Effect of required response force on rats’ performance on a VI+ schedule of reinforcement

Abstract

McDowell and Wixted (1986) introduced a schedule that has the molar properties of a variable ratio (VR) schedule (i.e., a strong function relating response rate to reinforcement rate) but the molecular properties of a variable interval (VI) schedule (i.e., one that differentially reinforces long interresponse times [IRTs]). This schedule can be called the variable-interval-plus-linear-feedback (VI ) schedule. This contingency reinforces high rates of responding by making rates of reinforcement depend on rates of responding. The VI schedule achieves this b by making the interval required for reinforcement vary inversely with response rate. However, the VI schedule also maintains the interval characteristics of the VI schedule that differentially reinforces long IRTs. Figure 1 shows a schematic representation of the feedb back function between response rate and reinforcement rate for three schedules (VR, VI, and VI ). The feedback function of the VI schedule is given by the following equation: a (i/n) * b. Here, a is the interval required before a response will produce reinforcement, i is the time since the last reinforcement, n the number of responses emitted during that period, and b the VR ratio to which the VI schedule is equivalent. Some numeric examples may serve to illustrate how this schedule works. Assume that the VI schedule was to have the molar characteristics of a VR-30 schedule. Assume also that the time from the last reinforcer was 60 d sec and that 60 responses had been made during that time. In this case, the interval to reinforcement would equal (60/60) * 30 30 sec. During this 30 sec, at 60 responses per minute, 30 responses would be emitted before the interval would time out. If 30 responses had been made during that 60-sec period, the interval would be (60/30) * 30 60 sec. During this 60 sec, if responses were being emitted at 30 responses per minute, 30 responses would be emitted prior to the reinforcer. Alternatively, if 120 responses had been made during the 60-sec period, the interval would become (60/120) * 30 15 sec. During this 15 sec, at 120 responses per minute, 30 responses would be emitted prior to the reinforcer. Thus, the interval varies inversely with the rate of responding, and each reinforcer is delivered for about 30 responses. McDowell and Wixted (1986) found that with human subjects, high response rates were produced by VR schedules and also by VI schedules that had the same feedback functions as the VR schedules. Reed (in press) replicated and extended the findings reported by McDowell and Wixted with human subjects. Reed (in press) noted that their human subjects responded at a similar rate on VR and VI schedules and more quickly on these schedules than on VI schedules yoked to the former contingencies in terms of rate of reinforcement. Importantly, the latter study indicated that subjects were more sensitive to the molar aspects of the VI schedule when they were d given a smaller response force requirement than that used by McDowell and Wixted.In contrast to these results...

A re-examination of responding on ratio and regulated-probability interval schedules

“…Although these results suggest that the RR-20 schedule is not able to sustain higher lever-pressing rates than the RPI-y schedule due to its low R-O rate correlation, there is a wealth of evidence implicating this factor in instrumental conditioning. Reed (2006), for example, compared response rates on the RI-plus-linear feedback schedule (RI+) (McDowell & Wixted, 1986;Reed, 2007aReed, , 2007bSoto, McDowell, & Dallery, 2006) and compared this performance to that of a ratio schedule yoking reward rates. Under the RI+ specification, higher reward probabilities for long IRTs are established in conjunction with positive R-O rate correlations.…”

“…Under the RI+ specification, higher reward probabilities for long IRTs are established in conjunction with positive R-O rate correlations. Using lever pressing in rats, Reed (2006) presented evidence of an RI+ schedule sustaining similar response rates to an RR schedule for matched reward rates.…”

Concurrent VR VI schedules: Primacy of molar control of preference and molecular control of response rates

“…Recently, Reed (2006Reed ( , 2007aReed ( , 2007b showed, in both rats and humans, that response rates on the VI schedule are higher than those on the VI when the force required to operate the operandum was low (see also McDowell & Wixted, 1986). As regards rats and the VR-VI rate difference lit-…”

The copyist model of response emission

“…One limitation is its inadequacy in explaining a subset of single-schedule effects in simulations 1 and 2, where some evidence of molar control of responding can be discerned (Dawson & Dickinson, 1990;Reed, 2007aReed, , 2007b, Experiment 2, lowforce data; see Tables 2 and 3). Because of these and other data sets where molar factors seem operative (Heyman & Tanz, 1995;Reed, 2006;Sakagami, Hursh, Christensen, & Silberberg, 1989;Shurtleff & Silberberg, 1990;Silberberg, Thomas, & Berendzen, 1991;Soto, McDowell, & Dallery, 2006), it appears that operant performances are not always controlled only at a molecular level. That having been said, the operation of molar factors in responding, at least to a degree, need not violate the copyist model, because it blends molar and molecular accounts by considering predecessor IRTs along with reinforced IRT in its definition.…”