• Action Research Arm Test (ARAT)
  
• Nine Hole Peg Test (9HPT)
  

• peak torque around the elbow joint in elbow flexion and extension
  
• grip force
  
• pinch force
  
• smoothness of movement during turning a cranked wheel
  
• reciprocal inneveration of biceps and triceps during turning of a cranked wheel.
  

Neuromuscular weakness occurs frequently after stroke and the processes underlying recovery are still poorly understood. Accepted practice in UK physiotherapy is to avoid training of muscle strength after stroke but there is preliminary evidence that it might be effective.

An observer-blind randomised pilot clinical trial (Phase II). Subjects will be within 3 months of first stroke with some voluntary movement in the paretic upper limb.

A Research Physiotherapist, blinded to measurement, will recruit subjects, allocate subjects to one of three intervention groups using sequentially numbered sealed envelopes containing previously allocated intervention cards and provide interventions. Ten participants will be recruited to each group therefore the trial will recuit 30 participants. The Research Assessor, blinded to intervention allocation will undertake all measurements.

Conventional therapy (control) will be provided as normal for the clinical setting, the additional conventional therapy (experimental 1)or functional strength training (experimental 2) will be provided for one hour, five times a week for six weeks. Subjects in the two experimental groups will also receive the conventional therapy standard in their clinical setting.

Blinded measurement will be made before randomisation, at the end of intervention and 12-weeks thereafter. Primary outcomes are the Action Research Arm Test (ARAT)and the Nine Hole Peg Test (9HPT). The secondary outcomes are peak torque around the elbow joint in elbow flexion and extension, grip force, pinch force, smoothness of movement during turning a cranked wheel and reciprocal inniveration of biceps and triceps during turning of a cranked wheel.

Data will be analysed using descriptive statistics and the estimation of standard deviations will be used to inform a power calculation to estimate sample size for a Phase II randomised controlled trial.