Feasibility of using acceleration-derived jerk to quantify bimanual arm use

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Wearable Technology

Background:

• Accelerometers are widely used in rehabilitation to assess patient progress, particularly for individuals with neurologic disorders.
• Existing metrics like use ratio (UR) and magnitude ratio (MR) rely on proprietary "activity counts," limiting cross-device comparability.
• There is a need for device-independent metrics derived from raw accelerometer data to analyze arm movement post-neurologic injury.

Purpose of the Study:

• To develop and validate novel metrics based on raw accelerometer data to quantify arm movement in individuals with neurologic disorders.
• To introduce jerk ratio (JR) and JR50 as complementary measures to existing activity count metrics.
• To assess the efficacy of these new metrics in differentiating movement patterns between children with cerebral palsy (CP) and typically developing (TD) peers.

Main Methods:

• Calculated jerk (the derivative of acceleration) from raw accelerometer data to analyze arm movement.
• Defined jerk ratio (JR) as the ratio of dominant to non-dominant arm jerk magnitude.
• Evaluated JR50 (the 50th percentile of JR) and compared it with activity count metrics (UR, MR) in five children with hemiplegic CP and five TD children.

Main Results:

• JR50 successfully differentiated between the CP and TD cohorts, indicating increased reliance on the dominant arm in the CP group (CP: 0.578 vs. TD: 0.506).
• Jerk metrics effectively quantified changes in arm use during and after Constraint-Induced Movement Therapy (CIMT) in the CP cohort.
• JR demonstrated strong correlations with UR and MR (r = -0.92, 0.89) in the CP cohort, and JR50 showed high repeatability in the TD cohort (0.945).

Conclusions:

• Acceleration-derived jerk metrics, specifically JR and JR50, can effectively capture differences in arm motion between TD and CP populations.
• These novel, device-independent metrics correlate well with existing activity count measures, offering a standardized approach to rehabilitation assessment.
• The open-source availability of the JR calculation code aims to foster collaboration and advance the use of wearable technologies in rehabilitation research.