Training and Testing Database for IMU Based Gait Analysis Methods (IMU-GAIT)

The goal of this study is to establish a high-quality, synchronised dataset of gait events (GE) by simultaneously collecting inertial measurement unit (IMU) data and validated ground truth detections using a Vicon motion capture system. The primary objective is to address existing limitations in GE detection - such as poor generalisability, limited data diversity, and lack of precise synchronisation - through a rigorous protocol that ensures accuracy and transparency. The experiment is structured in three phases. First, Vicon-derived GE will be validated and refined using complementary modalities (force plates and video recordings). Next, deep learning (DL) algorithms will be developed and evaluated for GE detection directly from IMU data, with Vicon annotations serving as ground truth. Finally, the impact of differences in GE timing on spatiotemporal gait parameters (SGP) will be analysed to assess the feasibility of using IMU-only systems for reliable gait analysis.

By achieving these objectives, the study aims to improve the accuracy of GE detection from wearable sensors and enable more accessible, scalable, and reliable gait analysis outside the laboratory environment.

Study Overview

• Status: Not yet recruiting
• Conditions: Healthy Subjects or Volunteers

Detailed Description

This project investigates the development of accurate and reliable gait event (GE) detection methods using wearable inertial measurement units (IMUs), validated against goldstandard motion capture data (Vicon). Gait analysis plays a central role in understanding human locomotion and has important clinical applications in rehabilitation, neurology, orthopaedics, and fall-risk assessment. However, current IMU-based approaches are limited by synchronisation issues, small or homogeneous datasets, and insufficient validation against ground truth. This study addresses these gaps by systematically collecting and validating gait data in healthy participants. Data collection will be performed at the Brubotics Rehabilitation Research Center (BRRC) motion analysis laboratory. Participants will complete walking trials at different speeds (slow, self-selected, and fast speeds) along a standardised 10 m pathway. Reflective markers will be placed on anatomical landmarks, and a sacrum-mounted IMU will capture inertial signals. GE will be simultaneously recorded with the Vicon system, complemented by video and force plate data for validation. The study is organised into three phases. Phase 1 validates and refines Vicon-detected GE using complementary modalities. Phase 2 develops and evaluates deep learning algorithms for IMU-based detection, including the exploration of self-supervised learning. Phase 3 examines how differences in GE timing influence spatiotemporal gait parameters (e.g., step time, cadence, asymmetry), with the goal of establishing whether IMU-only systems can serve as reliable alternatives to motion capture. Ultimately, this project will deliver a robust dataset and algorithmic framework that improve the precision and generalisability of IMU-based GE detection.

• Study Type: Observational
• Enrollment (Estimated): 150

Contacts and Locations

• Study Contact: Name: Redona Brahimetaj, Doctor of Engineering Sciences; Phone Number: 0032465597448; Email: redona.brahimetaj@vub.be

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: Yes

• Sampling Method: Non-Probability Sample

Study Population

Healthy Adults, age 18-65

Description

Inclusion Criteria:

• Healthy subjects with no motor impairments that disrupt the walking pattern.
• No history of pain in the lower limbs in the past 6 months.
• No history of lower limbs injuries/surgeries in the past 6 months.
• Language: Dutch and/or English and/or French speakers.
• Age 18-65 years old
• Subjects must be able to understand the instructions and to answer questions. Additionally, they should be able to; signal pain, fear, discomfort; give inform consent.

Exclusion Criteria:

• Persons with comorbidity that could hinder the study (e.g.: unstable cardiovascular system disorders, lung disorders, severe osteoporosis).
• Individuals with metal implants or skin conditions that would make sensor or marker placement difficult.

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort
Healthy subjects; Age 18-65 years old

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Heel-Strike Timing Error: Force Plates vs Vicon	Absolute time difference between heel-strike detected by synchronized force plates (reference) and heel-strike detected by the Vicon motion-capture system. Force-plate heel-strike is defined as the first frame with vertical ground-reaction force. Vicon heel-strike is the kinematic event labeled "heel-strike" per the lab's standard pipeline. Error is computed per step, then summarised as overall mean ± SD across participant(s). Lower values indicate better agreement. Unit of Measure: milliseconds (ms)	4 years
Toe-Off Timing Error: Force Plates vs Vicon	Absolute time difference between toe-off from force plates and toe-off detected by Vicon. Error is computed per step, averaged within participant, then summarized as overall mean ± SD. Lower values indicate better agreement. Unit of Measure: milliseconds (ms)	4 years
Event Agreement (%): Vicon vs Video - Heel-Strike	Percentage of heel-strike events for which Vicon and frame-by-frame video annotation indicate the same event within ms tolerance window (several tolerance windows will be explored). Agreement is computed per participant and summarized as mean ± SD across participant(s). Higher values indicate better agreement. Unit of Measure: percent (%)	4 years
Event Agreement (%): Vicon vs Video - Toe-Off	Percentage of toe-off events for which Vicon and video annotation match within a tolerance window (several tolerance windows will be assessed). Computed per participant; summarized as mean ± SD per participant(s). Higher values indicate better agreement. Unit of Measure: percent (%)	4 years

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
IMU Heel-Strike Detection Accuracy	Proportion of correctly classified heel-strike events by the IMU-based deep learning model relative to Vicon ground truth (correct classifications / total events), computed per participant and summarized as mean ± SD across participant(s). Reference & tolerance: Vicon event labels; a match is counted when IMU and Vicon events fall within same detection frame or within a tolerance window we will evaluate. Unit of Measure: percent (%)	4 years
Step Time Accuracy	Description: Absolute difference in step time between IMU-derived and Vicon-derived gait event detection. Step time is defined as the time between consecutive heel-strikes of opposite feet. Unit of Measure: milliseconds (ms)	4 years
IMU Toe-Off Detection Accuracy	Proportion of correctly classified toe off events by the IMU-based deep learning model relative to Vicon ground truth (correct classifications / total events), computed per participant and summarized as mean ± SD across participant(s). Reference & tolerance: Vicon event labels; a match is counted when IMU and Vicon events fall within same detection frame or within a tolerance window we will evaluate. Unit of Measure: percent (%)	4 years
IMU Heel-Strike Sensitivity (Recall)	True positives / (true positives + false negatives) for heel-strike detection vs Vicon, computed per participant(s); summarized as mean ± SD. Higher is better. Unit of Measure: percent (%) Reference & tolerance: Vicon labels; match within a tolerance window to be analysed	4
IMU Heel-Strike Specificity	True negatives / (true negatives + false positives) for heel-strike detection vs Vicon, computed per participant(s); summarized as mean ± SD. Higher is better. Reference & tolerance: Vicon labels; match within a tolerance window to be analysed Unit of Measure: percent (%)	4 years
IMU Toe-Off Sensitivity (Recall)	TP / (TP + FN) for toe-off detection vs Vicon, computed per participant(s); summarized as mean ± SD. Higher is better. Reference & tolerance: Vicon labels; match within a tolerance window to be analysed Unit of Measure: percent (%) Time Frame: 4 years	4 years
IMU Toe-Off Specificity	TN / (TN + FP) for toe-off detection vs Vicon, computed per participant(s); summarized as mean ± SD. Higher is better. Reference & tolerance: Vicon labels; match within a tolerance window to be analysed Unit of Measure: percent (%) Time Frame: 4 years	4 years
Cadence Accuracy	Description: Absolute difference in cadence between IMU-derived and Vicon-derived measurements. Cadence is defined as the number of steps per minute. Unit of Measure: steps per minute (steps/min)	4 years
Step Length Accuracy	Description: Absolute difference in step length between IMU-derived and Vicon-derived measurements. Step length is defined as the distance between heel-strikes of opposite feet. Unit of Measure: meters (mm)	4 years
Step-Time Asymmetry Accuracy	Description: Absolute difference in step-time asymmetry index between IMU-derived and Vicon-derived data. The asymmetry index is calculated as the relative difference between left and right step times. Unit of Measure: percent (%)	4 years
Stride Time Accuracy	Description: Absolute difference in stride time between IMU-derived and Vicon-derived data. Stride time is defined as the time between heel-strikes of the same foot. Unit of Measure: milliseconds (ms)	4 years
Stride Length Accuracy	Description: Absolute difference in stride length between IMU-derived and Vicon-derived measurements. Stride length is defined as the distance between consecutive heel-strikes of the same foot. Unit of Measure: meters (mm)	4 years
Stance Time Accuracy	Description: Absolute difference in stance time between IMU-derived and Vicon-derived data. Stance time is defined as the time from heel-strike to toe-off of the same foot. Unit of Measure: milliseconds (ms)	4 years
Double-Support Time Accuracy	Description: Absolute difference in double-support time between IMU-derived and Vicon-derived measurements. Double-support time refers to the duration during which both feet are in contact with the ground. Unit of Measure: milliseconds (ms)	4 years

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Weight	Body weight will be measured using a calibrated digital scale, with participants wearing light clothing and no shoes. Weight will be recorded once at baseline and at scheduled follow-ups. Results will be summarized as mean ± standard deviation across participants. Unit of Measure: kilograms (kg)	4 years
Height	Standing height will be measured using a wall-mounted stadiometer, with participants standing upright, without shoes, and heels together. Results will be summarized as mean ± standard deviation across participants. Unit of Measure: meters (mm)	4 years
Body Mass Index (BMI)	BMI will be calculated as weight (kg) divided by height squared (m²). This provides a standardized measure of relative weight adjusted for height. Results will be summarized as mean ± standard deviation across participants. Unit of Measure: kilograms per square meter (kg/m²)	4 years

Collaborators and Investigators

• Sponsor: Vrije Universiteit Brussel

Study record dates

Study Major Dates

• Study Start (Estimated): December 1, 2025
• Primary Completion (Estimated): November 30, 2029
• Study Completion (Estimated): December 31, 2029

Study Registration Dates

• First Submitted: August 27, 2025
• First Submitted That Met QC Criteria: November 17, 2025
• First Posted (Actual): November 19, 2025

Study Record Updates

• Last Update Posted (Actual): November 19, 2025
• Last Update Submitted That Met QC Criteria: November 17, 2025
• Last Verified: October 1, 2025

More Information

Terms related to this study

• Keywords: Healthy adults; Gait analysis; Wearable sensors; Deep learning; Machine learning; Gait events; Spatiotemporal gait parameters (SGP); Inertial measurement units (IMU); Motion capture (Vicon)
• Other Study ID Numbers: 25235_IMU-GAIT; 25235 (Registry Identifier: 25235)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED
• IPD Plan Description: Still undecided

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No