Introduction

The senses available to animals tend to be divided into at least two distinct portions. The eye is distinct from the visual cortex, the ear from the auditory cortex. It is possible to study the properties of the eye without reference to the visual areas of the brain. The study of vision as a whole would of course involve both, but the source of visual information is a distinct unit, isolated from the rest of the visual system.

The study of interval timing, how animals learn and remember short periods of time, offers many unique challenges, not least of which is the fact that there is no readily apparent sensory organ for the passage of time. We cannot dissect the "eye" of time to determine its properties and trace its connections to the rest of the mind. Rather, we are forced to infer its existence, properties, and functioning from behavior, and to theorize and test without any certain knowledge of the source of temporal information.

Several types of potential sources of temporal information have been proposed and tested with varying degrees of success. Another common metaphor for these sources is a clock or stopwatch, a mechanism that changes predictably over the course of the interval and can be started, stopped, and reset as necessary. This metaphor is quite natural, since a stopwatch is a mechanical solution to the same problem as our biological sense of time.

The two metaphors of a physical sense and a stopwatch both include the assumption that there is, in addition to the clock, a separate general learning mechanism. The eye itself does not learn to recognize faces; the stopwatch does not learn to turn itself on and off. This learning mechanism has historically received much less attention than the clock mechanism. For example, most timing models presuppose that the animal already knows the signal to start timing and that the problem being presented is exclusively one of timing.

This work is an attempt to cast new light on the field of animal timing by approaching the problem from the opposite direction. If a general learning mechanism is necessary for a complete explanation of animal timing, is it possible that it is sufficient as well? Theoretically, a general learning model should be able to learn patterns of responding in time if it is presented with stimuli that hold different values over the course of the interval. These timing stimuli have previously been produced by a special-purpose clock process, but what if the learning model could find such stimuli within itself? A dynamic model changes over time by definition, raising the possibility that these changes can be used as an internal clock.

The primary issue addressed by this work is whether a general learning model can fulfill the functions of a clock. If so, it opens the door to some remarkable questions about our understanding of interval timing. Is a separate special-purpose timing model necessary at all? If a general learning mechanism must exist and can be shown to be able to produce interval timing behavior, there would be no need for a separate interval timing clock to evolve. If a general learning model cannot perform any of the functions of a clock, can that failure tell us which aspects of timing are necessarily properties of the clock? If any learning mechanism can time, does that mean all learning mechanisms potentially play a role in interval timing? If successful, this project presents a serious challenge to our current theories of interval timing.

0 0

Post a comment