Therapy that moves with you.

NeuroCore is a smart hip brace that trains the muscles behind independent movement, using real-time biofeedback.

The belt is worn during physical therapy sessions. It is soft-shell, adjustable and quiet.

Read our story
The NeuroCore prototype belt on a dark studio background.

How it works

The NeuroCore prototype belt on a dark studio background.

Wear the belt

A lightweight belt, worn during a physical therapy session, places one inertial measurement unit (IMU) at the hip and another at the lower back.

Close-up of the hip sensor pod pocket on the NeuroCore prototype belt.

Read the pelvis

The hip sensor captures raw acceleration and rotation as the child moves.

Close-up of the lower-back sensor pod pocket on the NeuroCore prototype belt.

Read the spine

The second sensor tracks acceleration and rotation at the lower back.

Waveform lines rising from both sensor pods on the NeuroCore prototype belt.

Fuse the signal

A Madgwick filter fuses both streams into a stable, real-time picture of pelvic and spinal motion for the therapist dashboard.

Orange and blue signal traces above the NeuroCore prototype belt.

Score the movement

An AI model, the Isolation Forest, learns each child's movement baseline and calculates a live instability score.

Concentric rings shown at one sensor pod on the NeuroCore prototype belt.

Respond with feedback

Sustained instability triggers a gentle vibration cue, and the full event is logged so therapists can see how movement changed.

Product render of the NeuroCore brace worn at the waist, with arcs indicating vibration cues on both sides.

Haptic feedback at the site of deviation

The system detects a deviation as it happens, and a vibration cue fires at the exact spot on the body where the correction is needed.

Cue location
At deviation site
Response
Immediate self-correction
Reinforcement
Proper movement pattern
Product render of a clinician holding a tablet showing a real-time stability metrics dashboard.

Stability metrics therapists can read

Movement consistency and therapeutic progress are tracked continuously and streamed to therapists through an app.

RMS deviation
Per trial
Peak deviation
Per trial
Recovery time after peak
Per trial

What a clinician sees

A live two-axis view of trunk lean relative to the pelvis, with hip motion subtracted out so only the compensation that matters remains.

Relative lean monitor Good
SIDE LEAN FORWARD LEAN
0

Compensation score

Sum of relative side and forward lean. Under 8 reads as good form. 15 and above triggers a haptic cue.

Relative side lean

+0.0°

Relative forward lean

+0.0°

Simulated motion for demonstration, not a live sensor feed.

Standardized assessment trials

Six structured tasks a therapist can run in clinic, each with its own compensation threshold.

A 30-second quiet stand and a five-rep sit-to-stand should not be graded the same way.

The belt on a dark background with two curved traces rising from the sensors into a single waveform.

Standing

Quiet standing, 30 s
Feet together standing, 30 s

Seated

Seated unsupported hold, 30 s
Seated leg raise, left and right 10°

Dynamic

Sit-to-stand, 5 reps 12°
Standing march, 10 reps 12°
Every trial is scored on

Seven objective measures, computed the same way for every trial, so results compare session to session and side to side.

RMS
Root-mean-square deviation across the trial, the single best summary of overall control.
P95
95th-percentile deviation, so one outlier moment does not skew the read.
Peak
The single largest deviation recorded during the trial.
Outside %
Share of the trial spent beyond that task's compensation threshold.
Path and velocity
Total distance the trunk traveled off-center, and how fast.
Recovery time
Time to settle back within 3° of centered after the trial's peak deviation.
Asymmetry
Left versus right RMS difference on unilateral tasks. Shows which side compensates more.

Run a trial with us

We are looking for therapists and clinics to work with. Tell us about your practice.