Research Article
IMU-Based Approach to Detect Spastic Cerebral Palsy in Infants at Early Stages
@ARTICLE{10.4108/eetpht.10.5258, author={N Sukhadia and P Kamboj}, title={IMU-Based Approach to Detect Spastic Cerebral Palsy in Infants at Early Stages}, journal={EAI Endorsed Transactions on Pervasive Health and Technology}, volume={10}, number={1}, publisher={EAI}, journal_a={PHAT}, year={2024}, month={3}, keywords={Cerebral Palsy, Spastic Cerebral Palsy, Fidgety Movements, Inertial Measurement Unit, General Movement Assesment}, doi={10.4108/eetpht.10.5258} }
- N Sukhadia
P Kamboj
Year: 2024
IMU-Based Approach to Detect Spastic Cerebral Palsy in Infants at Early Stages
PHAT
EAI
DOI: 10.4108/eetpht.10.5258
Abstract
INTRODUCTION: Cerebral Palsy (CP) is a non-progressive neurological disorder affecting muscle control in early childhood, leading to permanent alterations in body posture and movement. Early identification is crucial for accurate diagnosis and therapy-based interventions. In recent years, an automated monitoring system has been developed to facilitate the health assessment of infants, enabling early recognition of neurological dysfunctions in high-risk infants. However, the interpretation of these assessments lacks standardization and is subject to examiner bias. OBJECTIVES: Many infants with CP exhibit increased tonic stretch reflexes due to Upper Motor Neuron Syndrome (UMNS), resulting from motor neuron damage that disrupts muscle signalling. METHOD: To detect abnormal muscle reactions, our team employed an Inertial Measurement Unit (IMU) sensor, comprising three tri-axial sensors (accelerometer, gyroscope, magnetometer) that capture movement data continuously and unobtrusively. IMU sensors are compact, cost-effective, and have low processing requirements, requiring attachment to the infant's body to measure inter-body part angles. Our team analyzed muscle activity and posture using IMU sensors, collecting tri-axial data from 43 infants in real-time. Additional factors like age, stride length, and leg length were incorporated into the dataset. RESULTS: Our team has applied various supervised machine learning approaches to predict CP in infants due to the limited dataset size, validating models through k-fold cross-validation. Among the models, Naive Bayes (NB) outperformed Logistic Regression (LR), Decision Tree (DT), Linear Discriminant Analysis (LDA), k-Nearest Neighbors (kNN), and Support Vector Machine (SVM), achieving an accuracy of 88%. CONCLUSION: This research contributes to the early detection and intervention of CP in infants, potentially improving their long-term outcomes.
Copyright © 2024 N. Sukhadia et al., licensed to EAI. This is an open access article distributed under the terms of the CC BY-NC-SA 4.0, which permits copying, redistributing, remixing, transformation, and building upon the material in any medium so long as the original work is properly cited.