (WD); and 3) walking at fast pace (WFP). Two machine learning approaches, an instance-based discriminative classifier ( -nearest neighbor) and a dynamical generative classifier (using Gaussian Process Latent Variable Model), are examined to distinguish between healthy and pathological gaits. Nested cross validation is implemented to evaluate the performance of the two classifiers using three metrics: F1-score, macro-averaged error, and micro-averaged error. The discriminative model outperforms the generative model in terms of the F1-score (discriminative: WSP > 0.95, WD > 0.96, and WFP > 0.95 and generative: WSP > 0.87, WD > 0.85, and WFP > 0.68) and macro-averaged error (discriminative: WSP < 0.08, WD < 0.1, and WFP < 0.09 and generative: WSP < 0.11, WD < 0.12, and WFP < 0.14). The dynamical generative model on the other hand obtains better micro-averaged error (discriminative: WSP < 0.37, WD < 0.3, and WFP < 0.35 and generative: WSP < 0.15, WD < 0.2, and WFP < 0.2). The high-dimensional gait features are divided into five subsets: lower limb, upper limb, trunk, velocity, and acceleration. An instance-based feature analysis method (ReliefF) is used to assign weights to each subset of features according to its discriminatory power. The feature analysis establishes the most informative features (upper limb, lower limb, and trunk) for identifying pathological gait.

Charcot: Past and present.

Jean-Martin Charcot (1825-1893) was the preeminent neurologist of the nineteenth century. Several of his major contributions remain fully relevant to contemporary neurology, and this essay highlights three areas of particular importance to the modern neurologist: the anatomo-clinical method that Charcot developed as the anchor of neurological study; the integration of new scientific discoveries from other fields as a core strategy for neurological advancement; and the role of heredity as the fundamental etiological focus to the understanding of the pathogenesis of primary neurological disorders. Further, Charcot left a strong tradition of visual skills as the core requirement for accurate neurological diagnosis and emphasized scientific humility in the face of difficult diseases. In spite of vast advances in neuroscience over the 20th and 21st centuries, the challenges faced by Charcot remain largely the same for the contemporary neurologist, and the lessons provided by Charcot retain their power and significance today.