|
Schedule. Choose the schedule, and specify any parameters associated with the schedule. Parameter descriptions will appear to help you. The schedules are:
| • | CRF - continuous reinforcement (FR-1). One reinforcer per response. |
| • | EXT - extinction. No reinforcers. |
| • | FR x - fixed ratio. One reinforcer per x responses. |
| • | VR x to y - variable ratio (specifying min, max). After a variable number of responses (randomly chosen from min to max inclusive), one reinforcer is delivered. |
| • | RR x - random ratio. P(reinforcer | response) = 1/x. |
| • | PROB p - probabilistic. P(reinforcer | response) = p. |
| • | FI x - fixed interval. The first response after x seconds is reinforced. The first response of the schedule is also reinforced. |
| • | RI x - random interval. Reinforcement is set up on a random-time schedule (see below); after reinforcement has been set up, the next response is reinforced. |
| • | VI x to y - variable interval (specifying min, max). After a variable time (from min to max seconds), the next response is reinforced. |
| • | FT x - fixed time (NONCONTINGENT). No lever is present. Reinforcement is delivered every x seconds. |
| • | VT x to y - variable time (specifying min, max) (NONCONTINGENT). No lever is present. The schedule waits for between min and max seconds, then delivers a reinforcer, then repeats. |
| • | RT x - random time (NONCONTINGENT). Every second, p(reinforcer delivered this second) = 1/x. Thus, on average, reinforcement is delivered once every x seconds, but the subject cannot predict the likelihood of reinforcement based on how long it has waited (unlike a typical VT schedule). |
| • | PR - progressive ratio - add one (1,2,3,4...) - progressive ratio schedule, adding one to the ratio requirement at each step. The schedule termination is determined by the parameter; if parameter is >0, then when parameter minutes have elapsed since the last reinforcer (or response - see below), the schedule stops. We suggest 60 as a sensible value. |
| • | PR - progressive ratio - double (1,2,4,8...) - progressive ratio schedule, doubling the ratio requirement at each step. The schedule termination is determined by the parameter; if parameter is >0, then when parameter minutes have elapsed since the last reinforcer (or response - see below), the schedule stops. We suggest 60 as a sensible value. |
| • | PR - progressive ratio - Fibonacci (1,1,2,3,5...) - progressive ratio schedule with a Fibonacci progression. The schedule termination is determined by the parameter; if parameter is >0, then when parameter minutes have elapsed since the last reinforcer (or response - see below), the schedule stops. We suggest 60 as a sensible value. |
| • | PR - progressive ratio - Roberts exponential (A * exp(reinfnum * B) - A) - progressive ratio schedule with an exponential progression, based on Roberts DCS & Richardson NR (1992), Self-administration of psychomotor stimulants using progressive ratio schedules of reinforcement, Neuromethods 24: 233-269 (eds Boulton A, Baker G, Wu PH; Humana Press). The ratio requirement is (A * exp(reinforcer number * B)) - A, rounded to the nearest integer. Typically, A is 5. A typical schedule might have B=0.2; these values yield ratio requirements {1, 2, 4, 6, 9, 12, 15, 20, 25, 32, 40, 50, 62, 77, 95, 118, 145, 178, 219, 268, 328, 402, 492, 603, 737, 901, 1102, 1347, ...}. A steeper PR schedule is obtained with B=0.25, giving {1, 3, 6, 9, 12, 17, 24, 32, 42, 56, 73, 95, 124, 161, 208, 268, 346, 445, 573, 737, 948, 1218, 1566, 2012, 2585, 3321, 4265, 5478, ...} The schedule termination is determined by the other parameter (on the left, labelled (min)); if this parameter is >0, then when this many minutes have elapsed since the last reinforcer (or response - see below), the schedule stops. We suggest 60 as a sensible value. |
| • | DELAYED_FR1 - FR1 with delayed reinforcement. This is an FR1 schedule, but there is a delay between responding and reinforcement. This delay is the sole parameter (specified in seconds). |
| • | PR - progressive ratio - double increment every A reinforcers. The increment starts at 1, and doubles every A reinforcers. If A is 8, then the ratio requirements are 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16, 18, 20, 22, 24, 28, 32, 36... The schedule termination is determined by the parameter; if parameter is >0, then when parameter minutes have elapsed since the last reinforcer (or response - see below), the schedule stops. We suggest 60 as a sensible value. |
| • | HAMMOND_CONTINGENCY - Hammond (1980) instrumental contingency. After Hammond (1980, J Exp Analysis Behav 34: 297). Time is windowed into 1-second bins. A decision about reinforcement is made at the end of each time bin (not when responding occurs). At the end of each time bin, if there has been at least one response during that time bin, reinforcement is delivered with probability P(reinf | response). If, instead, no responses were made in that time bin, reinforcement is delivered with probability P(reinf | ~response). You specify the two probabilities. The contingency is P(reinf | response) - P(reinf | ~response). Currently restricted to certain laboratories. |
| • | JACKSON_CONTINGENCY - Jackson et al. (2014 unpub). instrumental contingency. A modification of the Hammond instrumental contingency schedule. To a first approximation, Jackson is to Hammond what VR is to RR, or VT is to RT. Time is divided into bins (typically but not necessarily lasting 1 s each). Imagine each time bin generating a tick or "ping" when the bin finishes. The subject's behaviour during that time bin determines which of two schedules the "ping" is sent to. If the subject responded during the bin, the ping is sent to the contingent schedule (at the end of the time bin); if it didn't respond, the ping is sent to the noncontingent schedule. So far, that could describe the Hammond schedule. But then, rather than choose on a wholly probabilistic basis, the contingent part of the Jackson schedule assigns a target number of pings, with a flat probability distribution between a specified minimum and maximum (inclusive); when the schedule's target number of pings is reached, it delivers reinforcement. (This part of the schedule is therefore approximately a VR schedule, except that reinforcement is delivered at the end of the time bin, rather than immediately upon responding, and that if the subject makes >1 response during the bin, additional responses are ignored.) The noncontingent part of the schedule similarly assigns a target number of pings (with a minimum and maximum, and a flat probability distribution between them) and delivers reinforcement when its target is satisfied; this therefore approximates a VT schedule (except that the noncontingent schedule does not receive pings when the subject is responding). Whenever either schedule reaches its target, it picks a new target. If the contingent requirements are low and the noncontingent requirements are high, there will be a high instrumental contingency; if they are identical, there will be no contingency; and so on. Currently restricted to certain laboratories. Additional option (4 Sep 2014): enter 0 for the minimum and maximum settings to disable part of the schedule (either the contingent or noncontingent part). |
Special cases
| • | The first response on contingent interval schedules (FI, RI, VI) is always reinforced (SimpleSchedules) or this option is configurable (ConcurrentSchedules). |
See also notes on reinforcement schedule types.
|