Oscillator Models
There exist today a number of time-based (or partly time-based) approaches known as oscillator models. Here the assumption involves an entity, or many entities, that oscillate across time. Time is marked—kept track of—by these oscillations.
Burgess and Hitch (2006), developing a connectionist, oscillator-based model, described the various states as “internal context signals,” and noted that these had some properties that could be described as temporal, but also some that should be described as event based. The temporal aspect involved the pattern of the different signals following one another through time. But, under this model, it is the presentation of the item (an event) that triggers entry into each new (internal context) state, and not the temporal moment when each item appears. For instance, an item could be presented at the rate of one every 30 seconds, or of one every 2 minutes, and the critical factor (causing a change in context state) would not be a set amount of time between the appearance of a given item and the next item in the series, but the fact that it is, precisely, the next item. Thus, if a word is presented every half-second, the system could be operating in such a manner as to form an association following each half-second in time. Under the Burgess and Hitch model, the association would only be formed on the appearance of a new item. A somewhat similar model has been offered by Anderson and Matessa (1997).