InceptoFormer: A Multi-Signal Neural Framework for Parkinson's Disease Severity Evaluation from Gait

2025-08-06

AI Summary

This study addresses the challenge of automatic severity staging of Parkinson’s disease (PD) using gait dynamics. We propose InceptoFormer, a novel neural framework that jointly models multi-scale temporal patterns and long-range dependencies in gait sequences. Methodologically, it integrates a 1D Inception module—designed to capture multi-scale gait kinematic features—with a Transformer encoder that explicitly models inter-cycle temporal dependencies. To mitigate class imbalance inherent in the Hoehn & Yahr (H&Y) staging scale—particularly for rare severity grades—we incorporate a tailored oversampling strategy. Trained end-to-end with domain-specific preprocessing on a public gait dataset, InceptoFormer achieves 96.6% H&Y staging accuracy, substantially outperforming existing state-of-the-art methods. The implementation is publicly released. By enabling marker-free, low-cost, interpretable, and robust PD progression monitoring, this work establishes a new paradigm for clinical-grade digital biomarker development.

Technology Category

• Intelligent Robots: State Estimation
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

We present InceptoFormer, a multi-signal neural framework designed for Parkinson's Disease (PD) severity evaluation via gait dynamics analysis. Our architecture introduces a 1D adaptation of the Inception model, which we refer to as Inception1D, along with a Transformer-based framework to stage PD severity according to the Hoehn and Yahr (H&Y) scale. The Inception1D component captures multi-scale temporal features by employing parallel 1D convolutional filters with varying kernel sizes, thereby extracting features across multiple temporal scales. The transformer component efficiently models long-range dependencies within gait sequences, providing a comprehensive understanding of both local and global patterns. To address the issue of class imbalance in PD severity staging, we propose a data structuring and preprocessing strategy based on oversampling to enhance the representation of underrepresented severity levels. The overall design enables to capture fine-grained temporal variations and global dynamics in gait signal, significantly improving classification performance for PD severity evaluation. Through extensive experimentation, InceptoFormer achieves an accuracy of 96.6%, outperforming existing state-of-the-art methods in PD severity assessment. The source code for our implementation is publicly available at https://github.com/SafwenNaimi/InceptoFormer

Problem

• Evaluating Parkinson's Disease severity via gait dynamics analysis
• Addressing class imbalance in PD severity staging
• Improving classification performance for PD severity assessment

Innovation

• Inception1D captures multi-scale temporal gait features
• Transformer models long-range dependencies in gait sequences
• Oversampling strategy addresses PD severity class imbalance