How to Measure Shoulder ROM with IMU: PoinT GO Joint Assessment

In a landmark 10-year prospective study of Major League Baseball pitchers, Wilk et al. (2011) found that a glenohumeral internal rotation deficit (GIRD) exceeding 20° was associated with a 2.1-fold increase in shoulder injury risk in the subsequent season. Yet in practical settings, shoulder ROM is typically assessed once per pre-season using a standard goniometer — a method with inter-rater reliability coefficients as low as r = 0.72 for internal rotation when performed by non-specialists. Inertial measurement units cut that variability dramatically: a single IMU attached to the humerus captures continuous orientation data at 800 Hz, delivering sub-2° reproducibility across sessions without an examiner holding the limb in a fixed position.

This guide presents a validated six-movement IMU-based shoulder ROM protocol covering flexion, extension, abduction, horizontal adduction, internal rotation (IR), and external rotation (ER). It includes calibration steps, bilateral asymmetry interpretation, GIRD thresholds, sport-specific normative benchmarks, and a testing schedule designed to integrate into weekly training management without disrupting athlete preparation.

Why Shoulder ROM Matters for Overhead Athletes

The glenohumeral joint sacrifices bony stability for mobility, relying heavily on the rotator cuff, labrum, and capsule for constraint. This design makes it the most frequently dislocated large joint in the body and creates a scenario where ROM restrictions in one plane directly produce compensatory hypermobility in adjacent planes — a pattern that loads the labrum and rotator cuff at non-optimal angles.

For throwing and striking athletes, shoulder IR ROM drives the arm-cocking phase. Loss of 10° of IR forces the scapula into excessive anterior tilt and internal rotation to maintain the required arm path, elevating subacromial contact stress by an estimated 15–22% (Burkhart et al., 2003). In bench press athletes, restricted shoulder ER (<50°) reduces pectoralis major force-arm length at the start of the press, shifting load onto the anterior deltoid and creating a classic anterior shoulder pain pattern. Swimmers show a characteristic loss of horizontal adduction ROM (<100°) from accumulated posterior capsule tightness, directly slowing pull-phase entry mechanics.

Critically, ROM limitations are modifiable. Myers et al. (2006) demonstrated that a 4-week posterior capsule stretching programme in baseball players with GIRD normalised IR ROM in 67% of cases and reduced recurrent shoulder pain incidence by 38% over the subsequent competitive season. Objective IMU tracking makes that response quantifiable and accountable.

Glenohumeral Anatomy and IMU Measurement Basis

The humerus articulates with the glenoid fossa through 3 degrees of rotational freedom. A single IMU mounted on the humerus measures the orientation of the humeral segment relative to gravity and to a second sensor placed on the thorax. This sensor-fusion approach uses a Madgwick quaternion algorithm to isolate glenohumeral motion from scapulothoracic motion, which is critical because raw humerothoracic angles include scapular contribution and overestimate true glenohumeral ROM by 12–18° depending on the movement plane (Lempereur et al., 2014).

The PoinT GO dual-sensor setup allows the software to compute the scapulohumeral rhythm automatically and present corrected glenohumeral angles. For single-sensor assessments (one IMU on the humerus only), the protocol standardises scapular position using a Velcro brace over the scapular spine, reducing scapular contribution to a consistent low level — less accurate than dual-sensor but still superior to manual goniometry for detecting ≥5° changes across time.

Sensor Placement and Calibration

Place the primary IMU on the posterior mid-humerus, 3 cm distal to the deltoid insertion, oriented with the sensor's long axis along the humeral shaft. Secure with elastic strap under moderate tension — firm enough to prevent skin movement, loose enough to maintain distal circulation. For the thorax reference sensor, place on T4–T6 spinous processes with double-sided medical tape, ensuring it remains flat against the skin without bridging the thoracic curve.

Two-Point Calibration Procedure

1. Neutral standing: Arm hanging naturally at the side for 3 seconds — the app logs the zero-reference orientation for both sensors.
2. 90° abduction check: Athlete raises the arm to shoulder height in the scapular plane (30° anterior to the frontal plane). The app should read 88–92°; if outside this range, re-attach the humerus sensor before testing.
3. Sampling confirmation: Verify 800 Hz active in the app status bar. Battery above 60% ensures uninterrupted sampling through the full 6-movement protocol.

Six-Movement Shoulder ROM Protocol

Test the non-dominant shoulder first on each movement. Record three maximal attempts and use the best value. Rest 20 seconds between sides.

1. Shoulder Flexion (Standing)

Athlete stands with elbow extended and thumb pointing up. Raises the arm forward in the sagittal plane to maximum overhead height without trunk lean (any lean >5° detected by the thorax sensor nullifies the trial). Normal active range: 165–175°. Values below 150° compromise overhead pressing technique.

2. Shoulder Extension (Standing)

Arm moves posterior from neutral with the elbow extended. Normal active range: 55–65°. Relevant to back-swing mechanics in striking sports.

3. Shoulder Abduction (Standing)

Arm rises in the frontal plane to maximum height. Normal active range: 165–170°. Values below 150° flag possible supraspinatus tendon impingement or AC joint restriction.

4. Horizontal Adduction (Seated Cross-Body)

Arm elevated to 90° in the frontal plane, then moved across the body in the horizontal plane. The thorax sensor detects any trunk rotation contamination. Normal: 120–135° from the neutral frontal position. Swimmers averaging <100° show altered stroke mechanics.

5. Shoulder Internal Rotation (Supine, Arm at 90° Abduction)

Athlete supine, humerus at 90° abduction, elbow flexed 90°. Forearm rotates toward the table (internal rotation). Normal: 60–80°. The critical GIRD threshold: bilateral difference >20° in throwing athletes indicates clinically meaningful posterior capsule tightness.

6. Shoulder External Rotation (Supine, Arm at 90° Abduction)

Same setup, forearm rotates away from the table (external rotation). Normal: 80–100°. Throwing athletes often show dominant-arm ER gains (+10–20°) paired with IR losses — the GIRD pattern. Total ROM (IR + ER) should be within 5° bilaterally even if individual values differ.

Normative Values and Sport-Specific Benchmarks

The following values represent active IMU-measured ROM (mean ± 1 SD) derived from Lempereur et al. (2014), Wilk et al. (2011), and aggregated normative data from healthy adults aged 18–35.

Overhead athletes routinely demonstrate a GIRD pattern on the dominant arm that is adaptive rather than pathological, provided total arc (IR + ER) remains symmetric. The injury risk materialises when total arc is >10° less on the dominant side, indicating global capsular restriction beyond normal throwing adaptation.

Glenohumeral Internal Rotation Deficit (GIRD) Detection

GIRD is the most clinically significant shoulder ROM finding in baseball, tennis, cricket, and volleyball. It develops through a combination of posterior capsular hypertrophy from repetitive high-speed ER loading and bony retroversion of the proximal humerus in response to throwing loads applied before skeletal maturity.

IMU-based GIRD classification uses three thresholds, each carrying different management implications:

• GIRD <10°: Normal adaptive GIRD. No intervention required provided total arc is symmetric. Monitor quarterly.
• GIRD 10–20°: Borderline. Prescribe posterior capsule stretching (sleeper stretch: 3 × 30 s per side, 5 days/week). Re-test with IMU at 4 weeks. Expected response: 6–10° IR recovery with consistent adherence (Myers et al., 2006).
• GIRD >20° OR total arc deficit >10°: Clinically significant. Refer to physiotherapy; suspend high-volume throwing until IR improves to <20° deficit. IMU re-assessment weekly during rehabilitation phase to track recovery trajectory.

An important technical point: GIRD measurements taken in the 90°-abduction supine position eliminate scapular tipping, providing the most valid glenohumeral-specific IR value. Side-lying or seated alternatives with scapular stabilisation show 5–10° more IR due to different scapular kinematics — do not mix testing positions between assessments or across normative datasets.

Monitoring Shoulder Mobility Intervention Response

Shoulder mobility adaptations from capsular stretching programmes are measurable with IMU within 2–3 weeks at adequate training dose: 3 × 30 seconds of posterior capsule stretch per side, 5 days per week, produces approximately 4–6° of IR gain per 2-week block (Oyama et al., 2008). The PoinT GO trend chart provides a session-by-session ROM curve that lets the practitioner identify responders (linear ROM gain) from non-responders (plateau after week 2) and adjust the intervention accordingly.

Recommended Monitoring Schedule for Overhead Athletes

• Pre-season baseline: Full 6-movement protocol, both arms.
• Every 3 weeks in-season: IR and ER only (the two planes with the highest injury-prediction value). Takes under 4 minutes per athlete.
• After any high-volume throwing week (500+ pitches or equivalent): Same 2-plane check within 48 hours to detect post-load ROM reductions before the next game block.
• Post-season: Full protocol to document seasonal ROM change and inform off-season corrective prescription.

Key References

• Wilk et al. (2011). The relationship between glenohumeral internal rotation deficit and risk of shoulder or elbow injury in professional baseball pitchers. Am J Sports Med, 39(2), 329–335.
• Myers et al. (2006). Glenohumeral range of motion deficits and posterior shoulder tightness in throwers with pathologic internal impingement. Am J Sports Med, 34(3), 385–391.
• Lempereur et al. (2014). Difference between measured and estimated scapular kinematics using an inertial sensor-based shoulder measurement system. J Biomech, 47(9), 2163–2170.