Time course of linguistic information extraction from consecutive words during eye fixations in reading

Preview frequency effects in reading: evidence from Chinese

Studies about sentence reading have shown that visual and lexical information beyond the currently fixated word can be integrated across fixations. The gaze-contingent boundary paradigm has been used widely to explore the extent to which parafoveal information can be processed before a word is fixated on. However, a critical review of the current literature suggests that unrelated mask previews are an unlikely baseline control with zero lexical activation, blurring the nature of experimental effects observed in the paradigm. The present study, therefore, aimed at shedding light on the effect of parafoveal mask properties through a manipulation of preview word frequency. Low-frequency preview words that are unrelated to target words elicited a larger interference than high-frequency preview words. We discuss implications of the preview frequency effect for computational models of eye-movement control in reading.

A Theoretical Analysis of the Perceptual Span based on SWIFT Simulations of the n + 2 Boundary Paradigm

Eye-movement experiments suggest that the perceptual span during reading is larger than the fixated word, asymmetric around the fixation position, and shrinks in size contingent on the foveal processing load. We used the SWIFT model of eye-movement control during reading to test these hypotheses and their implications under the assumption of graded parallel processing of all words inside the perceptual span. Specifically, we simulated reading in the boundary paradigm and analysed the effects of denying the model to have valid preview of a parafoveal word n + 2 two words to the right of fixation. Optimizing the model parameters for the valid preview condition only, we obtained span parameters with remarkably realistic estimates conforming to the empirical findings on the size of the perceptual span. More importantly, the SWIFT model generated parafoveal processing up to word n + 2 without fitting the model to such preview effects. Our results suggest that asymmetry and dynamic modulation are plausible properties of the perceptual span in a parallel word-processing model such as SWIFT. Moreover, they seem to guide the flexible distribution of processing resources during reading between foveal and parafoveal words.

Age-related sparing of parafoveal lexical processing

BACKGROUND/STUDY CONTEXT: This study examined the role of increased adult age in the processing of lexical information presented in foveal and parafoveal areas of the retina (right and left visual fields). Previous research has shown that older adults are able to compensate for age-related changes though a highly practiced skill (Salthouse, 1984 , Journal of Experimental Psychology: General, 113, 345-371). The authors examined if older adults would show a lexical-only parafoveal benefit.

METHODS

Two experiments were conducted on both younger and older adults. In Experiment 1, participants completed a lexical decision task that presented words in the fovea and parafovea in both visual fields. In Experiment 2, the task was a font discrimination task (nonlexical) with foveal and parafoveal presentation in both visual fields.

RESULTS

In Experiment 1, the authors observed word frequency effects for both foveal and right parafoveal presentation locations. This effect was present for both older and younger adults. Experiment 2 was a font discrimination task and there was no right parafoveal advantage for older adults on this task, suggesting that this effect observed in Experiment 1 was lexical in nature due to the highly overlearned nature of word recognition.

CONCLUSION

These results suggest that older adults may compensate for slower encoding time in reading by encoding text to the right of fixation more efficiently than younger adults. This suggests an asymmetrical change in the useful field of view that is lexical in nature.

Using reinforcement learning to examine dynamic attention allocation during reading

A fundamental question in reading research concerns whether attention is allocated strictly serially, supporting lexical processing of one word at a time, or in parallel, supporting concurrent lexical processing of two or more words (Reichle, Liversedge, Pollatsek, & Rayner, 2009). The origins of this debate are reviewed. We then report three simulations to address this question using artificial reading agents (Liu & Reichle, 2010; Reichle & Laurent, 2006) that learn to dynamically allocate attention to 1-4 words to "read" as efficiently as possible. These simulation results indicate that the agents strongly preferred serial word processing, although they occasionally attended to more than one word concurrently. The reason for this preference is discussed, along with implications for the debate about how humans allocate attention during reading.

Evidence for direct control of eye movements during reading

It is well established that fixation durations during reading vary with processing difficulty, but there are different views on how oculomotor control, visual perception, shifts of attention, and lexical (and higher cognitive) processing are coordinated. Evidence for a one-to-one translation of input delay into saccadic latency would provide a much needed constraint for current theoretical proposals. Here, we tested predictions of such a direct-control perspective using the stimulus-onset delay (SOD) paradigm. Words in sentences were initially masked and, on fixation, were individually unmasked with a delay (0-, 33-, 66-, 99-ms SODs). In Experiment 1, SODs were constant for all words in a sentence; in Experiment 2, SODs were manipulated on target words, while nontargets were unmasked without delay. In accordance with predictions of direct control, nonzero SODs entailed equivalent increases in fixation durations in both experiments. Yet, a population of short fixations pointed to rapid saccades as a consequence of low-level information at nonoptimal viewing positions rather than of lexical processing. Implications of these results for theoretical accounts of oculomotor control are discussed.

Contextual constraint and N + 2 preview effects in reading

Extracting linguistic information from locations beyond the currently fixated word is a core component of skilled reading. Recent debate on this topic is focused on the question of whether useful linguistic information can be extracted from more than one (parafoveally visible) word to the right of a fixated word (N). The current study examined this issue through the use parafoveal previews with a short and high-frequency next (N + 1) word, as this should increase the opportunity for the extraction of useful information from the subsequent (N + 2) word. Pairs of N + 2 words were selected so that contextual constraint was either high or low. Using saccade contingent display manipulations, preview of a N + 2 target word during word N viewing consisted of either a visually dissimilar nonword or a word. The results revealed a substantial drop in fixation probability for word N + 1 when the N + 2 preview was masked with a nonword. Furthermore, the masking of word N + 2 influenced its viewing duration even when word N + 1 was fixated prior to word N + 2 viewing. These results provide compelling evidence for the view that the linguistic processing can encompass more than one word at a time.

Eye movements in reading: some theoretical context

The study of eye movements has proven to be one of the most successful approaches in research on reading. In this overview, it is argued that a major reason for this success is that eye movement measurement is not just a methodology--the control of eye movements is actually part and parcel of the dynamics of information processing within the task of reading itself. Some major developments over the last decade are discussed with a focus on the issue of spatially distributed word processing and its relation to the development of reading models. The survey ends with a description of two newly emerging trends in the field: the study of continuous reading in non-Roman writing systems and the broadening of the scope of research to encompass individual differences and developmental issues.

The zoom lens of attention: Simulating shuffled versus normal text reading using the SWIFT model

Assumptions on the allocation of attention during reading are crucial for theoretical models of eye guidance. The zoom lens model of attention postulates that attentional deployment can vary from a sharp focus to a broad window. The model is closely related to the foveal load hypothesis, i.e., the assumption that the perceptual span is modulated by the difficulty of the fixated word. However, these important theoretical concepts for cognitive research have not been tested quantitatively in eye movement models. Here we show that the zoom lens model, implemented in the SWIFT model of saccade generation, captures many important patterns of eye movements. We compared the model's performance to experimental data from normal and shuffled text reading. Our results demonstrate that the zoom lens of attention might be an important concept for eye movement control in reading.

How preview space/time translates into preview cost/benefit for fixation durations during reading

Eye-movement control during reading depends on foveal and parafoveal information. If the parafoveal preview of the next word is suppressed, reading is less efficient. A linear mixed model (LMM) reanalysis of McDonald (2006) confirmed his observation that preview benefit may be limited to parafoveal words that have been selected as the saccade target. Going beyond the original analyses, in the same LMM, we examined how the preview effect (i.e., the difference in single-fixation duration, SFD, between random-letter and identical preview) depends on the gaze duration on the pretarget word and on the amplitude of the saccade moving the eye onto the target word. There were two key results: (a) The shorter the saccade amplitude (i.e., the larger preview space), the shorter a subsequent SFD with an identical preview; this association was not observed with a random-letter preview. (b) However, the longer the gaze duration on the pretarget word, the longer the subsequent SFD on the target, with the difference between random-letter string and identical previews increasing with preview time. A third pattern-increasing cost of a random-letter string in the parafovea associated with shorter saccade amplitudes-was observed for target gaze durations. Thus, LMMs revealed that preview effects, which are typically summarized under "preview benefit", are a complex mixture of preview cost and preview benefit and vary with preview space and preview time. The consequence for reading is that parafoveal preview may not only facilitate, but also interfere with lexical access.

Word-Initial Letters Influence Fixation Durations during Fluent Reading

The present study examined how word-initial letters influence lexical access during reading. Eye movements were monitored as participants read sentences containing target words. Three factors were independently manipulated. First, target words had either high or low constraining word-initial letter sequences (e.g., dwarf or clown, respectively). Second, targets were either high or low in frequency of occurrence (e.g., train or stain, respectively). Third, targets were embedded in either biasing or neutral contexts (i.e., targets were high or low in their predictability). This 2 (constraint) × 2 (frequency) × 2 (context) design allowed us to examine the conditions under which a word's initial letter sequence could facilitate processing. Analyses of fixation duration data revealed significant main effects of constraint, frequency, and context. Moreover, in measures taken to reflect "early" lexical processing (i.e., first and single fixation duration), there was a significant interaction between constraint and context. The overall pattern of findings suggests lexical access is facilitated by highly constraining word-initial letters. Results are discussed in comparison to recent studies of lexical features involved in word recognition during reading.

Eye Movements During Mindless Reading

Mindless reading occurs when the eyes continue moving across the page even though the mind is thinking about something unrelated to the text. Despite how commonly it occurs, very little is known about mindless reading. The present experiment examined eye movements during mindless reading. Comparisons of fixation-duration measures collected during intervals of normal reading and intervals of mindless reading indicate that fixations during the latter were longer and less affected by lexical and linguistic variables than fixations during the former. Also, eye movements immediately preceding self-caught mind wandering were especially erratic. These results suggest that the cognitive processes that guide eye movements during normal reading are not engaged during mindless reading. We discuss the implications of these findings for theories of eye movement control in reading, for the distinction between experiential awareness and meta-awareness, and for reading comprehension.

The 35th Sir Frederick Bartlett Lecture: Eye movements and attention in reading, scene perception, and visual search

Eye movements are now widely used to investigate cognitive processes during reading, scene perception, and visual search. In this article, research on the following topics is reviewed with respect to reading: (a) the perceptual span (or span of effective vision), (b) preview benefit, (c) eye movement control, and (d) models of eye movements. Related issues with respect to eye movements during scene perception and visual search are also reviewed. It is argued that research on eye movements during reading has been somewhat advanced over research on eye movements in scene perception and visual search and that some of the paradigms developed to study reading should be more widely adopted in the study of scene perception and visual search. Research dealing with “real-world” tasks and research utilizing the visual-world paradigm are also briefly discussed.

Parafoveal magnification: visual acuity does not modulate the perceptual span in reading

Models of eye guidance in reading rely on the concept of the perceptual span-the amount of information perceived during a single eye fixation, which is considered to be a consequence of visual and attentional constraints. To directly investigate attentional mechanisms underlying the perceptual span, we implemented a new reading paradigm-parafoveal magnification (PM)-that compensates for how visual acuity drops off as a function of retinal eccentricity. On each fixation and in real time, parafoveal text is magnified to equalize its perceptual impact with that of concurrent foveal text. Experiment 1 demonstrated that PM does not increase the amount of text that is processed, supporting an attentional-based account of eye movements in reading. Experiment 2 explored a contentious issue that differentiates competing models of eye movement control and showed that, even when parafoveal information is enlarged, visual attention in reading is allocated in a serial fashion from word to word.

Using E-Z Reader to model the effects of higher level language processing on eye movements during reading

Although computational models of eye-movement control during reading have been used to explain how saccadic programming, visual constraints, attention allocation, and lexical processing jointly affect eye movements during reading, these models have largely ignored the issue of how higher level, postlexical language processing affects eye movements. The present article shows how one of these models, E-Z Reader (Pollatsek, Reichle, & Rayner, 2006c), can be augmented to redress this limitation. Simulations show that with a few simple assumptions, the model can account for the fact that effects of higher level language processing are not observed on eye movements when such processing is occurring without difficulty, but can capture the patterns of eye movements that are observed when such processing is slowed or disrupted.

Eye movements and non-canonical reading: comments on

Kennedy and Pynte [Kennedy, A., & Pynte, J. (2008). The consequences of violations to reading order: An eye movement analysis. Vision Research, 48, 2309-2320] presented data that they suggested pose problems for models of eye movement control in reading in which words are encoded serially. They focus on situations in which pairs of words are fixated out of order (i.e., the first word is skipped and the second fixated prior to a regression back to the first word). We strongly disagree with their claims and contest their arguments. We argue that their data set was obtained selectively and the events they believe are problematic do not occur frequently during reading. Furthermore, we do not consider that Kennedy and Pynte's arguments pose serious difficulties for serial models of reading such as E-Z Reader.

Immediate and delayed effects of word frequency and word length on eye movements in reading: a reversed delayed effect of word length

Three experiments examined the effects in sentence reading of varying the frequency and length of an adjective on (a) fixations on the adjective and (b) fixations on the following noun. The gaze duration on the adjective was longer for low frequency than for high frequency adjectives and longer for long adjectives than for short adjectives. This contrasted with the spillover effects: Gaze durations on the noun were longer when adjectives were low frequency but were actually shorter when the adjectives were long. The latter effect, which seems anomalous, can be explained by three mechanisms: (a) Fixations on the noun are less optimal after short adjectives because of less optimal targeting; (b) shorter adjectives are more difficult to process because they have more neighbors; and (c) prior fixations before skips are less advantageous places to extract parafoveal information. The viability of these hypotheses as explanations of this reverse length effect on the noun was examined in simulations using an updated version of the E-Z Reader model (A. Pollatsek, K. Reichle, & E. D. Rayner, 2006c; E. D. Reichle, A. Pollatsek, D. L. Fisher, & K. Rayner, 1998).

The cost of question concealment: eye-tracking and MEG evidence

Although natural language appears to be largely compositional, the meanings of certain expressions cannot be straightforwardly recovered from the meanings of their parts. This study examined the online processing of one such class of expressions: concealed questions, in which the meaning of a complex noun phrase (the proof of the theorem) shifts to a covert question (what the proof of the theorem is) when mandated by a sub-class of question-selecting verbs (e.g., guess). Previous behavioral and magnetoencephalographic (MEG) studies have reported a cost associated with converting an entity denotation to an event. Our study tested whether both types of meaning-shift affect the same computational resources by examining the effects elicited by concealed questions in eye-tracking and MEG. Experiment 1 found evidence from eye-movements that verbs requiring the concealed question interpretation require more processing time than verbs that do not support a shift in meaning. Experiment 2 localized the cost of the concealed question interpretation in the left posterior temporal region, an area distinct from that affected by complement coercion. Experiment 3 presented the critical verbs in isolation and found no posterior temporal effect, confirming that the effect of Experiment 2 reflected sentential, and not lexical-level, processing.

Eye position changes during reading fixations are spatially selective

Intra-fixation location changes were measured when one-line sentences written in lower or aLtErNaTiNg case were read. Intra-fixation location changes were common and their size was normally distributed except for a relatively high proportion of fixations without a discernible location change. Location changes that did occur were systematically biased toward the right when alternating case was read. Irrespective of case type, changes of the right eye were biased toward the right at the onset of sentence reading, and this spatial bias decreased as sentence reading progressed from left to right. The left eye showed a relatively stable right-directed bias. These results show that processing demands can pull the two fixated eyes in the same direction and that the response to this pull can differ for the right and left eye.

Temporal overlap in the linguistic processing of successive words in reading: reply to Pollatsek, Reichle, and Rayner (2006a)

A. Pollatsek, E. D. Reichle, and K. Rayner argue that the critical findings in A. W. Inhoff, B. M. Eiter, and R. Radach are in general agreement with core assumptions of sequential attention shift models if additional assumptions and facts are considered. The current authors critically discuss the hypothesized time line of processing and indicate that the success of Pollatsek et al.'s simulation is predicated on a gross underestimation of the pretarget word's viewing duration in Inhoff et al. and that the actual data are difficult to reconcile with the strictly serial attention shift assumption. The authors also discuss attention shifting and saccade programming assumptions in the E-Z Reader model and conclude that these are not in harmony with research in related domains of study.

Tests of the E-Z Reader model: Exploring the interface between cognition and eye-movement control

This paper is simultaneously a test and refinement of the E-Z Reader model and an exploration of the interrelationship between visual and language processing and eye-movements in reading. Our modeling indicates that the assumption that words in text are processed serially by skilled readers is a viable and attractive hypothesis, as it accounts not only for "normal" reading data, but also for the pattern of decrements that occur when text is withheld from view during certain periods of time, such as in the boundary paradigm and the disappearing text paradigm. Our analyses also indicate (a) that lexical processing during reading is essentially continuous and (b) that the trigger for eye movements has to be a different (and prior) event to the trigger for shifts of covert spatial attention. In addition, the parameter values in our model are in accordance with what is known about such processes and should be taken as serious hypotheses for how long these processes last. Although a parallel model may be able to duplicate the predictions of the E-Z Reader model, we think that it is likely to be far less parsimonious and not nearly as good a heuristic device for using eye movements to understand language processing in reading.