In lieu of an abstract, here is a brief excerpt of the content:

27 The Potential Link between Temporal Averaging and Drug-Taking Behavior Allison N. Kurti, Dale N. Swanton, and Matthew S. Matell The capacity to perceive time in the seconds to minutes range, or interval timing, allows organisms to develop temporal expectations about when significant events should occur, therein promoting the efficient organization of behavior. However, disruptions in temporal perception, such as those that have been seen following drug use, for example amphetamine (Eckerman et al., 1987), methamphetamine (Maricq, Roberts, & Church, 1981; Matell, Bateson, & Meck, 2006), cocaine (Matell, King, & Meck, 2004), marijuana (Mathew et al., 1998), MDMA (Frederick & Paule, 1997), and other drugs of abuse (Paule et al., 1999), can have dramatic impacts on the temporal organization of behavior. Furthermore, since it has been demonstrated that temporal expectations play a role in choice behavior (e.g., intertemporal choice based on delay discounting), alterations in temporal perception could lead individuals to pursue alternative goals. In the present chapter, we discuss the formation of a novel temporal expectation resulting from the averaging of temporal memories, and we end with the hypothesis that temporal memory averaging may play a role in drug abuse and addiction. 27.1 General Framework of Interval Timing Prior to delving into the data obtained in our laboratory on temporal memory averaging, we provide a brief framework of an interval-timing model, which we refer to in subsequent presentations of our experimental findings. The large majority of interval-timing models can be broken down into three information-processing components: a clock, memory, and decision stage (Church, 1997). While there are important differences in the form of these stages across models, the basic framework is that the clock component provides an isomorphic representation of elapsed time; temporal memory is a storehouse of experienced clock values at biologically relevant times; and temporally controlled behaviors are produced when the decision stage registers that the current clock representation is “similar enough” to the value(s) stored in temporal memory. For example, in scalar expectancy theory (Gibbon, 1977), a popular model of interval timing, the clock component is instantiated as a pacemaker-accumulator system in which the subjective representation of elapsed time 600 Allison N. Kurti, Dale N. Swanton, and Matthew S. Matell grows as a linear function of objective time. Upon occurrence of a biologically relevant event, such as reinforcement, the accumulated value of the clock is stored as an element in a distribution of temporal memories. Temporally controlled behaviors are emitted when the current accumulator value is similar enough to a value selected from temporal memory according to a proportional rule. Due to variability in these components across trials (i.e., clock speed, memory storage processes, and decision thresholds), temporal estimates are roughly normally distributed. Further, due to the proportional similarity rule of the decision stage, errors in estimation are directly proportional to the interval being timed, a characteristic of interval timing known as the scalar property (Gibbon, 1977). 27.2 Temporal Averaging We have recently reported that simultaneously presenting rats with two stimuli, each specifying its own tone-food reinforcement delay, results in maximal response at a time midway between the conflicting intervals (Swanton, Gooch, & Matell, 2009). Specifically, rats were trained on a dual-duration, peak-interval procedure in which one modal stimulus (e.g., a 4 kHz tone) signaled probabilistic reinforcement on a fixed-interval 10 s schedule (i.e., on a portion of trials, the first nose-poke response after 10 s was reinforced, and the stimulus terminated), and a different modal stimulus (e.g., a house light) signaled probabilistic reward on a fixed-interval 20 s schedule. The stimulus-duration relations were counterbalanced across rats. On a proportion of trials, “probes” were provided in which one of the cues commenced , but no reinforcement was delivered, and the cue terminated independently of response after three to four times the duration associated with that cue. Plotting the rate of nose-poking on probe trials as a function of elapsed time following the onset of each cue resulted in the typical Gaussian-shaped “peak functions” in which maximal responding occurred at approximately the time of the criterion interval associated with each cue. In order to keep peak response rates equivalent between the short (10 s) and long (20 s) cues, the reinforcement probability was twice as large for the long cue (50 percent) as for the short cue (25 percent). After sufficient training to establish reliable peak functions for the discriminative stimuli signaling each of the two anchor durations, rats were presented with a compound...

Share