Better assessment of abnormal neuro-mechanics can be obtained via designed probing motions. (a) Stereotypical probing motions executed by experimental subjects. Non-instructed subjects manipulated the hypertonic arm model mainly with random motions. Instructed subjects were encouraged to use the two task-specific probing motions (pulling and semi-elliptical motions - see ‘Selection of specific probing motions’). (b) Mean and standard error (n =6 per group) of the assessment reliability as calculated in ; specific probing motions significantly improved the assessment. Note that a random guess approach has a reliability of 1/3.