Acquired pathological gait patterns and their relationship to structural brain lesions in children with spastic cerebral palsy through Bayesian reasoning: a methodological framework.

Cerebral Palsy (CP) is one of the most common neurological pathologies in children and is caused by a broad range of pre-, peri- or postnataly acquired structural brain lesions, such as periventricular leucomalacia or intraventricular hemorrhage. These brain lesions lead to several topographical involvements (hemiplegia, diplegia, ), primary neuromotor deficits (spasticity, muscle weakness, ) and heterogeneous pathological movement patterns during functional activities, such as walking. The latter has a huge impact on the activities of daily life and consequently patients are often dependent on others. Therapy aiming at the improvement of pathological movement patterns therefore has a socio-economic relevance.Previous research mainly focused on unraveling general connections between brain lesions and motor outcome in children with CP. These studies showed significant relations between (classes of) brain lesions and 1. the (in)ability to walk, 2. a general measure of motor involvement, 3. topographical involvement, 4. muscle tone in premature infants and 5. (classes of) muscle strength. Until now, the relation between pathological movement patterns and brain lesions has not been described. Insight in this connection is crucial for the multidisciplinary treatment of CP as it enables early prediction of movement patterns and individual tuning of the multidisciplinary therapy. The search for relevant connections is hindered by the broad range of brain lesions and the heterogeneity of the gait patterns in CP. Moreover, CP gait is strongly influenced by the primary deficits that interact in a highly complex way. Therefore, it is important to reduce the high-dimensionality of the CP motor data in order to facilitate the search for connections with brain lesions. The overall aim of this PhD was to elucidate the complex relation between the wide range of brain anomalies, primary motor deficits and heterogeneous gait pathologies observed in children with CP (Figure 5) thereby providing a framework for the future construction of a prognostic (Bayesian) model. Such a prognostic model would enable early identification of the main problems in the long term lower limb motor development of children with CP based on brain anomalies, and thus a timely individual goal-setting and tuning of the multidisciplinary therapy. To develop the framework for the future construction of a prognostic (Bayesian) model, this PhD project focused on the development of methodological concepts that allowed (1) modeling of CP gait pathologies; (2) quantifying the role of primary motor deficits in these gait pathologies; and (3) comprehensively evaluating the brain anomalies to further quantify the relation between these anomalies and gait pathology in CP. The crucial aspects and innovations in the field of gait pathology, primary motor deficits and brain anomalies are addressed through several studies, which are outlined below. The first research phase developed a generic workflow for the application of Bayesian reasoning in the preprocessing of three-dimensional gait analysis (3DGA) data. The ultimate goal is the classification of 3DGA data into multiple gait patterns. However, modeling multiple patterns within one patient population (CP) is very complex and Bayesian reasoning has not been applied for this before. Hence, this research phase focused on establishing the methodological workflow for gait classification through Bayesian reasoning and defined the feasibility and efficacy of the Bayesian approach based on a first network modeling the movement patterns of the ankle and knee in the sagittal plane only. A crucial keystone is thereby provided for the future development of a full Bayesian model of CP gait. The second research phase developed high-demand tasks to gain further insight in the role of primary motor deficits in gait. Within this PhD project, a high-demand task for lower limb muscle strength was implemented in the standard 3DGA to assess the added value of this task and to identify 3DGA markers for underlying muscle weakness.The final research phase drafted a standardized and systematic workflow for the collection and analysis of MRI data on brain anomalies. The large set of neurological parameters that resulted from this comprehensive evaluation was then related to the gait pathologies in CP.Table of Contents Dankwoord i Frequently used abbreviations v Chapter 1 General introduction and outline 1 Chapter 2 Probabilistic gait classification in children with cerebral palsy: A Bayesian approach 35 Chapter 3 To what extent is mean EMG frequency during gait a reflection of functional muscle strength in children with cerebral palsy 63 Chapter 4 High-demand tasks to highlight the relationship between muscle weakness and walking in children with cerebral palsy 77 Chapter 5 Macro- and microstructural brain lesions relate to gait pathology in children with cerebral palsy 91 Chapter 6 General discussion 113 Summary 138 Dutch Summary (Samenvatting) 141 Appositions (Bijstellingen) 144 Curriculum Vitae 145 Publications 146nrpages: 148status: publishe