Knee flexion amplitude in swing phase was related to the size of the knee velocity at the end of stance phase and in turn to paresis of the ankle plantarflexors and passive stiffness in the knee extensors. In contrast there was no association with measures of stretch reflex hyperexcitability or spasticity 
Muscle paresis affected multiple muscle groups in people with SP, being greatest in the ankle dorsiflexors. Muscle paresis may reflect both a decrease in corticospinal drive and secondary muscle atrophy 
and is often the main determinant in limiting functional movements in people with an upper motor neuron lesion [15,16]
. Indeed there was a significant negative correlation between the degree of strength in the knee extensors and the sit-to-stand time (R2
= 0.47) and a positive correlation between the strength in the ankle plantarflexors and normal walking speed (R2
= 0.61). The presence of significant muscle weakness that can limit functional movements in SP is in contrast with the view that paresis is relatively mild in this patient group and that function is mainly limited by spasticity 
. This may in part reflect the fact that active movements as opposed to objective tests of muscle strength have been assessed to date.
We found that a high knee extensor moment during pre-swing limited swing phase knee flexion and was associated with a higher degree of passive stiffness in the knee extensors. Increases in passive stiffness could reflect changes in the connective tissue, muscle architecture and/or intrinsic muscle proteins 
and have been reported in other conditions affecting the central nervous system such as multiple sclerosis or stroke [18–21]
. The change in passive stiffness was more marked for the plantarflexors than for the knee extensors; this may reflect different patterns of use and differences in the amount or pattern of intramuscular connective tissue. However, in those people with SP an above average passive stiffness in the knee extensors was associated with a further limitation in knee flexion.
Hyperexcitable reflexes were seen when higher velocity stretches were applied at rest in people with SP, this is a hallmark of spasticity 
. Stretch reflex size can decrease following a stretch 2–10 s earlier, termed post activation depression. Post activation depression can be reduced in people with spasticity 
and given the inter-stretch interval of 6.5 s, this could also contribute to the high stretch reflex size observed. The increase in stretch-evoked muscle activity and reduction in post-activation depression reflects in part a decrease in inhibitory activity within spinal cord circuits 
. The degree of reciprocal inhibition, for example, from the ankle dorsiflexors to the ankle plantarflexors is reduced in people with spasticity 
. In healthy participants the level of reciprocal inhibition is not static. With the onset of contraction of the plantarflexors the reciprocal inhibition of that muscle decreases 
. This could explain why, compared to the resting condition, identical stretches of a pre-activated muscle resulted in EMG evoked activity and total stiffness that was similar between groups. Such a normalisation of muscle stiffness and stretch reflex activity in people with spasticity [26,27]
raises the question as to the role of spasticity in limiting movement, particularly if the stretched muscle is pre-activated such as during an eccentric contraction.
No relationship between spasticity of the knee extensors and the degree of knee flexion was seen. Stretch reflex size is normally modulated by postural set and the phase of walking 
. In people with spasticity such modulation is decreased 
. Therefore, a lack of correlation between stretch reflex size recorded in supine and parameters of walking may reflect differences in stretch reflex properties during functional tasks such as walking. There may also be differences between how the response to unexpected perturbations (as delivered in the current study) and expected perturbations as occurs during voluntary movement. Stretch reflexes were only recorded at one velocity (60°/s). Non-linearity in response to different stretch velocities may further mean that the contribution of stretch reflexes at the velocities achieved while walking (~140°/s) was underestimated.
The current work found that the limitation in knee flexion velocity and peak to peak knee amplitude during swing phase was associated with muscle paresis and passive stiffness; explaining ~50% of the variance, rather than static measures of spasticity. This is in keeping with recent work showing that excessive activation of the rectus femoris in pre-swing while walking is less common in children with SP compared to children with cerebral palsy and spastic diplegia [8,29]
. This study has only investigated associations between variables and it remains unclear whether ankle strength and knee extensor passive stiffness actually cause a reduction in knee movement while walking. However, this study highlights the need to determine muscle strength objectively, to assess the contribution of stretch reflexes to limb stiffness in both the resting and active participant and to differentiate between limb stiffness caused passive stiffness and spasticity; this will have a direct impact on potential therapeutic approaches. While spasticity may be amenable to pharmacological interventions; physical interventions such as stretching or splinting may be more applicable to target changes in passive stiffness.