Locomotion in natural environment is a complex task, particularly without visual cues. A universal sense that can be used for solving this problem is touch. In insects, the antennae are the main sensory organ of touch. They are actively movable and allow animals to maneuver in their surroundings. Stick insects of the species Carausius morosus, have poor vison, are wingless, live in cluttered environment and show remarkable climbing abilities. They mostly rely on information gathered by their antennae to find footholds or to cross gaps. They also show a tactually induced reach-to-grasp behaviour with targeted movement of a front leg towards the position of antennal contact. The underlying mechanism is fast and relies on the information transfer from the antennae to thoracic locomotor networks. Descending interneurons that potentially contribute to this behaviour have been identified and provide a nice substrate for the study of control of locomotion in response to tactile cues. However, the properties of these neurons, their role in the control of this behaviour is not well understood. In an attempt to fill this gap, I studied the spiking activity of the pair of contralateral ON-type velocity sensitive descending interneurons (cONv) of the stick insect Carausius morosus.

In chapter one I introduce general aspects of adaptative locomotion and the role of descending neuronal information. I expose the complexity of the control of locomotion under history- and context- dependent neuronal activity. I argue that stick insects are particularly suitable model to study these aspects because they show a reach-to-grasp movement of the front legs in response to antennal contact.

In chapter two, I use bilateral extracellular neck-connective recordings and analyze the modulation of spontaneous activity of cONv in resting animals, during antennal movement or substrate vibration. I found that the spontaneous activity of cONv is asymmetrically reduced during antennal motion whereas substrate vibration reduced the spontaneous activity of both cONv in a similar fashion. I show that cross-modal interactions occur in this system and relate to the effect of substrate vibration on the reduction of spontaneous to the footfall frequency during walking. I argue that walking-induced vibration reduces the spontaneous activity of cONv, thus reducing its impact on velocity encoding in this neuron.

In the third chapter I focus on stimulus-induced responses of cONv. Since stick insects continuously move their antennae during walking and most sensory system are subject to adaptation, I investigate the effect of adaptation in cONv. I took advantage of the bilateral extracellular neck-connective recordings and the bimodality of cONv to study stimulus-specific and cross-modal adaptation. I found that adaptation is cross-modal because substrate vibration pre-adapted cONv. Adaptation in cONv could improve detection of changes in antennal velocity and because antennal contact with an obstacle will produce a change in velocity, I propose that adaptation, could aid the detection of antennal contacts.

In chapter four, I combine motion capture of antennal movements with single-unit recordings of cONv in self-generated antennal movements. I reconstruct the 3D movements of the antennae during self-generated exploratory movement and during imposed deflection of the flagellum. Other than during rest, the joint-angle velocity of self-generated antennal movements were not encoded by cONv. Consistent with a context dependency of neuronal activity, antennal contacts occurring during exploratory movements were signalled by cONv. Because the putative antennal input to cONv are proprioceptive hair fields, I propose that an efference copy mechanism could take place in this system. Finally, I argue that under behaviourally relevant conditions, cONv could serve as a contact detector because it signals unintended movement.

In chapter 5, I conclude this thesis and show the limit of the extracellular recording technique to study cONv depending on the behavioural context. I further develop on the idea of an efference copy mechanism taking place in the stick insect antennal system and discuss the potential antennal inputs to cONv. Finally, I use the knowledge gathered in the previous chapters to propose potential output of cONv and its role in stick insect motor behaviour.