One of the major research goals of our department is to investigate the elementary processes underlying learning and memory in animal and human subjects and to make between-species comparisons. A problem associated with these comparisons is to identify equivalent tasks. In many cases, human and animal tasks differ to such an extent that direct comparisons are severely invalidated. For example, operant conditioning tasks in rats are usually performed in Skinner boxes, whereas in human studies, seemingly comparable paradigms are presented using a computer or slide projector. Especially computerised tasks have become increasingly popular, despite the fact that this approach has widened the gap between animal and human conditioning paradigms, hindering extrapolation of outcomes.

Figure 1.Illustration of the human Skinner box.

In order to enhance the comparability between animal and human operant conditioning tasks, a human analogue of the Skinner box was designed after an idea by Dr. W.H.I.M. Drinkenburg. As can be seen in Figure 1, this Skinner box shows a strong resemblance to the traditional animal Skinner box. The apparatus allows for the presentation of visual, auditive, and spatial stimuli as well as of reinforcements. The intelligence panel contains two speakers, one left and one right, while two green lights are placed below the speakers, and two 'levers' at the bottom of the panel. Between the two buttons a 'feeder' is placed containing two sensors on each side wall to detect a 'feeder visit'. The reinforcer, a coin, is delivered behind a magnetically regulated flap and can be obtained by pushing the flap backwards. The coins are maintained in a reservoir, which can be filled up through a slot. Two lights, a green and a red one, are placed above the 'feeder'. Additionally, the possibility to use a score counter instead of coins is enabled by a display at the top of the panel. A Macintosh computer with specially designed software runs the experiments. Lights and sounds can be used as conditioned stimuli, pressing one of the buttons or visiting the feeder as a response and receiving a point or coin can function as a reinforcer. It is possible to measure response latencies, response accuracy, and set performance criteria. All well-known experimental designs can be programmed and run. For instance, fixed- or variable ratio schedules, time interval tasks, conditional discriminations, occasion-setting paradigms, and countless other tasks and schedules. The most appealing feature of this human Skinner box panel, next to its physical resemblance, is the correspondence in task demands with the animal Skinner box. This gives the experimenter the opportunity to observe and compare learning and memory processes involved in operant conditioning in humans and animals. The apparatus has been used to study effects of aging on problem solving, while further studies are planned.

© 2000 Noldus Information Technology b.v.