Week 11-eye movement and reading Flashcards
What does reading require?
Requires different perceptual and cognitive processes combined with knowledge of language and grammar
Define Orthography
The study of word letters and word spelling
Define Phonology
The study of sounds of words and parts of words
Define Semantics
The meaning conveyed by words, phrases and sentences
Define Lexicon
The Language / knowledge –Vocabulary of a person (e.g.,
English)
What is the normal pattern of
eye movements during reading?
*Needs highly detailed (central) vision for accurate perception
of word form
*I.e. needs to make a fixation on (almost) every word
*Makes short saccades between the words
Define saccades
A rapid movement of the eye between fixation points.
Define regression in relation to eye movement
About 10%-25% of eye movements move backwards in the text to reprocess information when we read/misread it (can also be known as regression saccades)
Define foveal vision
Highly detailed central vision
Define parafovea
A region in the retina that circumscribes (restricts) the fovea and is part of the macula lutea.
What’s the oculomotor pattern?
Just a demonstration of how we read a sentence
Define refixation
It’s fixation but the 2nd+ time you go back to the word
Define a return sweep (AKA large regressive saccade)
Return to first part of the text but second line (called a return sweep) only ever make this for the new line of a text if it’s just going back to first sentence that’s called regression
What are some key features of the Oculomotor pattern?
*Fixations ~200ms
*Most words receive at least one direct fixation
*Skips ~20%
*Saccades ~15-40ms, ~5-9 chars
*Mostly progressive ~10-15% regressions
*= average reading speed ~250-350wpm (treatments for blind/deaf people consider reading speed, so important)
What factors can influence EM behaviour?
1.Characteristics of the visual system
2.Attention
3.Online cognitive processing of text e.g., in an exam would pay attention to words longer
4.Personal characteristics
5.Task differences (reading exam vs casual leisurely reading)
6.Text differences
What are the key features of rod cells?
*Lower acuity
*Peripheral retina
*Monochromatic
*Works at lower light intensities
What are the key features of cone cells?
*Better acuity
*Central retina
*Trichromatic (colour vision)
Where is the fovea located?
It’s pit at the back of the eye (which has a higher conc. of cone photoreceptor cells)
True or false: directly looking at
something = high detail
True, the distance goes out to the periphery of the vision + the further out to the side from where you’re looking = much less
detail (so more difficult to identify things)
Visual system: Define visual crowding
Letters compressed with no gaps
Visual system: Explain what is meant by “Greater further into the periphery”
Further into the periphery=loss of spatial resolution (e.g., loss of detail=more pixelated) and greater convergence to other cells
Visual system: What does Bouma’s (1970) graph show?
-shows difficulty distinguishing individual letters (experiment shows visual crowding)
-found letter moved further from central letter to periphery, letters flanked (2 x’s) by other letters, makes it harder to identify middle later so decrease in accuracy due to crowding effect
-further out of periphery=worse resolution so harder determining/identifying word due to visual crowding
What’s the role of attention in reading?
*Mostly deployed to upcoming text (ahead of the point of fixation) to begin processing
*Small allocation behind the fixation
*Single line
Define word identification span
the area where your focused attention is going on and tells us about the area we’re focused enough on to identify words (slightly asymmetric (~4L-7R)– close to
average progressive saccade size (7 characters ahead seen in BCP))
How did Underwood & McConkie,(1985) investigate with boundary change paradigm?
-used an eyetracker where when they passed a word on the boundary, it triggered a change in words in the sentence
-either explicitly ask them if they noticed a change OR implicitly (seeing whether processing is effective e.g., longer fixation duration in the word that’s changed)
-not a lot of people noticed this different=7 characters ahead pay attention to
What factors in attention can help to plan eye movements and to start constraining your lexical identity search to process the word quicker? (According to the Gaze-contingent moving window paradigm (McConkie & Rayner 1975))
*Perceptual span
*Asymmetric – spans around 4 characters to left and 15 characters to right of fixation
*Constrained under higher foveal load
*Information about word form, inter-word spacing
What’s the Gaze-contingent moving window paradigm? (McConkie & Rayner 1975)
-uses eye tracker where wherever you’re reading, it has a certain number of characters ahead/behind the bit you’re on (aka a window) BUT the rest of it is either blurred or covered by other letters e.g. XXXX
-shows word identity not necessarily used
How do we know this is based on attention and not the visual system with what we perceive?
*Reversed for orthographies that are read the other way (e.g. Hebrew)
*Shrunk for more dense (/complicated orthographies; e.g. Chinese, perceptual span=1L (1 character to the left) and 3R) showing a lot more attention required to process hence why the span shrunk
*Not affected by parafoveal magnification (making the text size larger the same is seen regardless of text size)
What 2 competing models are there in attention?
1.E-Z Reader – serial
2.SWIFT – parallel
Explain the E-Z Reader model (Reichle et al. 2003, 2009; Reichle 2011)
*Attention deployed to one word at a time (aka serially)
2 stages of lexical processing:
1st stage – Familiarity check (n) (saccade planned after first stage is complete via oculomotor system)
*2nd stage completed as well before saccade executed= attention shifts to the next word – -both stages of lexical processing of word (n+x) can’t be completed before the saccade programme is implemented (intervening words are ‘skipped’)
*If 2nd stage isn’t completed before the saccade is executed= a regression may be programmed to return and complete processing
Explain the SWIFT model (Engbert & Kliegl, 2011; Engbert et al. 2005)
*Parallel model (several words attended to/processed at a time)
*Words around point of fixation/visual scene activated to different extents, dependent on early low-level processing which occurs in parallel (ones in foveal vision have highest activation as they have the best/most sensory info coming in)
*Saccadic shifts are programmed at random intervals to the word receiving highest activation
*Shift might be delayed up to a point by higher processing difficulty on foveated word n than planned saccade target – if delay isn’t sufficient to complete processing before a saccade, activation for this word will be increased and therefore increase likelihood of later regressive saccade
What does SWIFT stand for?
Saccade generation With Inhibition by Foveal Targets
What has online processing shown?
*Clear effects of certain text characteristics on metrics such as fixation duration
*Looking at effect of different manipulations on different measures can help us understand how text is processed during reading
*Manipulate a target word itself or sentence context and compare oculomotor behaviour on target word e.g., word length, word frequency, sentence predictability, sentence plausibility, syntactic ambiguity etc.
Single fixation duration, first fixation duration, gaze duration, go-past duration, skipping probability, refixation probability,: What do these mean?
Basically, higher duration numbers and lower probabilities mean it’s more difficult to process
What’s involved in early processing?
-single fixation duration
-first fixation duration
-skipping probability –the initial familiarity/identification stages
What’s involved in late processing?
-total gaze duration
-regression probability
-go-past duration – the stuff that comes after
-integration into wider representation of the text
What’s the effect of word length on reading?
*Short words are processed quicker
*Perceptual effect: same pattern X-strings
Opening night was held at a RED theatre in the centre of London. (XXX)
Opening night was held at a SPECIAL theatre in the centre of London. (XXXXXXX)
What’s the effect of word frequency on reading?
*Less frequently encountered words take longer to process (e.g. high freq=town/low freq=cove)
*It’s a measure of successful word identification
What did Fitzsimmons, Weal, & Drieghe 2014 WebSci find on the effect of word frequency on reading when it involved reading on websites?
*Compared normal reading with skim-reading in P’s
*Typical text or text with hyperlinks in it
F:
*Loss of word frequency effect for unlinked words demonstrates that when skimming text with hyperlinks, readers only tend to
process the linked words properly
*Real-world application: Don’t put
important information in between
hyperlinks!
Wahat’s the effect of predictability on reading?
Words that are easier to predict from preceding context are
processed quicker
E.g.:
High predictability:
Russell had hurt his hand in the door of the car. He had trapped his finger while playing.
Neutral:
Russell had to go to the hospital. He had trapped his finger while playing. (single fixation duration longer for this than HP)
What more general effects of text can impact reading?
Font, language, difficulty of content and display format
