One possible approach to treating localized brain tissue loss is substituting the affected brain area with a synthetic neuroprosthetis interfaced to the affected area's inputs and outputs, thus creating a closed-loop system. In the EU-based ReNaChip project, we successfully replaced a dysfunctional cerebellar microcircuit in rats with a biomimetic synthetic system (Prueckle et al, 2011; Bamford et al, 2012). A pivotal part of designing a biomimetic closed-loop system is characterizing the inputs to the neuroprosthesis and studying their relevance to various aspects of its function. In classical conditioning, the unconditioned stimulus (US) is relayed to the cerebellum by the inferior olive (IO). The IO-US signal is often referred to as an "error signal" (representing a sensorimotor error) or as a "teaching signal" (i.e., driving cerebellar plasticity involved in the acquisition of conditioned responses). Moreover, cerebellar output can modulate IO activity - including firing rate and synchronicity, and IO activity can directly modulate cerebellar output.

In the current study, we examined the temporal dynamics of the IO's response to sustained somatosensory stimuli. We demonstrated that the IO's population response to sustained periorbital airpuff-USs is independent of stimulus duration, features a sharp peak 20-30 ms after stimulus onset, followed by a period of reduced activity (post-peak depression) and a subsequent return to baseline levels. We used a computer simulation to explore the involvement of intrinsic IO neuron mechanisms and IO-cerebellar network properties in IO response shaping. Results indicated that the IO signal activates a negative IO-cerebellar feedback loop. The strong peak in IO activity elicited strong cerebello-olivary inhibition, suppressing IO firing and causing the post-peak depression. Following the post-peak depression, IO activity elicited weak cerebello-olivary inhibition, which desynchronizes the firing of IO neurons by reducing their electrotonic coupling. These findings suggested that disruption of the cerebello-olivary pathway would result in sustained IO responses to sustained somatosensory stimuli. An in vivo experiment in which the cerebellar output was temporarily inactivated pharmacologically confirmed this prediction.

We conclude that the IO's response to sustained stimuli is shaped by IO-cerebellar interactions, with potential implications for the acquisition and execution of conditioned motor responses. From the perspective of a closed-loop system, these olivocerebellar interactions must be taken into account in designing brain-neuroprosthetis interactions.