Upper limb sensory processing: development of a robotic paradigm

Stroke is the third leading cause of death and disability in the world. Following stroke, upper limb motor and somatosensory impairments are frequently present. In the acute stage after stroke, up to 50% of survivors show exteroceptive impairments, and up to 63% show proprioceptive and sensory processing impairments. Exteroception is the detection of touch, pain, and temperature on the skin. Proprioception is the detection of joint position and movement. Sensory processing is interpretation of exteroceptive and proprioceptive signals to discriminate between and recognize somatosensory information. At six months after stroke, exteroception already shows nearly full recovery, while proprioception and sensory processing only show a proportional recovery of 69-86%. Somatosensory impairments negatively affect functional outcome after stroke, so accurate assessment and effective treatment is crucial. However, there is a lack of standardized assessments and evidence-based treatments in the literature. Robotics are interesting tools to achieve these aims, as they allow objective and detailed assessment, as well as intensive treatment.

In the first chapter, we filled an important gap in the literature by developing and validating a robot-based assessment of upper limb sensory processing. A three-step assessment was developed, consisting of exploration, reproduction, and identification of geometrical shapes. We demonstrated discriminative validity by showing that 20 participants with chronic stroke performed significantly worse compared to 60 healthy participants. Convergent validity was established for participants with chronic stroke by showing moderate to high correlations with a sensitive clinical assessment of sensory processing. In the second chapter, we also demonstrated that performance on this novel assessment decreases with increasing age and worse cognitive function in 96 healthy participants.

In the third chapter, we described a novel clinical assessment of the combined sensorimotor function. A gold-standard assessment of upper limb activity, namely the action research arm test (ARAT), was adapted to be performed with the eyes closed. In this adaptation, called the sensorimotor ARAT or sARAT, we evaluate the effect of somatosensory impairments on upper limb motor activity capacity in the absence of vision. Discriminative validity was established by showing worse performance in 22 participants with chronic stroke compared to 60 healthy participants. In addition, we also found high correlations with upper limb motor and somatosensory function in participants with chronic stroke, confirming convergent validity.

In the fourth chapter, we described a novel robot-assisted upper limb somatosensory treatment protocol which was based on the SENSe training principles. Ten participants with chronic stroke and residual upper limb sensorimotor impairments completed 10 one-hour training sessions over four weeks, consisting of four novel robot-assisted serious games targeting upper limb proprioception and sensory processing. After completing the training program, we found small, but favorable effects on upper limb motor, somatosensory, and combined sensorimotor function.

In conclusion, we described novel assessment and treatment protocols for upper limb sensory processing. This doctoral thesis forms a strong basis for future power-based research on the recovery of upper limb sensory processing following stroke.