Yes, I’m going to enjoy this.
Expectancy, attention, and time
R Barnes, MR Jones – Cognitive psychology,
“In the present research, we use auditory sequences to pursue the possibility that sequence time structure controls attending.”
Yes, I’m going to like this. Expectation without priors? Right up my alley.
“In the context of more familiar deﬁnitions of expectancy that draw heavily upon prior knowledge this claim may seem unusual. To provide a context for it, we refer to the insightful analysis of Kahneman and Tversky (1982), who challenge a common view of expectancy: ‘‘. . . rules that govern perceptual expectancies differ from rules of probability theory’’ (p. 148).”
Pattern-based approaches to expectancy go beyond lo- cal cues and ?rst-order conditional probabilities to view contextual informa- tion in terms of relationships among features or elements. For instance, in auditory sequences, deterministic arrangements of particular pitches (e.g., ascending verus descending pitch sequences) may contribute to the guidance of attending (Boltz et. al., 1981; Howard, O’Toole, Parasuraman, & Bennet, 1984; 1986; Spiegel & Watson, 1981; see Leek, 1987). In fact, some expec- tancy theories link effects of serial pattern constraints on performance di- rectly to expectancies (Garner, 1974; Jones, 1974; Narmour, 1992; Restle, 1970; Simon & Kotovsky, 1963). A digit sequence such as: 1-2-3-2-3- X tends to evoke the expectancy that X will be ‘‘4,’’ although prior exposure to ‘‘4’’ is nil (for reviews see Jones, 1974, 1978, 1981).
The present research suggests a third way to view expectancy. It builds upon the pattern-based view of expectancy, but expresses a more dynamic interpretation of expectancy by explicitly connecting it to stimulus time rela- tionships within a pattern context. By stimulus time relationships we mean the rate and rhythm created by elements comprising a sequence, i.e., the pattern’s temporal layout. We claim that people tacitly rely on these relation- ships to anticipate the ‘‘when’’ as well as the ‘‘what’’ of future elements (Jones, 1976). Time relationships have been shown to affect an individual’s monitoring of sequence elements (pitch, duration, and other features) (Boltz, 1989; Boltz, Jones, & Kidd, 1981; Cutler, 1976; Dowling, Lung, & Herrbold, 1987; Jones, Boltz, & Klein, 1993; Klein & Jones, 1996; Martin, 1972; Shields, McHugh, & Martin, 1974; Skelly, Jones, Goodyear, & Roe, 1999; Wright & Dai, 1994). Much of this research is consistent with the general idea that expectancy involves anticipations about ‘‘when’’ something will occur in the future; furthermore, some experiments indicate that these tempo- ral anticipations can affect judgments about ‘‘what’’ is expected. It is possi- ble that when sequences of elements are involved, the time relationships among successive elements contribute to a dynamic pattern-based control of focal attending which features temporal anticipations about future elements (Jones, 1976; Jones & Boltz, 1989; Martin, 1972). In sum, bottom-up expec- tancies may exist that are determined by pattern relationships, including time relationships, with a local context. In the next section, we ?esh out this third approach to expectancy found in a dynamic attending approach.
IOI INTERONSET INTERVAL. – yes, I’m finding the missing concepts.
Thus, moment-to-moment attending to events such as speech and music is controlled, in part, by their relational properties, e.g., rate and rhythm.Successive IOIs are assumed to engage internal attending rhythms which, in turn, direct attending (Jones, 1976).Attending rhythms are instantiated as oscillators capable of entraining, i.e., ‘‘locking into’’ the ongoing time structure. Essentially, dynamic attending models present an entrainment approach to attention and expectancy (but see Desain, 1992 for a nonentrainment, view).These models share the fol- lowing: (1) enlistment of one or more underlying periodicities (oscillators) that de?ne a relevant time metric on a ratio scale, e.g., 1:1, 1:2; (2) reliance on stimulus properties, such as relational dependencies among successive context IOIs, as a source of low-level attentional control; and (3) assumptions that momentary attending, based on internal oscillations, is adaptive, en- abling more or less synchronous attentional tracking of both regular and irregular temporal sequences.
back to entrainment
“At any point in a pattern, a temporal expectancy is determined by the current period and phase of an oscillator. “
“But memory is intrinsic to the process as well; it is expressed as working memory in that it is captured by the current period of the oscillator; this provides an internal estimate of sequence rate, a running memory of the sequence’s time intervals. Finally, perception is critical to adaptation of the oscillator; it occurs at the time of any tone onset where the disparity, i.e., temporal contrast, between an actual onset and the expected onset provides a violation of expectancy. When this disparity distinguishes the ending of a to-be-judged time interval, it forms the basis for a time estimation response as well as triggering adapta- tive activity (see also Jones & Boltz, 1989).