Within Sessions Task

The within sessions task is one of the more commonly used variants of the delay discounting task in the animal literature. It is often a top choice when the study calls for an acute manipulation because one can attain a complete discounting curve within a single session (hence the name). This therefore requires fewer administrations of the acute manipulation of interest.

Design
The within sessions task is an operant procedure and therefore is typically held in an operant chamber. In this task, the animal has to make a choice between one of two levers in order to receive reinforcement. Pressing on one lever always yields a small, immediate reinforcer. Pressing on the other lever yields a large reinforcer after a specific delay. This delay typically starts out at zero and increases periodically throughout the session. For example, in Figure 1 the session contains 60 trials. During the first 12 trials, the delay to the large reinforcer is 0 seconds. During the second 12 trials, it becomes 10 seconds, then 20, 40, and 60 seconds. As can be seen in Figure 2, the animals initially have a strong preference for the large reinforcer when the delay is 0 seconds, but have switched their preference by the time the delay is 60 seconds.

Trial Format
There are two types of trials in a within sessions task: Forced choice trials and free choice trials. During a forced choice trial, animals only have the option to press one of the levers. There are typically at least two forced choice trials administered each time the delay to the large reinforcer is changed; one for each lever. The forced choice trials ensure that the animal experiences the changing contingencies (i.e., the fact that delay has increased for the large reinforcer). Following the forced choice trials are free choice trials in which both levers are available.

The design of any given trial is typically similar to the example shown in Figure 3. There is often a signal for the start of the trial, such as the houselight being turned on or levers being extended into the chamber. The program then awaits a response from the animal. In many studies, if the animal does not make a response within a specific time window, then the trial is scored as an 'omission' and the trial advances directly to the intertrial interval (ITI). Contrarily, if the animal does respond on a lever, then after the delay associated with that lever expires the animal receives the reinforcer associated with that lever. The trial then advances to the ITI. The length of the ITI depends on the study, but often it is adjusted so that the length of each trial is held constant regardless of the delay to the reinforcer. Following the ITI, the next trial begins.

Details
Several things need to be kept in mind when designing a within sessions experiment. First, the experimenter needs to address or acknowledge all of the caveats associated with operant conditioning. Second, the experimenter needs to counterbalance which side the small, immediate lever and large, delayed lever is on. Third, the experimenter needs to determine whether a cue is lit for the duration of the delay, thereby signaling the delay (see role of cue in delay discounting).

The aforementioned are either part of good experimental design of have been explicitly demonstrated to affect experimental results. The following are elements that often vary between studies, but the effect of this variance has not been explicitly investigated. One element is the range of delays and reinforcer magnitudes used, which can vary widely between studies. Additionally, several studies require the animal to start each trial by making a response, such as a nosepoke into the food magazine, before extending levers into the chamber. Other considerations are the use of retractable levers, the use of a houselight, the number of trials, the length of the ITI, the length of the session, and the requirement (or not) that the animal respond within a certain time frame.

Control tasks
There are a few control tasks that are sometimes used with the within sessions task. One is the magnitude estimation task. This task is identical to the normal within sessions task, but there is no delay to the large reinforcer. This is used to ensure that the manipulation is not affecting the animal's ability to discriminate the magnitude difference between the two reinforcers. Another task is a reversal task. In this task, the order of delays is switched (for example, in Figure 1 it would be 60, 40, 20, 10, and finally 0 seconds). This is done to ensure that there are no satiation effects and to ensure that the manipulation isn't simply altering bias.

Dependent variables
The primary dependent variable is the number of choices of the large, delayed reinforcer for each delay. This is typically displayed in a graph such as seen in Figure 2. Additional dependent variables of interest are the reaction time to press the lever, reaction time to retrieve the reinforcer, number of omitted/incomplete trials, and number of extraneous responses.

Review of the literature
Studies have investigated the effect of a variety of manipulations on behavior in the within sessions task.

Pharmacological manipulations
Several pharmacological agents have been tested on animals run in the within sessions task.

Implicated brain regions
Several brain regions have been implicated in delay discounting