Non-invasive recording from the human cerebellum during a classical conditioning paradigm using the otolith-evoked blink reflex

Abstract

We studied nine normal volunteers with a classical conditioning paradigm using a mastoid tap, believed to activate otolith receptors, as an unconditional stimulus (US) and the consequent blink as the unconditioned response (UR). Both visual (alternation of stripes) and an auditory tone were used as conditional stimuli (CS). Recordings were made below the eyes at IO1 and IO2, from over the frontal eye fields (C3′ and C4′) and over the posterior fossa, the latter at sites we have previously reported that we were able to record an evoked climbing fibre response (CFR) at short latency. Behavioural analysis confirmed that weak conditioning did occur early, which subsequently showed extinction on repeated CS alone trials. Further, a UR was more likely to occur following a preceding CFR when preceded by a CS, supporting a correlation between the CFRs and behaviour. For further statistical analysis, the time period of interest was divided into a series of epochs, based around the events occurring at the time. Grand averages, plus analysis of variance, confirmed evidence of weak conditioning for the blink response following both modalities. The EMG associated with the eyeblink for the UR occurred at a similar time to the expected post-CFR pause in the spontaneous cerebellar activity, or electrocerebellogram (ECeG), while hypothesised conditioned pausing in the ECeG was also observed in CS alone trials. A correlation was found between the size of the CFR and the RMS amplitudes of the segments covering the ocular vestibular evoked myogenic potential oVEMP, (periocular) EMG and the EOG. The slope was greater for the non-oVEMP segments than for the oVEMP segment suggesting the correlation was not simply due to differing sizes of the vestibular volley. We suggest that these recorded events fit with the proposed role of the CFR in Purkinje neurons in classical conditioning, gating the excitability of the cerebellar nuclei, and thereby neurons in the reticular formation mediating the otolith blink reflex. This effect appears to apply to polysynaptic reflexes only as there was no evidence of changes to the oVEMP.