Introduction

Accurate timing is a ubiquitous aspect of mental processes. How does the central nervous system solve the demands involved in the temporal aspects of information processing? One solution would be that timing is handled by subsystems specialized for domain-specific processing. For example, the timing required for producing well-articulated speech would be solved by areas involved in speech production, whereas the timing demands for the coordination of manual actions would be controlled by brain areas that also control the force and spatial aspects of these movements. Alternatively, humans are capable of producing rather arbitrary behaviors that exhibit accurate timing. We can produce periodic movements over a considerable range of durations. These actions can be achieved with different parts of the body and, indeed, do not even require overt actions; we can covertly maintain an internal beat. We can also detect and judge rhythmicity in a wide variety of sensory signals. While our sense of rhythm may be most accurate for auditory events, we can readily detect temporal perturbations in a sequence of visual or tactile events. Thus, there likely exists some general system specialized to represent temporal information, a system that is recruited for tasks that require this form of computation.

There is ample evidence that temporal acuity correlates across different domains and behaviors (Ivry and Hazeltine, 1995). Furthermore, a direct relationship is observed between interval duration and temporal accuracy over a wide variety of intervals, organisms, and tasks (for a review, see Gibbon et al., 1997). These observations suggest at least one common underlying system for the representation of temporal information.

Three major challenges for research on temporal processing then become apparent. The first is primarily psychological, involving the distinction and characterization of different timing systems. What behaviors and perceptual skills share a common system and which functional domains engage domain-specific processes? For example, it has been proposed that a distinction can be made between repetitive movements for which timing is explicitly represented and repetitive movements in which temporal regularities are an emergent property (Ivry et al., in press; Robertson et al., 1999). A second distinction that has been considered is based on the idea that different systems may be engaged, depending on the temporal extent of the timed intervals (Gibbon et al., 1997; Ivry, 1996).

The second challenge is to generate a process model or models that make explicit the component operations involved in tasks that require temporal processing. For example, the scalar timing model specifies a series of component parts associated with the accumulation of clock pulses and the comparison of this sum to stored representations in long-term memory (Gibbon et al., 1997; Treisman, 1963). Similarly, the Wing-Kristofferson model postulates distinct processes that contribute to the variability observed during repetitive tapping tasks (Wing and Kristofferson, 1973). In this chapter, we consider models of this type, analyzing the psychological processes involved in synchronizing an internal timing mechanism to external, rhythmic events.

The last and ultimate challenge is to provide a mapping between psychological operations and neural circuits. This mapping need not be in a one-to-one correspondence.

While some operations may be localized to particular neural structures, it is possible that the operations we describe at a computational level of explanation are implemented in a distributed manner within the brain. Exploring this mapping not only provides a first step toward developing a mechanistic explanation at the neural level, but also can help shape our understanding of the psychological operations.

We will focus on the contribution of two subcortical structures that have been proposed to be the cornerstone of an internal timing system: the cerebellum and the basal ganglia (Ivry, 1997; Meck, 1996). Both structures form reciprocal loops with many cortical areas (Alexander et al., 1986; Middleton and Strick, 1997; Strick et al., 1995), which would enable them to provide the precise representation of temporal information across a range of different task domains. We first review neuropsycho-logical studies that investigate the role of the cerebellum and basal ganglia in the production and perception of timed events. We then report a new experiment, examining the contribution of these structures to synchronization behavior.

0 0

Post a comment