NEUROGAIT
MULTIMODAL OBSERVER
Multimodal movement assessment

NEUROGAIT

Movement is information.

↓ SCROLL TO OBSERVE
VIDEO + IMU
02

MODALITIES

Two synchronized views of the same movement.

08

FUSED FEATURES

Synchronized eight-dimensional movement state vector

left_ankle_velocity
right_ankle_velocity
left_knee_angle
right_knee_angle
stride_width
accel_rms
gyro_x_var
gyro_z_var
VIDEO
IMU
±0.1s

TEMPORAL ALIGNMENT

MOVEMENT TIMELINE EPISODE WINDOW: 05.2s — 06.8s
05.2s 06.8s
MODEL-DETECTED EPISODE

POSSIBLE FoG EPISODE

05.2s — 06.8s

Non-diagnostic assessment derived from continuous gait window classification.

92%
CONFIDENCE
Mean class-1 probability across episode windows
STRUCTURED EVIDENCE
type: FoG
confidence: 0.92
interval: 05.2s — 06.8s
primary_cue: left_ankle_velocity
↓

AMAZON BEDROCK

ML detects. AI explains.
“A possible freezing episode was detected during this interval (05.2s — 06.8s), primarily associated with altered left ankle movement.”
Explaining structured machine evidence. Grounded explanation without diagnostic claims.
Movement.
Measured.
Explained.

NEUROGAIT

Made with love by

Devika Anup × Shriram
2026

BEGIN YOUR ASSESSMENT

Upload your files and run the multimodal assessment

1. Kinematic Video (.mp4) ⓘ
Choose video file
MP4 (Max 100 MB)
📹 PDFE01_1.mp4 (80.0 MB)
2. Inertial IMU Sensor (.txt, .csv) ⓘ
Choose sensor file
TXT or CSV (Max 10 MB)
📊 SUB01_1.txt (1.8 MB)