**1. Introduction**

Cerebral palsy (CP) with a prevalence of about 2.2/1000 live births is the most common cause of motor disability in childhood [1]. Changes in gait of children with spastic cerebral palsy are often affected by symptoms of spasticity and lower extremity muscle weakness, which limit the patient's ability to walk [2].

The use of ankle–foot orthoses (AFOs) is widely recommended to prevent the development or progression of deformity and to improve dynamic e fficiency of the child's gait [3]. There are a wide variety of AFOs used in clinical practice, which are di fferentiated depending on their design, the material used and the sti ffness of that material. Changing any of these three elements will alter the AFO control and this may have influence on the patient's gait [4]. Because of heterogeneity among study designs, AFO types, control conditions, and outcome measures, main e ffects of AFO on gait of children with CP are unclear. Furthermore, most studies in this area are focused on evaluating a single, specific outcome measure (e.g., step length, ankle range of motion, or knee flexion) rather than the overall e ffect of AFO use on patient's gait [5].

Three-dimensional gait analysis (3DGA) is commonly used in clinical practice and scientific research for the purpose of objective assessment and description of gait disorders, as well as to plan and evaluate the treatment of children with CP. This method provides a large amount of interdependent data and variables concerning spatio-temporal gait parameters, joint motions (kinematics), as well as joint movement and power (kinetics) in three planes [6]. Given the di fferent gait patterns and pathologies in children with CP, a global analysis is essential in clinical practice. For this purpose, it is useful to summarize and present the results from 3DGA as a single, quantitative, numerical measure that reflects the patient's gait.

One of the models designed to obtain a single measure of the quality of a gait pattern is the gait deviation index (GDI), that measures the subject's gait deviation from a reference, normative data. It is a score derived from 3DGA, which provides a numerical value that expresses overall gait pathology (range 0–100, where 100 and above indicates absence of gait pathology). Every 10-point decrease in the GDI corresponds to one standard deviation from the mean of healthy controls. GDI is calculated based on kinematic parameters [7].

Another gait index, the gait variability index (GVI), is computed based on spatio-temporal parameters. The GVI was intended to be applicable to di fferent patient's groups and diseases severities. It was constructed as a complex measure of nine weighted spatio-temporal gait variables seen in relation to a reference population. GVI value ≥ 100 indicates a similar level of gait variability as the reference population, and each 10-point reduction in the score corresponds to one standard deviation from the reference mean, where lower scores indicate increased gait variability [8,9].

Danino et al. have demonstrated that improvement in temporal and kinematic parameters in spastic diplegic CP children using AFO are not reflected by some gait indices, including GDI [10]. Therefore, the aim of this study was to assess whether the GDI and the GVI reflect changes in temporal and kinematic parameters in spastic hemiplegic CP children wearing AFO.

### **2. Materials and Methods**
