Subjects with First-time ischemic stroke who can follow the commands to perform the activities for the study, having motor recovery of Brunnstrom stage 2 and below in lower extremity and having the ability to maintain unsupported sitting in bed were included in the study. Subjects with any other neurological or orthopedic impairments that would limit the ability to undertake the activity, presence of profound spasticity and having history of lower limb injury that can affect the outcome of the study were not included in the study. Subjects with profound proprioception deficit in the paretic lower extremity were also excluded.

Sample size was calculated with a sample of 10 stroke survivors, meeting the inclusion criteria. They were given training in supine similar to the study protocol, to stimulate muscle contraction in gluteus medius. Pre and post intervention surface electromyographic measurement was taken as per the standard protocol. The difference was considered for sample size calculation, with alpha error 0.05, effect size 0.9. The calculated number was 28 subjects in each group.

Subjects meeting these inclusion criteria were included in the study after obtaining a written informed consent from willing participants. Surface electromyography (SEMG) recording were taken from paretic Gluteus medius, Tensor fascialata and Quadriceps femoris muscles for all the participants on the day of inclusion into the study. Baseline SEMG for tensor fascialata and Gluteus medius muscles was recorded with participants in supine lying and for Quadricep femoris muscle, the participants were positioned in high sitting. Electrode position for EMG was standardised based on Surface Electromyography for the Non-Invasive Assessment of Muscles (SENIAM) guidelines  www.seniam.org .As per the guidelines, for Tensor fascialata, the active electrode was placed on lateral aspects of thigh, the reference electrode was placed on Anterior superior iliac spine and the ground electrode was placed on the lateral condyle of the femur and the participants were instructed to attempt to abduct their paretic limb. For Gluteus medius, the active electrode was placed on posterolateral aspect of the muscle belly, the reference electrode was placed on the posterior superior iliac spine, the ground electrode was placed on lateral condyle of femur and the participants were instructed to attempt to abduct their paretic limb.  For Quadricep femoris, the active electrode was placed at the muscle belly, the reference electrode was placed 5cm below the anterior superior iliac spine and ground electrode was placed above the patella and participants were instructed to attempt to extend their paretic limb. SEMG was recorded using Neuro trac software 4.0 from VM (Verity Medical Ltd, United Kingdom) using silver chloride adhesive electrodes.

Participants were randomly allocated by block randomization of four into two groups. The control group had the routine care and the experimental group were given the routine care along with training to hold their paretic lower limb on unstable surfaces in supine lying and high sitting. In routine care participants in both the group received 45 minutes of physiotherapy including facilitatory techniques for upper extremity, lower extremity and trunk and functional training. In addition, the participants in experimental group were instructed to stabilize/hold their paretic limb on a ball, while ensuring the ball remained in place. The interventions were given for 10session over 10 days.

Post-intervention the SEMG recordings from Gluteus medius, tensor fascialata and quadriceps femoris muscles were taken 6 hours after the completion of 10 sessions of therapy, in the similar procedure as base line recording.

Data analysis:

The average SEMG values of each muscle is taken for data analysis. The maximum value among the three trials was considered for testing. The within group and between the group data was tested using paired and un-paired “t” test, with significance level at p≤0.05.