Abstract TMP102: Robotic Assessment Of Proprioceptive Dysfunction In Children With Perinatal Stroke

Abstract

Objective: Perinatal stroke causes most hemiplegic cerebral palsy. Sensory dysfunction has been ignored and objective measurement tools are limited. Robotic technology can quantify complex sensory function in adult stroke but has not been applied to kids. Methods: Children from the Alberta Perinatal Stroke Project had MR confirmed unilateral perinatal stroke and upper extremity functional deficit. A bilateral exoskeletal robot (KINARM) capable of testing planar upper limb movements in an augmented reality environment was employed. Primary robotic outcomes were 2 dimensional variability, shift, and contraction/expansion scores of a position-matching task (Figure). Blinded clinical measures of sensory function (touch, proprioception, graphesthesia, stereognosis) were scored. Matched controls (age/gender) were tested. Results: Five children (median 14 yrs, 3 male) with perinatal stroke (3 PVI, 2 arterial) were compared to 7 controls. Stroke children demonstrated marked impairment in position matching including variability (6.48±1.4 vs 3.89±0.7cm, p= 0.001) and shift (5.05±2.2 vs 2.00±1.3cm). Contraction/expansion ratios were also abnormal (0.56±0.27 vs 0.31±0.22; p=0.09). Clinical sensory scores were lower but correlated poorly with robotic measures and motor function. Assessments were well tolerated with no adverse events. Conclusion: Robotic quantification of proprioception is feasible in perinatal stroke. Sensitivity and quantification appear superior to clinical exam. Disordered proprioception is an under-recognized component of disability and a novel therapeutic target.