This thesis provides morphological and functional information needed for further development of the implantable neural prostheses for locomotion. Data obtained from series of morphometric studies and retrograde tracing experiments was analyzed using novel statistical and imaging approaches developed to study the organization of the myelinated nerve fibers in the peripheral nervous system.

Research findings can be summarized in two propositions concerning the organization of the peripheral nerves and spinal roots in the lumbar region. By means of statistical modeling of the size distributions enabling to discriminate different populations of fibers in an otherwise "continuous" size spectrum of fibers, it was found that fiber populations have a conserved and readily discernable composition throughout their course along the peripheral nerves. Moreover, it was possible to discern the functional modality of these populations - i.e. their sensory or motor character. With regard to neuroprostheses, it is shown that the peroneal nerve is a suitable target for electric stimulation due to the separation of its large motor fibers from the smaller ones. This composition can explain the empirical efficiency of the implantable drop-foot stimulator. In contrast, solely from a structural point of view, the tibial and sciatic nerves are difficult targets for stimulation due to the worse separation of the large motor fibers from the smaller presumably sensory fibers.

Using spatial statistical technique developed to delineate regions of clustered fibers, it is demonstrated that the A-alpha motor fiber populations, which innervate gastrocnemius muscles are topographically organized (i.e. forming a clustered pattern in the spatial statistical sense) from the level of the ventral roots and remain in this state as they form dedicated motor muscle branches. A hypothesis describing the origin of the observed clustering of the motor axons is proposed. In view of the findings described in this thesis, the ventral spinal roots L4 - L6 are proper targets for electrical stimulation due to their musculotopical organization.

Chapter 1: General Inroduction to the thesis

Chapter 2: Functional Electric Stimulation for sensory and motor functions: progress and problems

This chapter gives an overview on the neuroprosthetic field (e.g. Functional Electrical Stimulation). A brief historical overview is presented in the light of its relevance for the development of the major clinical applications of the implantable neuroprostheses, such as pacing of the heart and diaphragm, the restoration of locomotion and grasping, and the restoration of bladder and bowel functions. Critical analysis is performed on the neuroprosthetic restoration of lost locomotor functions. The structural and functional information needed for further development of the neuroprostheses for locomotion and for urological functions is identified based on the analysis of their contemporary clinical applications.

Published as D. Prodanov, E. Marani, and J. Holsheimer, Biomedical Reviews (2003) 14, 23-50

Chapter 3: Three-dimensional topography of the motor endplates of the rat gastrocnemius muscle

This chapter presents a three-dimensional (3D) map of the motor endplates in the medial and the lateral gastrocnemius muscles of the rat assembled from thick histological sections. Recommendations are given for optimal placement of tracer injections in these muscles based on the presented 3D maps. Gastrocnemius muscles are used further in the retrograde tracing experiments described in Chapters 5 and 6 to label the sciatic nerve and its founding ventral roots. In addition, the use of the presented 3D maps for the realistic simulation of surface EMG recordings is discussed.

Published as D. Prodanov, M.-A. Thil, J. Delbeke, E. Marani, and J. Holsheimer, Muscle & Nerve (2005), 32(3), 292-302

Chapter 4: Morphometric analysis of the fiber populations of the rat sciatic nerve, its spinal roots, and its major branches

This chapter provides reference values for the fiber density, interspace, and cross-sectional area. The sciatic nerve, its founding dorsal and ventral spinal roots, and its major branches, the tibial, peroneal, and sural nerves, were measured (semi-) automatically on semi-thin histological sections. Fiber diameter distributions are modeled statistically using multi-component lognormal models and an optimal model for each nerve or root is selected using information theory criteria. The functional identities of so modeled fiber populations are established using calculations of conduction velocities and anatomical considerations for each studied nerve or root. Findings are discussed in view of their importance for development of neural prostheses.

Published as D. Prodanov and HKP Feirabend. J Comp. Neurol 503, 85-100, 2007

Chapter 5: Automatic morphometry of synaptic boutons of cultured cells using granulometric analysis of digital images

This chapter presents a novel technique for automatic identification and morphometry of fluorescently labeled neuronal structures based on non-linear filtering of digital images. The algorithm is tested in a task for automatic identification of synaptic boutons in microscopic images of cultured cells and further compared to the performance of human observers. Its further use for morphometric analysis of nerve fibers is discussed. A brief overview of the mathematical morphology, the underlying theory used for the algorithm development, is given in an appendix.

Published as D Prodanov, JH Heeroma, & E Marani. J Neurosci Methods, 151(2), 168-177, 2006

Chapter 6: Automated morphometric analysis of the nerve fiber population innervating the rat gastrocnemius muscles

In this part, the method presented in Chapter 5 is elaborated further to encompass automatic identification of tracer in axonal profiles in histological cross sections. Using paired sets of measurements of the tracer signal within the axonal profile and of the corresponding fiber diameter (i.e. axonal profile and its myelin sheath) in sections of ventral roots and peripheral nerves, a relationship between the fiber diameter and the signal diameter is established. Using this relationship the total fiber population is reconstructed from the axonal fluorescence measurements and the functional type of the predominantly labeled fibers is identified.

 

Chapter 7: Local spatial analysis of the motor axonal clustering in rat ventral spinal roots

This chapter presents the functional topography of the motor axons in the L6 ventral spinal root revealed from the performed tracing experiments using rat gastrocnemius muscles. Maps of the locations of the tracer-positive fibers in the ventral root L6 were automatically constructed using the method developed in Chapter 6. A new spatial statistical function is introduced to test for the occurrence of clusters locally and its behavior is tested on simulated data by means of serial Monte-Carlo simulations. The function is further applied to the spatial data from the retrograde tracing experiments and used to reveal the locations of the clusters of labeled fibers. A brief overview on the theory and methodology of spatial statistics is given in an appendix.

Published as Prodanov D, N Nagelkerke, & E Marani, J Neurosci Methods, 160(1), 93-108, 2007

general discussion (Chapter 8).