Identification of the optimal parameters for electrical stimulation to generate locomotor patterns in the rat isolated spinal cord

Recently, an innovative protocol of electrical stimulation, named “fictive locomotion induced stimulation” (FListim), which consists of an intrinsically variable noisy waveform, has been obtained from a segment of chemically-induced fictive locomotion (FL) sampled from the ventral root (VR) of an in vitro preparation of neonatal rat spinal cord. FListim delivered at sub-threshold intensities to a dorsal root (DR) has been shown to optimally activate the central pattern generators (CPGs) for locomotion (Taccola, 2011). In an attempt to introduce novel and improved protocols of stimulation in combination with neurochemicals, the current PhD project aims to identify the features that make sub-threshold noisy waveforms effective in activating locomotor patterns. In an attempt to introduce novel and improved protocols of stimulation in combination with neurochemicals, the current PhD project aims to identify the features that make sub-threshold noisy waveforms effective in activating locomotor patterns. To reach this aim, locomotor-like patterns in response to different noisy waveforms were compared. In order to obtain a wide palette of noisy protocols electromyographic (EMG) recordings were performed from leg muscles of adult volunteers during walking. These recordings were then delivered as stimulating patterns called real locomotion-induced stimulation (ReaListim). To reach this aim, locomotor-like patterns in response to different noisy waveforms were compared. In order to obtain a wide palette of noisy protocols electromyographic (EMG) recordings were performed from leg muscles of adult volunteers during walking. These recordings were then delivered as stimulating patterns called real locomotion-induced stimulation (ReaListim). ReaListim protocols, sampled during different motor behaviours, are equally able to induce an epoch of locomotor-like oscillations. Conversely, smooth kinematic profiles and non-phasic noisy patterns such as standing and isometric contraction, are unable to activate the locomotor CPGs. The complexity of noisy waveforms was then reduced at motoneuronal level, by recording electrical activity of a single motoneuron during FL. Long-lasting episode of FL, were evoked in response to intracellular patterns delivered at sub-threshold intensities. The analysis of motoneuronal firing during FL was used to identify four recurrent frequency values that optimally activated the locomotor CPGs when applied simultaneously in a multifrequency protocol. Different permutations were tried to further simplify the multifrequency protocol while isolating the most effective components of the four identified frequencies. The simplest asynchronous paradigm that can induce locomotor-like episodes consists of a train of rectangular pulses that contain two frequencies: 35 and 172 Hz. This protocol resulted already effective at subthreshold intensity even when delivered for a very short time (500 ms). The role of oxytocin in the modulation of neuronal networks is explored here on spinal networks. Intracellular recordings demonstrate that oxytocin dosedependently depolarizes single motoneurons with the appearance of sporadic bursts with superimposed firing. By applying the selective blocker of sodium channels, tetrodotoxin (TTX), the effects of oxytocin can be completely abolished, which suggest a premotoneuronal-level origin. The neuropeptide is capable to induce VRs depolarization with superimposed synchronous bursts of activity, while reflex responses induced by single pulses are depressed depending on the stimulus strength and peptide-concentration. The disinhibited bursting evoked by the pharmacological blockade of glycine and GABAA receptors blockers, strychnine and bicuculline, respectively, is accelerated by oxytocin, an effect that is suppressed by the selective oxytocin receptor antagonist atosiban. On spinal locomotor networks oxytocin facilitates the emergence of FL episodes in response either to weak noisy waveforms protocols or to the conjoint application of NMDA and 5HT at sub-threshold concentrations, even if the periodicity of a stable FL is not significantly affected by the neuropeptide. Interestingly, the facilitation of the locomotor CPGs by oxytocin is dependent on the endogenous release of 5HT, as is demonstrated by incubation with the inhibitor of 5HT synthesis, pchlorophenilalanine (PCPA). Low-frequency trains of stereotyped pulses (0.33 and 0.67Hz) delivered with a controlled time interval (delays 0.5 to 2 s) to multiple DRs converged on spinal locomotor circuits to generate locomotor rhythm. The same finding is confirmed by the phase resetting that is induced by single afferent stimuli during a simultaneous train of pulses delivered to another DR. Staggered protocols fail to elicit FL when simultaneously applied to multiple DRs, while a multi-site randomized pulse train is still effective in eliciting locomotor-like patterns. This thesis outlines new strategies for optimizing the reactivation of spinal locomotor networks after spinal damage. Though the technology that is currently available in clinics does not allow for the delivery of highly-variable stimulating patterns, experiments reported here indicate a way to overcome these limitations. Indeed, protocols that contain few distinct frequencies that are isolated from the spectrum of noisy waves can activate the CPGs even when delivered with a multisite approach. This suggests that it may be possible to separately supply multiple trains of pulses to several cord sites using different electrostimulators. The yield of stimulation in activating locomotor circuits will be further improved by the association with the neuropeptide oxytocin.