Linear and nonlinear analysis of brain dynamics in children with cerebral palsy

This study was carried out to determine linear and nonlinear changes of brain dynamics and their relationships with the motor dysfunctions in CP children. For this purpose power of EEG frequency bands (as a linear analysis) and EEG fractality (as a nonlinear analysis) were computed in eyes-closed resting state and statistically compared between 26 CP and 26 normal children. Based on these characteristics accuracy of the classification between the two groups was obtained by enhanced probabilistic neural network (EPNN). Severity of gross motor and manual disabilities was determined by standard systems and the relation between the deficient brain dynamics and severity of the motor dysfunctions was obtained by Pearson's correlation coefficient. A definitely higher delta and lower theta and alpha powers, and higher EEG complexity in CP patients. As such a high accuracy of 94.8% in distinguishing the two groups was obtained. Moreover significant positive correlations were found between beta power and severity of manual disabilities and gross motor dysfunctions in the boys with CP. It is concluded that the obtained brain dynamics’ characteristics are useful in diagnosis of CP. Furthermore severity of the motor dysfunctions in boys with CP could be evaluated by the beta activity.

► Power of EEG bands and EEG fractality were compared between children with CP and normal children.
► The first study of nonlinear brain dynamics in CP children.
► A slow wave brain and higher EEG complexity in CP patients, compared with normal children.
► A positive correlation between beta power and motor dysfunctions in the boys with CP.
► A high accuracy of enhanced probabilistic neural network (94.8%) in diagnosis of CP.