Spontaneous Movements during Sleep Guide Spinal Self-organization: Formation and Expression of a Memory Trace

During development, information about the three-dimensional shape and mechanical properties of the body is laid down in the synaptic connectivity of sensorimotor systems through adaptive mechanisms. In spinal reflex systems, this enables the fast transformation of complex sensory information into adequate correction of movements. In this thesis, it is shown, in a computer simulation, that an unsupervised correlation-based learning mechanism, guided by spontaneous muscle twitches, can account for the functional adaptation of the nociceptive withdrawal reflex system (NWR). This new learning principle was termed Motor-Directed Somatosensory Imprinting (MDSI). MDSI proved to be a highly efficient mechanism, far better than conventional Hebbian feedforward learning, and to be relatively independent on such parameters as learning rate and noise level. By developing and using a new fast optical analysis system for detection and classification of spontaneous movements, it was demonstrated that tactile feedback resulting from spontaneous muscle twitches indeed modifies sensorimotor transformation in young rats in a predictable manner. This learning occurs during “active sleep” which is similar to REM sleep in adults, indicating a novel role for sleep in learning and memory. A new analysis system for rapid imaging of receptive fields, termed RFI, was developed to characterize the differences between strong and weak connections in the NWR in adult rats. Connections of different strengths differed with regard to gain, onset latency and relative variability. These differences represent the preserved end product of MDSI. Neither the inhibitory input to NWR nor differences in glutamate receptors could explain the differences in strengths between strong and weak connections, although it became clear that NMDA receptors are important in setting the overall gain in nociceptive transmission. In conclusion, this thesis demonstrates for the first time, that spontaneous twitches during sleep, corresponding to human foetal movements, play a key role in spinal self-organization and tentatively suggests that this learning results in structural changes, such as elimination or pruning of erroneous connections, in the spinal circuits. Since a variety of spontaneous movements are present concomitant with the maturation of motor systems, it is conceivable that spontaneous movements reflect a general mechanism whereby motor systems are functionally adapted during development