The role of stroke-induced immunosuppression as a predictor of functional outcome in the neurorehabilitation setting.

Stroke affects the interconnection between the nervous and immune systems, leading to a down-regulation of immunity called stroke-induced immunosuppression (SII). The primary aim of this study is to investigate SII role as a predictor of functional, neurological, and motor outcomes in the neurorehabilitation setting (NRB). We conducted a prospective observational study enrolling post-acute stroke patients hospitalized for neurorehabilitation. At NRB admission (T0) and discharge (T1), we assessed presence of SII (defined by a neutrophil-to-lymphocyte ratio ≥ 5) and we evaluated functional independence (Functional Independence Measure-FIM, Barthel Index-BI), motor performances (Tinetti Score, Hauser Ambulation Index) and neurological impairment (NIHSS). We enrolled 96 patients (45.8% females, 70.6 ± 13.9 years, 88.5% ischemic stroke). At T0, 15.6% of patients (15/96) had SII. When compared to immunocompetent patients (IC), the SII group was characterized by worse baseline functional independence, motor performances and neurological disability. The same was confirmed at T1 (FIM p = 0.012, BI p = 0.007, Tinetti p = 0.034, NIHSS p = 0.001). Neurological disability demonstrated a less pronounced improvement in SII (ΔNIHSS: SII: - 2.1 ± 2.3 vs. IC: - 3.1 ± 2.5, p = 0.035). SII group presented a higher percentage of infectious complications during the neurorehabilitation period (SII 80% vs. IC 25.9%; p = 0.001). SII may represent a negative prognostic factor in the neurorehabilitation setting. SII patients were characterized by poorer functional, motor, neurological performances and higher risk of infectious complications. ClinicaTrial registration: NCT05889169.

The Lower Limb Muscle Co-Activation Map during Human Locomotion: From Slow Walking to Running.

The central nervous system (CNS) controls movements and regulates joint stiffness with muscle co-activation, but until now, few studies have examined muscle pairs during running. This study aims to investigate differences in lower limb muscle coactivation during gait at different speeds, from walking to running. Nineteen healthy runners walked and ran at speeds ranging from 0.8 km/h to 9.3 km/h. Twelve lower limb muscles' co-activation was calculated using the time-varying multi-muscle co-activation function (TMCf) with global, flexor-extension, and rostro-caudal approaches. Spatiotemporal and kinematic parameters were also measured. We found that TMCf, spatiotemporal, and kinematic parameters were significantly affected by gait speed for all approaches. Significant differences were observed in the main parameters of each co-activation approach and in the spatiotemporal and kinematic parameters at the transition between walking and running. In particular, significant differences were observed in the global co-activation (CIglob, main effect F(1,17) = 641.04, p < 0.001; at the transition p < 0.001), the stride length (main effect F(1,17) = 253.03, p < 0.001; at the transition p < 0.001), the stride frequency (main effect F(1,17) = 714.22, p < 0.001; at the transition p < 0.001) and the Center of Mass displacement in the vertical (CoMy, main effect F(1,17) = 426.2, p < 0.001; at the transition p < 0.001) and medial-lateral (CoMz, main effect F(1,17) = 120.29 p < 0.001; at the transition p < 0.001) directions. Regarding the correlation analysis, the CoMy was positively correlated with a higher CIglob (r = 0.88, p < 0.001) and negatively correlated with Full Width at Half Maximum (FWHMglob, r = -0.83, p < 0.001), whereas the CoMz was positively correlated with the global Center of Activity (CoAglob, r = 0.97, p < 0.001). Positive and negative strong correlations were found between global co-activation parameters and center of mass displacements, as well as some spatiotemporal parameters, regardless of gait speed. Our findings suggest that walking and running have different co-activation patterns and kinematic characteristics, with the whole-limb stiffness exerted more synchronously and stably during running. The co-activation indexes and kinematic parameters could be the result of global co-activation, which is a sensory-control integration process used by the CNS to deal with more demanding and potentially unstable tasks like running.

Gait analysis, trunk movements, and electromyographic patterns after minimally invasive spine surgery for lumbar instability: An observational prospective study.

Objective: The aim of the present study was to investigate trunk kinematics and spine muscle activation during walking after minimally invasive surgery in patients with L4-L5 degenerative spondylolisthesis suffering from lumbar instability (LI). Methods: Eleven patients suffering from LI and 13 healthy controls (HC) were enrolled. Trunk kinematics and spine muscle activation patterns during walking were collected. Maximal trunk ranges of motion were also recorded from standing position. Assessments were performed pre-operatively (T0), 1 month (T1) and 3 months (T2) after MIS. Results: We found significant improvement in spine muscle activation during walking at T2 compared to T0, mainly involving right/left symmetry at the operated level (L4-L5) and up-down synchronization from L3 to S1. Significant improvements in trunk rotation nearing to the HC group during walking were also found at T2 after surgery, though no changes were observed in the maximal range of motion of the trunk during standing. Furthermore, trunk rotation improvement correlated with a lower grade of residual disability. Conclusions: Our findings indicate that trunk rotation improves after surgery, and impaired aspects of spine muscle activation can be improved with surgery. These biomechanical parameters could represent novel tools for monitoring the effect of surgery in LI and preventing impaired spine mobility and muscle activation.

Local Dynamic Stability of Trunk During Gait is Responsive to Rehabilitation in Subjects with Primary Degenerative Cerebellar Ataxia.

This study aimed to assess the responsiveness to the rehabilitation of three trunk acceleration-derived gait indexes, namely the harmonic ratio (HR), the short-term longest Lyapunov's exponent (sLLE), and the step-to-step coefficient of variation (CV), in a sample of subjects with primary degenerative cerebellar ataxia (swCA), and investigate the correlations between their improvements (∆), clinical characteristics, and spatio-temporal and kinematic gait features. The trunk acceleration patterns in the antero-posterior (AP), medio-lateral (ML), and vertical (V) directions during gait of 21 swCA were recorded using a magneto-inertial measurement unit placed at the lower back before (T0) and after (T1) a period of inpatient rehabilitation. For comparison, a sample of 21 age- and gait speed-matched healthy subjects (HSmatched) was also included. At T1, sLLE in the AP (sLLEAP) and ML (sLLEML) directions significantly improved with moderate to large effect sizes, as well as SARA scores, stride length, and pelvic rotation. sLLEML and pelvic rotation also approached the HSmatched values at T1, suggesting a normalization of the parameter. HRs and CV did not significantly modify after rehabilitation. ∆sLLEML correlated with ∆ of the gait subscore of the SARA scale (SARAGAIT) and ∆stride length and ∆sLLEAP correlated with ∆pelvic rotation and ∆SARAGAIT. The minimal clinically important differences for sLLEML and sLLEAP were ≥ 36.16% and ≥ 28.19%, respectively, as the minimal score reflects a clinical improvement in SARA scores. When using inertial measurement units, sLLEAP and sLLEML can be considered responsive outcome measures for assessing the effectiveness of rehabilitation on trunk stability during walking in swCA.

Central sensitization mechanisms in chronic migraine with medication overuse headache: a study of thalamocortical activation and lateral cortical inhibition.

BACKGROUND: It is unclear whether cortical hyperexcitability in chronic migraine with medication overuse headache (CM-MOH) is due to increased thalamocortical drive or aberrant cortical inhibitory mechanisms. METHODS: Somatosensory evoked potentials (SSEP) were performed by electrical stimulation of the median nerve (M), ulnar nerve (U) and simultaneous stimulation of both nerves (MU) in 27 patients with CM-MOH and, for comparison, in 23 healthy volunteers (HVs) of a comparable age distribution. We calculated the degree of cortical lateral inhibition using the formula: 100 - [MU/(M + U) × 100] and the level of thalamocortical activation by analyzing the high frequency oscillations (HFOs) embedded in parietal N20 median SSEPs. RESULTS: Compared to HV, CM-MOH patients showed higher lateral inhibition (CM-MOH 52.2% ± 15.4 vs. HV 40.4% ± 13.3; p = 0.005), which positively correlated with monthly headache days, and greater amplitude of pre-synaptic HFOs (p = 0.010) but normal post-synaptic HFOs (p = 0.122). CONCLUSION: Our findings suggest that central neuronal circuits are highly sensitized in CM-MOH patients, at both thalamocortical and cortical levels. The observed changes could be due to the combination of dysfunctional central pain control mechanisms, hypersensitivity and hyperresponsiveness directly linked to the chronic intake of acute migraine drugs.

Neurophysiology of cerebellar ataxias and gait disorders.

There are numerous forms of cerebellar disorders from sporadic to genetic diseases. The aim of this chapter is to provide an overview of the advances and emerging techniques during these last 2 decades in the neurophysiological tests useful in cerebellar patients for clinical and research purposes. Clinically, patients exhibit various combinations of a vestibulocerebellar syndrome, a cerebellar cognitive affective syndrome and a cerebellar motor syndrome which will be discussed throughout this chapter. Cerebellar patients show abnormal Bereitschaftpotentials (BPs) and mismatch negativity. Cerebellar EEG is now being applied in cerebellar disorders to unravel impaired electrophysiological patterns associated within disorders of the cerebellar cortex. Eyeblink conditioning is significantly impaired in cerebellar disorders: the ability to acquire conditioned eyeblink responses is reduced in hereditary ataxias, in cerebellar stroke and after tumor surgery of the cerebellum. Furthermore, impaired eyeblink conditioning is an early marker of cerebellar degenerative disease. General rules of motor control suggest that optimal strategies are needed to execute voluntary movements in the complex environment of daily life. A high degree of adaptability is required for learning procedures underlying motor control as sensorimotor adaptation is essential to perform accurate goal-directed movements. Cerebellar patients show impairments during online visuomotor adaptation tasks. Cerebellum-motor cortex inhibition (CBI) is a neurophysiological biomarker showing an inverse association between cerebellothalamocortical tract integrity and ataxia severity. Ataxic gait is characterized by increased step width, reduced ankle joint range of motion, increased gait variability, lack of intra-limb inter-joint and inter-segmental coordination, impaired foot ground placement and loss of trunk control. Taken together, these techniques provide a neurophysiological framework for a better appraisal of cerebellar disorders.

Multiscale Entropy Algorithms to Analyze Complexity and Variability of Trunk Accelerations Time Series in Subjects with Parkinson's Disease.

The aim of this study was to assess the ability of multiscale sample entropy (MSE), refined composite multiscale entropy (RCMSE), and complexity index (CI) to characterize gait complexity through trunk acceleration patterns in subjects with Parkinson's disease (swPD) and healthy subjects, regardless of age or gait speed. The trunk acceleration patterns of 51 swPD and 50 healthy subjects (HS) were acquired using a lumbar-mounted magneto-inertial measurement unit during their walking. MSE, RCMSE, and CI were calculated on 2000 data points, using scale factors (τ) 1-6. Differences between swPD and HS were calculated at each τ, and the area under the receiver operating characteristics, optimal cutoff points, post-test probabilities, and diagnostic odds ratios were calculated. MSE, RCMSE, and CIs showed to differentiate swPD from HS. MSE in the anteroposterior direction at τ4 and τ5, and MSE in the ML direction at τ4 showed to characterize the gait disorders of swPD with the best trade-off between positive and negative posttest probabilities and correlated with the motor disability, pelvic kinematics, and stance phase. Using a time series of 2000 data points, a scale factor of 4 or 5 in the MSE procedure can yield the best trade-off in terms of post-test probabilities when compared to other scale factors for detecting gait variability and complexity in swPD.

Whole brain surface-based morphometry and tract-based spatial statistics in migraine with aura patients: difference between pure visual and complex auras.

Background: The migrainous aura has different clinical phenotypes. While the various clinical differences are well-described, little is known about their neurophysiological underpinnings. To elucidate the latter, we compared white matter fiber bundles and gray matter cortical thickness between healthy controls (HC), patients with pure visual auras (MA) and patients with complex neurological auras (MA+). Methods: 3T MRI data were collected between attacks from 20 patients with MA and 15 with MA+, and compared with those from 19 HCs. We analyzed white matter fiber bundles using tract-based spatial statistics (TBSS) of diffusion tensor imaging (DTI) and cortical thickness with surface-based morphometry of structural MRI data. Results: Tract-based spatial statistics showed no significant difference in diffusivity maps between the three subject groups. As compared to HCs, both MA and MA+ patients had significant cortical thinning in temporal, frontal, insular, postcentral, primary and associative visual areas. In the MA group, the right high-level visual-information-processing areas, including lingual gyrus, and the Rolandic operculum were thicker than in HCs, while in the MA+ group they were thinner. Discussion: These findings show that migraine with aura is associated with cortical thinning in multiple cortical areas and that the clinical heterogeneity of the aura is reflected by opposite thickness changes in high-level visual-information-processing, sensorimotor and language areas.

Characterization of prosthetic knees through a low-dimensional description of gait kinematics.

The characterization of both limbs' behaviour in prosthetic gait is of key importance for improving the prosthetic components and increasing the biomechanical capability of trans-femoral amputees. When characterizing human gait, modular motor control theories have been proven to be powerful in providing a compact description of the gait patterns. In this paper, the planar covariation law of lower limb elevation angles is proposed as a compact, modular description of prosthetic gait; this model is exploited for a comparison between trans-femoral amputees walking with different prosthetic knees and control subjects walking at different speeds. Results show how the planar covariation law is maintained in prostheses users, with a similar spatial organization and few temporal differences. Most of the differences among the different prosthetic knees are found in the kinematic coordination patterns of the sound side. Moreover, different geometrical parameters have been calculated over the common projected plane, and their correlation with classical gait spatiotemporal and stability parameters has been investigated. The results from this latter analysis have highlighted a correlation with several parameters of gait, suggesting that this compact description of kinematics unravels a significant biomechanical meaning. These results can be exploited to guide the control mechanisms of prosthetic devices based purely on the measurement of relevant kinematic quantities.

Translation of surface electromyography to clinical and motor rehabilitation applications: The need for new clinical figures.

Advanced sensors/electrodes and signal processing techniques provide powerful tools to analyze surface electromyographic signals (sEMG) and their features, to decompose sEMG into the constituent motor unit action potential trains, and to identify synergies, neural muscle drive, and EEG-sEMG coherence. However, despite thousands of articles, dozens of textbooks, tutorials, consensus papers, and European and International efforts, the translation of this knowledge into clinical activities and assessment procedures has been very slow, likely because of lack of clinical studies and competent operators in the field. Understanding and using sEMG-based hardware and software tools requires a level of knowledge of signal processing and interpretation concepts that is multidisciplinary and is not provided by most academic curricula in physiotherapy, movement sciences, neurophysiology, rehabilitation, sport, and occupational medicine. The chasm existing between the available knowledge and its clinical applications in this field is discussed as well as the need for new clinical figures. The need for updating the training of physiotherapists, neurophysiology technicians, and clinical technologists is discussed as well as the required competences of trainers and trainees. Indications and examples are suggested and provide a basis for addressing the problem. Two teaching examples are provided in the Supplementary Material.

A Proof of Concept Combined Using Mixed Reality for Personalized Neurorehabilitation of Cerebellar Ataxic Patients.

BACKGROUND: Guidelines for degenerative cerebellar ataxia neurorehabilitation suggest intensive coordinative training based on physiotherapeutic exercises. Scientific studies demonstrate virtual exergaming therapeutic value. However, patient-based personalization, post processing analyses and specific audio-visual feedbacks are not provided. This paper presents a wearable motion tracking system with recording and playback features. This system has been specifically designed for ataxic patients, for upper limbs coordination studies with the aim to retrain movement in a neurorehabilitation setting. Suggestions from neurologists and ataxia patients were considered to overcome the shortcomings of virtual systems and implement exergaming. METHODS: The system consists of the mixed-reality headset Hololens2 and a proprietary exergaming implemented in Unity. Hololens2 can track and save upper limb parameters, head position and gaze direction in runtime. RESULTS: Data collected from a healthy subject are reported to demonstrate features and outputs of the system. CONCLUSIONS: Although further improvements and validations are needed, the system meets the needs of a dynamic patient-based exergaming for patients with cerebellar ataxia. Compared with existing solutions, the mixed-reality system is designed to provide an effective and safe therapeutic exergaming that supports both primary and secondary goals of an exergaming: what a patient should do and how patient actions should be performed.

Effects of Botulinum Toxin Type A on the Nociceptive and Lemniscal Somatosensory Systems in Chronic Migraine: An Electrophysiological Study.

(1) Background: OnabotulinumtoxinA (BoNT-A) is a commonly used prophylactic treatment for chronic migraine (CM). Although randomized placebo studies have shown its clinical efficacy, the mechanisms by which it exerts its therapeutic effect are still incompletely understood and debated. (2) Methods: We studied in 15 CM patients the cephalic and extracephalic nociceptive and lemniscal sensory systems using electrophysiological techniques before and 1 and 3 months after one session of pericranial BoNT-A injections according to the PREEMPT protocol. We recorded the nociceptive blink reflex (nBR), the trigemino-cervical reflex (nTCR), the pain-related cortical evoked potential (PREP), and the upper limb somatosensory evoked potential (SSEP). (3) Results: Three months after a single session of prophylactic therapy with BoNT-A in CM patients, we found (a) an increase in the homolateral and contralateral nBR AUC, (b) an enhancement of the contralateral nBR AUC habituation slope and the nTCR habituation slope, (c) a decrease in PREP N-P 1st and 2nd amplitude block, and (d) no effect on SSEPs. (4) Conclusions: Our study provides electrophysiological evidence for the ability of a single session of BoNT-A injections to exert a neuromodulatory effect at the level of trigeminal system through a reduction in input from meningeal and other trigeminovascular nociceptors. Moreover, by reducing activity in cortical pain processing areas, BoNT-A restores normal functioning of the descending pain modulation systems.

Consensus Paper: Ataxic Gait.

The aim of this consensus paper is to discuss the roles of the cerebellum in human gait, as well as its assessment and therapy. Cerebellar vermis is critical for postural control. The cerebellum ensures the mapping of sensory information into temporally relevant motor commands. Mental imagery of gait involves intrinsically connected fronto-parietal networks comprising the cerebellum. Muscular activities in cerebellar patients show impaired timing of discharges, affecting the patterning of the synergies subserving locomotion. Ataxia of stance/gait is amongst the first cerebellar deficits in cerebellar disorders such as degenerative ataxias and is a disabling symptom with a high risk of falls. Prolonged discharges and increased muscle coactivation may be related to compensatory mechanisms and enhanced body sway, respectively. Essential tremor is frequently associated with mild gait ataxia. There is growing evidence for an important role of the cerebellar cortex in the pathogenesis of essential tremor. In multiple sclerosis, balance and gait are affected due to cerebellar and spinal cord involvement, as a result of disseminated demyelination and neurodegeneration impairing proprioception. In orthostatic tremor, patients often show mild-to-moderate limb and gait ataxia. The tremor generator is likely located in the posterior fossa. Tandem gait is impaired in the early stages of cerebellar disorders and may be particularly useful in the evaluation of pre-ataxic stages of progressive ataxias. Impaired inter-joint coordination and enhanced variability of gait temporal and kinetic parameters can be grasped by wearable devices such as accelerometers. Kinect is a promising low cost technology to obtain reliable measurements and remote assessments of gait. Deep learning methods are being developed in order to help clinicians in the diagnosis and decision-making process. Locomotor adaptation is impaired in cerebellar patients. Coordinative training aims to improve the coordinative strategy and foot placements across strides, cerebellar patients benefiting from intense rehabilitation therapies. Robotic training is a promising approach to complement conventional rehabilitation and neuromodulation of the cerebellum. Wearable dynamic orthoses represent a potential aid to assist gait. The panel of experts agree that the understanding of the cerebellar contribution to gait control will lead to a better management of cerebellar ataxias in general and will likely contribute to use gait parameters as robust biomarkers of future clinical trials.

Identification of Gait Unbalance and Fallers Among Subjects with Cerebellar Ataxia by a Set of Trunk Acceleration-Derived Indices of Gait.

This study aimed to assess the ability of 25 gait indices to characterize gait instability and recurrent fallers among persons with primary degenerative cerebellar ataxia (pwCA), regardless of gait speed, and investigate their correlation with clinical and kinematic variables. Trunk acceleration patterns were acquired during the gait of 34 pwCA, and 34 age- and speed-matched healthy subjects (HSmatched) using an inertial measurement unit. We calculated harmonic ratios (HR), percent recurrence, percent determinism, step length coefficient of variation, short-time largest Lyapunov exponent (sLLE), normalized jerk score, log-dimensionless jerk (LDLJ-A), root mean square (RMS), and root mean square ratio of accelerations (RMSR) in each spatial direction for each participant. Unpaired t-tests or Mann-Whitney tests were performed to identify significant differences between the pwCA and HSmatched groups. Receiver operating characteristics were plotted to assess the ability to characterize gait alterations in pwCA and fallers. Optimal cutoff points were identified, and post-test probabilities were calculated. The HRs showed to characterize gait instability and pwCA fallers with high probabilities. They were correlated with disease severity and stance, swing, and double support duration, regardless of gait speed. sLLEs, RMSs, RMSRs, and LDLJ-A were slightly able to characterize the gait of pwCA but failed to characterize fallers.

Characterizing the Gait of People With Different Types of Amputation and Prosthetic Components Through Multimodal Measurements: A Methodological Perspective.

Prosthetic gait implies the use of compensatory motor strategies, including alterations in gait biomechanics and adaptations in the neural control mechanisms adopted by the central nervous system. Despite the constant technological advancements in prostheses design that led to a reduction in compensatory movements and an increased acceptance by the users, a deep comprehension of the numerous factors that influence prosthetic gait is still needed. The quantitative prosthetic gait analysis is an essential step in the development of new and ergonomic devices and to optimize the rehabilitation therapies. Nevertheless, the assessment of prosthetic gait is still carried out by a heterogeneous variety of methodologies, and this limits the comparison of results from different studies, complicating the definition of shared and well-accepted guidelines among clinicians, therapists, physicians, and engineers. This perspective article starts from the results of a project funded by the Italian Worker's Compensation Authority (INAIL) that led to the generation of an extended dataset of measurements involving kinematic, kinetic, and electrophysiological recordings in subjects with different types of amputation and prosthetic components. By encompassing different studies published along the project activities, we discuss the specific information that can be extracted by different kinds of measurements, and we here provide a methodological perspective related to multimodal prosthetic gait assessment, highlighting how, for designing improved prostheses and more effective therapies for patients, it is of critical importance to analyze movement neural control and its mechanical actuation as a whole, without limiting the focus to one specific aspect.

Harmonic ratio is the most responsive trunk-acceleration derived gait index to rehabilitation in people with Parkinson's disease at moderate disease stages.

BACKGROUND: Harmonic ratios (HRs), recurrence quantification analysis in the antero-posterior direction (RQAdetAP), and stride length coefficient of variation (CV) have recently been shown to characterize gait abnormalities and fall risk in people with Parkinson's disease (pwPD) at moderate disease stages. RESEARCH QUESTION: This study aimed to i) assess the internal and external responsiveness to rehabilitation of HR, RQAdetAP, and CV, ii) identify the baseline predictors of normalization of the gait stability indexes, and iii) investigate the correlations between the gait indexes modifications (∆) and clinical and kinematic ∆s in pwPD at Hoehn and Yahr disease staging classification 3. METHODS: The trunk acceleration patterns of 21 pwPD and 21 age- and speed-matched healthy subjects (HSmatched) were acquired during gait using an inertial measurement unit at baseline (T0). pwPD were also assessed after a 4-week rehabilitation period (T1). Each participant's HR in the antero-posterior (HRAP), medio-lateral (HRML), and vertical directions, RQAdetAP, CV, spatio-temporal, and kinematic variables were calculated. RESULTS: At T1, HRAP and HRML improved to normative values and showed high internal and external responsiveness. Lower HRs and higher pelvic rotation values at baseline were predictors of ∆HRs. A minimal clinically important difference (MCID) ≥ 21.5 % is required to normalize HRAP with 95 % probability. MCID ≥ 36.9 % is required to normalize HRML with 92 % probability. ∆HRAP correlated with ∆HRML and both correlated with ∆stride length and ∆pelvic rotation, regardless of ∆gait speed. RQAdetAP and step length CV were not responsive to rehabilitation. SIGNIFICANCE: When using inertial measurement units, HRAP and HRML can be considered as responsive outcome measures for assessing the effectiveness of rehabilitation on trunk smoothness during walking in pwPD at moderate disease stages.

Machine Learning Approach to Support the Detection of Parkinson's Disease in IMU-Based Gait Analysis.

The aim of this study was to determine which supervised machine learning (ML) algorithm can most accurately classify people with Parkinson's disease (pwPD) from speed-matched healthy subjects (HS) based on a selected minimum set of IMU-derived gait features. Twenty-two gait features were extrapolated from the trunk acceleration patterns of 81 pwPD and 80 HS, including spatiotemporal, pelvic kinematics, and acceleration-derived gait stability indexes. After a three-level feature selection procedure, seven gait features were considered for implementing five ML algorithms: support vector machine (SVM), artificial neural network, decision trees (DT), random forest (RF), and K-nearest neighbors. Accuracy, precision, recall, and F1 score were calculated. SVM, DT, and RF showed the best classification performances, with prediction accuracy higher than 80% on the test set. The conceptual model of approaching ML that we proposed could reduce the risk of overrepresenting multicollinear gait features in the model, reducing the risk of overfitting in the test performances while fostering the explainability of the results.