A complete understanding of sensory and motor processing requires characterization of how the nervous system processes time in the range of tens to hundreds of milliseconds (ms). Temporal processing on this scale is required for simple sensory problems, such as interval, duration, and motion discrimination, as well as complex forms of sensory processing, such as speech recognition. Timing is also required for a wide range of motor tasks from eyelid conditioning to playing the piano. Here we review the behavioral, electrophysiological, and theoretical literature on the neural basis of temporal processing. These data suggest that temporal processing is likely to be distributed among different structures, rather than relying on a centralized timing area, as has been suggested in internal clock models. We also discuss whether temporal processing relies on specialized neural mechanisms, which perform temporal computations independent of spatial ones. We suggest that, given the intricate link between temporal and spatial information in most sensory and motor tasks, timing and spatial processing are intrinsic properties of neural function, and specialized timing mechanisms such as delay lines, oscillators, or a spectrum of different time constants are not required. Rather temporal processing may rely on state-dependent changes in network dynamics.

The phenomenon of long-term potentiation (LTP), a long lasting increase in the strength of synaptic transmission which is due to brief, repetitive activation of excitatory afferent fibres, is one of the most striking examples of synaptic plasticity in the mammalian brain. In the CA1 region of the hippocampus, the induction of LTP requires activation of NMDA (N-methyl-D-aspartate) receptors by synaptically released glutamate with concomitant postsynaptic membrane depolarization. This relieves the voltage-dependent magnesium block of the NMDA-receptor ion channel, allowing calcium to flow into the dendritic spine. Although calcium has been shown to be a necessary trigger for LTP (refs 11, 12), little is known about the immediate biochemical processes that are activated by calcium and are responsible for LTP. The most attractive candidates have been calcium/calmodulin-dependent protein kinase II (CaM-KII) (refs 13-16), protein kinase C (refs 17-19), and the calcium-dependent protease, calpain. Extracellular application of protein kinase inhibitors to the hippocampal slice preparation blocks the induction of LTP (refs 21-23) but it is unclear whether this is due to a pre- and/or postsynaptic action. We have found that intracellular injection into CA1 pyramidal cells of the protein kinase inhibitor H-7, or of the calmodulin antagonist calmidazolium, blocks LTP. Furthermore, LTP is blocked by the injection of synthetic peptides that are potent calmodulin antagonists and inhibit CaM-KII auto- and substrate phosphorylation. These findings demonstrate that in the postsynaptic cell both activation of calmodulin and kinase activity are required for the generation of LTP, and focus further attention on the potential role of CaM-KII in LTP.

Comparison of two seemingly quite different behaviors yields a surprisingly consistent picture of the role of the cerebellum in motor learning. Behavioral and physiological data about classical conditioning of the eyelid response and motor learning in the vestibulo-ocular reflex suggests that (i) plasticity is distributed between the cerebellar cortex and the deep cerebellar nuclei; (ii) the cerebellar cortex plays a special role in learning the timing of movement; and (iii) the cerebellar cortex guides learning in the deep nuclei, which may allow learning to be transferred from the cortex to the deep nuclei. Because many of the similarities in the data from the two systems typify general features of cerebellar organization, the cerebellar mechanisms of learning in these two systems may represent principles that apply to many motor systems.

Among the many issues surrounding the involvement of the cerebellum in motor learning, the relative roles of the cerebellar cortex and cerebellar nuclei in Pavlovian conditioning have been particularly difficult to assess. While previous studies have investigated the effects of cerebellar cortex lesions on the acquisition and retention of conditioned movements, we have examined the effects of these lesions on the timing of Pavlovian eyelid responses. The rationale for this approach arises from previous studies indicating that this timing is a component of Pavlovian eyelid responses that is learned and that involves temporal discrimination. To permit within-animal comparisons, rabbits were trained to produce differently timed responses to high- and low-frequency auditory conditioned stimuli (CSs). Before the lesion the conditioned responses to both CSs were appropriately timed--each peaked near the time at which the unconditioned stimulus was presented for that CS. However, after the lesion both CSs could elicit similarly timed conditioned responses that peaked inappropriately at very short latencies. The changes in responses timing were sensitive to the size of the lesion, particularly its rostral-caudal extent. Similar results were obtained in animals trained with one CS, indicating that the disruption of response timing is not related to impaired auditory discrimination. Because response timing is learned and therefore requires synaptic plasticity, these data suggest that there are at least two sites of plasticity involved in the motor expression of Pavlovian eyelid responses. Plasticity at one site is necessary for the learned timing of conditioned responses, while plasticity at another site is revealed by the inappropriately timed responses observed following removal of the cerebellar cortex. This lesion-induced dissociation of the expression of motor responses and their learned timing supports a synthesis of competing views by suggesting that motor learning involves both the cerebellar cortex and cerebellar nuclei. We hypothesize that motor learning involves a decrease in strength of the granule cell-Purkinje cell synapses (e.g., Ito and Kano, 1982) in the cerebellar cortex and an increase in strength of the mossy fiber-cerebellar nuclei synapses (e.g., Racine et al., 1986). Finally, these data suggest that the cerebellar cortex may mediate the temporal discriminations that are necessary for the learned timing of conditioned responses.

We used large-scale computer simulations of eyelid conditioning to investigate how the cerebellum generates and makes use of temporal information. In the simulations the adaptive timing displayed by conditioned responses is mediated by two factors: (1) different sets of granule cells are active at different times during the conditioned stimulus (CS), and (2) responding is not only amplified at reinforced times but also suppressed at unreinforced times during the CS. These factors predict an unusual pattern of responding after partial removal of the cerebellar cortex that was confirmed using small, electrolytic lesions of cerebellar cortex. These results are consistent with timing mechanisms in the cerebellum that are similar to Pavlov's "inhibition of delay" hypothesis.