Abstract:Background: Deep learning has always been at the forefront of scientific research. It has also been applied to medical research. Hereditary spinocerebellar ataxia (SCA) is characterized by gait abnormalities and is usually evaluated semi-quantitatively by scales. However, more detailed gait characteristics of SCA and related objective methods have not yet been established. Therefore, the purpose of this study was to evaluate the gait characteristics of SCA patients, as well as to analyze the correlation between gait parameters, clinical scales, and imaging on deep learning. Methods: Twenty SCA patients diagnosed by genetic detection were included in the study. Ten patients who were tested via functional magnetic resonance imaging (fMRI) were included in the SCA imaging subgroup. All SCA patients were evaluated with the International Cooperative Ataxia Rating Scale (ICARS) and Scale for the Assessment and Rating of Ataxia (SARA) clinical scales. The gait control group included 16 healthy subjects, and the imaging control group included seven healthy subjects. Gait data consisting of 10 m of free walking of each individual in the SCA group and the gait control group were detected by wearable gait-detection equipment. Stride length, stride time, velocity, supporting-phase percentage, and swinging-phase percentage were extracted as gait parameters. Cerebellar volume and the midsagittal cerebellar proportion in the posterior fossa (MRVD) were calculated according to MR. Results: There were significant differences in stride length, velocity, supporting-phase percentage, and swinging-phase percentage between the SCA group and the gait control group. The stride length and stride velocity of SCA groups were lower while supporting phase was longer than those of the gait control group. SCA group's velocity was negatively correlated with both the ICARS and SARA scores. The cerebellar volume and MRVD of the SCA imaging subgroup were significantly smaller than those of the imaging control group. MRVD was significantly correlated with ICARS and SARA scores, as well as stride velocity variability. Conclusion: SCA gait parameters were characterized by a reduced stride length, slower walking velocity, and longer supporting phase. Additionally, a smaller cerebellar volume correlated with an increased irregularity in gait. Gait characteristics exhibited considerable clinical relevance to hereditary SCA. We conclude that a combination of gait parameters, ataxia scales, and MRVD may represent more objective markers for clinical evaluations of SCA.

Scalogram-Based Gait Abnormalities Classification Using Deep Convolutional Networks for Neurological and Non-Neurological Disorders

Classification of Abnormal Gait Based on Improved Two-Stream CNN

GaitRec-Net: A Deep Neural Network for Gait Disorder Detection Using Ground Reaction Force

Gait Phase Recognition of Lower Limb Based on GA Optimized BP Neural Network

Gait characteristics and clinical relevance of hereditary spinocerebellar ataxia on deep learning

Benchmarking Reliability of Deep Learning Models for Pathological Gait Classification

Deep Learning based Gait Abnormality Detection using Wearable Sensor System

Neural Networks for Pathological Gait Classification Using Wearable Motion Sensors

Detection of Gait Abnormalities caused by Neurological Disorders

Simultaneous time-frequency analysis of gait signals of both legs in classifying neurodegenerative diseases

Deep Neural Network-Based Gait Classification Using Wearable Inertial Sensor Data

Gait disorder classification based on effective feature selection and unsupervised methodology

Multimodal Gait Abnormality Recognition Using a Convolutional Neural Network–Bidirectional Long Short-Term Memory (CNN-BiLSTM) Network Based on Multi-Sensor Data Fusion

Abnormal Gait Recognition Algorithm Based on LSTM-CNN Fusion Network

Classification of Gait Anomalies from Kinect

Cognitive driven gait freezing phase detection and classification for neuro-rehabilitated patients using machine learning algorithms

Emerging trends in gait recognition based on deep learning: a survey

SAGA: Stability-Aware Gait Analysis in constraint-free environments

FP-GCN: Frequency Pyramid Graph Convolutional Network for Enhancing Pathological Gait Classification

Classification of Neurological Gait Disorders Using Multi-task Feature Learning

Towards a diagnostic tool for neurological gait disorders in childhood combining 3D gait kinematics and deep learning