Prior research proposed that temporal control over the pattern of operant wheel running on a fixed interval (FI) schedule of sucrose reinforcement is a function of automatic reinforcement generated by wheel running and the experimentally arranged sucrose reinforcement. Two experiments were conducted to assess this prediction. In the first experiment, rats ran for different durations (0, 30, 60, and 180 min) prior to a session of operant wheel running on a FI 120-s schedule. In the second experiment, the concentration of sucrose reinforcement on a FI 180-s schedule was varied across values of 0, 5, 15, and 25%. In Experiment 1, as the duration of pre-operant running increased, the postreinforcement pause before initiation of running lengthened while wheel revolutions in the latter part of the FI interval increased. In Experiment 2, wheel revolutions markedly increased then decreased to a plateau early in the FI interval. Neither manipulation increased temporal control of the pattern of wheel running. Instead, results indicate that operant wheel running is regulated by automatic reinforcement generated by wheel activity and an adjunctive pattern of running induced by the temporal presentation of sucrose. Furthermore, the findings question whether the sucrose contingency regulates wheel running as a reinforcing consequence.

Three experiments assessed whether changes in rates of wheel running and lever pressing on a schedule of wheel-running reinforcement were controlled by local or overall constraint of the contingent response. In Experiment 1 rats responded on a response-initiated variable interval 15-s schedule with wheel running as reinforcement. Overall constraint (total revolutions by session) was held constant at 600 revolutions while local constraint (revolutions per reinforcement) varied from 5 to 60 revolutions. Results showed that rates of wheel running and lever pressing decreased as local constraint decreased. In Experiment 2, rats responded on the same schedule, but local constraint (revolutions per reinforcement) was held constant at 10 revolutions. Overall constraint (total revolutions) was varied from 100 to 800 revolutions. Results showed no systematic changes in rates of wheel running and lever pressing with variation in overall constraint. In Experiment 3, rats responded on the same VI schedule for 30 revolutions as reinforcement. A 4.5-s interruption during the wheel-running reinforcement (30 revolutions) occurred after 2, 5, 10, and 15 revolutions as a manipulation of local constraint. Results showed that with overall constraint (total revolutions) and total reinforcement held constant, rates of wheel running before the interruption and overall lever-pressing rates varied in a bitonic relationship with local constraint. Comparison of the results across experiments demonstrates that local constraint of contingent wheel running controls the rate of wheel running and the rate of instrumental responding for wheel-running reinforcement.

Does the effect of amphetamine on behavior (wheel running) differ depending on the functional role (operant, reinforcement) of that behavior? This study addressed this question using a multiple schedule of reinforcement in which wheel running served as reinforcement for lever pressing in one component and as operant behavior for sucrose reinforcement in the other component. Seven female Long-Evans rats were exposed to a multiple schedule in which pressing a lever on a variable ratio 10 schedule produced the opportunity to run for 15 revolutions in one component and running 15 revolutions produced a drop of 15% sucrose solution in the other component. Doses of 0.5, 1.0, and 2.0 mg/kg d-amphetamine were administered by intraperitoneal injection 20 min prior to a session. As amphetamine dose increased, wheel running decreased in both components - showing no evidence that the effect of the drug on wheel running depended on the function of wheel activity. Notably, lever pressing for wheel-running reinforcement also decreased with amphetamine dose. Drug dose increased the initiation of operant lever pressing, but not the initiation of operant wheel running. We propose that amphetamine dose had common effects on wheel running regardless of its function (reinforcement vs. operant) because wheel-running generates automatic reinforcement and the automatic-reinforcement value of wheel activity is modulated by drug dose.

The current study compared the development of response patterns for operant wheel-running and lever-pressing on fixed-interval schedules. Eleven female Long-Evans rats were exposed to fixed-interval (FI) 15-s, 30-s, and 60s schedules with wheel revolutions as the operant behavior and sucrose solution as reinforcement. Subsequently, a lever was mounted in each wheel and rats responded on an FI-30 s schedule of sucrose reinforcement. Operant lever-pressing on average developed a scalloping pattern of low responding early in the reinforcement interval followed by an increase in pressing to the moment of reinforcement. In contrast, average operant wheel-revolutions peaked early in the reinforcement interval followed by a plateau, a pattern that did not change over sessions. Variation in the FI-schedule value (interval size) with operant wheel-running did not alter the pattern of running throughout the reinforcement interval, but merely parsed this pattern at different points. Cumulative records for the last session showed long postreinforcement pauses (PRP) for lever pressing. Wheel running, however, rose quickly after reinforcement and continued throughout the reinforcement interval. Overall and local wheel-running rates decreased and PRP duration increased as the interval size of the FI schedule increased. We propose that the automatic reinforcement generated by wheel running, but not lever pressing, provides an account of the poor temporal regulation of operant wheel-running in our study.

The current study investigated the relationship between wheel-running reinforcement and operant lever pressing when an opportunity to run is defined by the number of wheel revolutions rather than duration of wheel access. Twelve female Long-Evans rats responded on response-initiated variable interval 15-s schedules for the opportunity to run for 1, 3, 5, 10, 20, 30, or 40 revolutions. Half the rats received an ascending order of revolutions/reinforcement; the other half, received a descending order. Results showed that wheel-running and lever-pressing rates were described by a bitonic relationship with revolutions/reinforcement, initially rising then falling, as revolutions/reinforcement increased. Long postreinforcement pause (PRP) durations occurred with few revolutions, short PRPs with an intermediate number, and long PRPs again with a high number of revolutions; thus, PRP showed an inverted U-shaped function. Rather than a reinforcement-magnitude interpretation, our findings suggest that number of revolutions/reinforcement (or duration of wheel access) be conceptualized as constraint on an automatically-reinforcing behavior (wheel running). Specifically, we propose that the automatic-reinforcement value of contingent wheel running varies with its rate of occurrence, which causes instrumental lever-pressing rates to vary with wheel-running rates. The current analysis of constraint and the automatic reinforcement of the contingent behavior is shown to further extend the response-deprivation hypothesis to wheel-running reinforcement.

Previous research using resistance to extinction to assess response strength has shown slower attenuation of responding maintained by larger than smaller reinforcement magnitudes. The current study sought to generalize this reinforcement-magnitude effect to extinction of lever-pressing rates maintained by wheel-running reinforcement of different durations. Rats responded on a response-initiated variable interval 20-s schedule for the opportunity to run in a wheel for 5s, 30s, and 90s. Following training on each reinforcement duration, lever pressing for wheel running was placed on extinction. Results showed that responding maintained by 5s of wheel-running reinforcement declined slower in extinction than for the 90-s duration, while 30s of wheel-running reinforcement produced inconsistent effects on attenuation of response rates. These results suggest that shorter periods of wheel-running reinforcement have higher reinforcement value than longer durations—findings incompatible with the earlier research showing more resistance to extinction with large reinforcement magnitudes. We offer an interpretation based on response deprivation arising from restriction of the wheel-access interval and its motivational impact on wheel running as a reinforcing consequence and on responding for wheel-running reinforcement. An alternative analysis proposes that discrimination of the shift to extinction is more salient for the long wheel-running duration—accounting for the rapid attenuation of responding in extinction with the long periods of wheel-running reinforcement.