Info

FIGURE 2.11 Learning constant during nonreinforcement. This figure demonstrates the effects of changing the learning constant that controls the magnitude of the changes in the weights during time steps where the model is not being reinforced. Each line represents the average of five independent peak procedure simulations. The heavy black line represents the standard value for this parameter in all other simulations in this work.

Timestep

FIGURE 2.12 Fixed-interval simulation, cumulative record. This graph represents the total amount of activity of the output node over the course of the FI. The scallop shape of the function is similar to that found in animals, with a very low level of activity early in the interval and building up to a high, steady rate late in the interval.

Timestep

FIGURE 2.12 Fixed-interval simulation, cumulative record. This graph represents the total amount of activity of the output node over the course of the FI. The scallop shape of the function is similar to that found in animals, with a very low level of activity early in the interval and building up to a high, steady rate late in the interval.

shown in Figure 2.12. The scallop is distinct, showing the early pausing typical of FI performance and the steadily increasing rate of response as the interval progresses. The response rate over the course of the interval is plotted in Figure 2.13.

2.4.1.2 Peak-Interval Timing

While the FI schedule is a good simple measure of timing, it only lets us see the subject's behavior leading up to the interval. Once the subject has been reinforced, its internal clock presumably resets, so we cannot see what it would have done after the point where it is normally reinforced. The peak-interval (PI) procedure developed by Catania (1970) and further explored by Roberts (1981) solves this problem as follows. The start of each trial is signaled by the presentation of an explicit timing signal such as a light or a tone, which stays on for the length of the trial. The first response after the interval produces food and the timing signal goes off. There is then a brief intertrial interval (ITI) before the next trial begins. These are referred to as food trials.

0 0

Post a comment