Timing in Eyeblink Classical Conditioning and Timed-Interval Tapping

The cerebellum is implicated in interval timing for diverse tasks including eyeblink classical conditioning (EBCC) and repetitive tapping. We examined performance on both tasks across identical intervals ranging from 325 to 550 ms. In five weekly sessions, 23 participants used a different interval each week, both as the target for tapping and as the delay interval in EBCC. Changes in variability as a function of the tapping or delay interval were assessed using regression analyses. The slope for repetitive tapping was comparable to two measures of temporal acuity in EBCC, onset and peak latency of the conditioned response. Each of 80 additional participants was assessed in one session at one of four tapping and delay intervals. Results were similar to those observed in the repeated measures group. These findings provide further evidence that EBCC and repetitive tapping utilize common mechanisms for representing temporal information. Two tasks that depend on the cerebellum are eyeblink classical conditioning (EBCC) and timed-interval tapping. In EBCC, each trial consists of presentation of a neutral stimulus, the conditioned stimulus (CS), followed by a reflex-eliciting stimulus, the unconditioned stimulus (US). For example, the CS might be a tone and the US a corneal airpuff. Over trials, the organism learns to produce a conditioned response (CR) in anticipation of the corneal airpuff. EBCC has proven to be one of the most fruitful model tasks for studying the neural mechanisms of learning and memory (see Thompson, 1990). Lesions of the cerebellum produce severe impairments in EBCC in both rabbits and humans, although the motor response, the unconditioned response (UR), remains intact (see reviews in Steinmetz, 1996; Woodruff-Pak, 1997). The cerebellum receives inputs conveying representations of both the CS and the US. However, EBCC does not simply require that these two stimuli be associated. The organism must be able to represent the precise temporal relationship between the CS and US so that the CR occurs just prior to the onset of the US. Lesion studies (Perrett, Ruiz, & Mauk, 1993; WoodruffPak, Lavond, Logan, Steinmetz, & Thompson, 1993), as well as computational models (Bartha, Thompson, & Gluck, 1992; Buonomano & Mauk, 1994; Fiala, Grossberg, & Bullock, 1996), indicate that although the cerebellar nuclei are essential for forming the critical associative link, precise timing is dependent on the cerebellar cortex. Precise timing is also required for the production of coordinated movement. Lesions of the cerebellum produce impairments on a range of experimental tasks that directly assess timing control (see Ivry, 1997). One such task is the repetitive tapping task (Wing & Kristofferson, 1973), in which participants attempt to produce a series of equally spaced intervals, first with a pacing signal (synchronization phase) and then when the pacing signal is terminated (continuation phase). Patients with cerebellar lesions demonstrate increased temporal variability on this task, and this deficit has been attributed to increased noise in the central control of the timing of the responses (Ivry & Keele, 1989; Ivry, Keele, & Diener, 1988; Woodruff-Pak, Papka, & Ivry, 1996). The cerebellum has been hypothesized to operate as an internal timing system, given its link to EBCC, repetitive tapping, and other tasks requiring the precise representation of temporal information. To examine interactions between these two tasks, we assessed EBCC learning rates under various dual-task conditions, finding that participants who engaged in tapping while they underwent EBCC showed a reduced percentage of CRs compared with participants who engaged in control tasks while they underwent EBCC (Papka, Ivry, & Woodruff-Pak, 1995). Results indicated that simultaneous tapping interfered with EBCC because the two tasks shared a common cerebellar substrate. Timing variability is a constant proportion of the interval being timed, at least for intervals ranging from 200 ms to 1.5 s (see Getty, 1975; Gibbon, 1991). This relationship, a temporal form of Weber’s law, is described by the equation variance = k2 * interval2 + c (1) The slope, k2, provides a measure of duration-dependent variability, assumed to reflect noise in an internal timing system. The square root of the slope corresponds to the Weber fraction. The intercept, c, provides a measure of duration-independent variability, such as noise related to sensory processing or motor implementation. If two tasks share a common timing component, then the slopes should be similar. Ivry and Hazeltine (1995) applied Equation 1 to variability data obtained on tapping and perception tasks with intervals ranging from 325 to 550 ms. Across a series of experiments, the slope terms for the motor and perceptual tasks were correlated and affected by similar manipulations. In this study, we employed this methodology to explore similarities in EBCC and repetitive tapping. A range of intervals was used, either as the delay between the CS and US or as the target interval on the repetitive tapping task. We predicted that the slopes would be similar for the two tasks. We did not expect the intercepts to be comparable because the two tasks entail different perceptual and motor pathways.