eyetracking in clinical settings: a pilot study · eyetracking in clinical settings: a pilot study...
TRANSCRIPT
Eyetracking in clinical settings: A pilot study
Ana Matić[email protected]
Faculty of Education and Rehabilitation Sciences, Department of Speech and Language PathologyLaboratory for Psycholinguistic Research
University of Zagreb
Child Language & Eyetracking: Analyses and Rationale (CLEAR 2016)
Potsdam, 12 Mar 2016
Laboratory for Psycholinguistic Research • Since 1999
• Associates from different professions (speech and language pathologists,psychologists, linguists, croatists)
• Interdisciplinary research - the aim of improving the understanding of speech andlanguage development, processing, disorders and their causes with the use oftechnologically advanced methods
• Equipment for recording evoked potentials, eye movement monitoring devices,equipment for measuring reaction time, a large number of linguistic,psychological, and neuropsychological tests
Eyetracking methodology in Speech and Language Pathology
• Eyetracking methodology is slowly becoming more and more recognised in the field of speech and language pathology (SLP)
• It can be used as a diagnostic instrument and also for clinical purposes
Language impairment
• Impaired comprehension and/or use of spoken, written and/or other symbol systems
• It may involve (1) the form of language (phonology, morphology, syntax), (2) the content of language (semantics), and/or (3) the function of language in communication (pragmatics) in any combination (ASHA, 1993)
• Increases risk for reading deficits (decoding and reading comprehension) and writing disability (Schuele, 2004)
!!!• Early detection helps reduce later negative consequenses mostly manifested in
academic settings
How language impairment influences reading?
• Language impairment can be the underlying cause of dyslexia
• Dyslexia - reading performance is significantly behind what may be expected based on person’s intelligence and the followed reading education and there are no neurological damages (Lyon et al., 2003)
• Eye movements reflect the difficulties that dyslexic individuals have reading and are not the cause of the reading problem (Rayner, 1998)
Dslxᴉc vs. good reader(Adler-Grinberg and Stark, 1978; Eden et al., 1994;
Fischer and Weber, 1990)
Good reader Dyslexic reader
Skilled at recognising printed words/letters/phrases Struggles with each letter of a word
Skilled in interpreting peripheral information (word length and topography)
Does not interpret peripheral information: characteristics of a child still developing reading skills
Smaller number of fixations and shorter fixation duration
More fixations, longer fixation duration and smaller fixation span
More eye movement stability during fixations Worse eye movement stability
More words per minute Less words per minute: more time – less text
Normal saccade reaction times Abnormal saccade reaction times
Some studies indicate there are no differences in the saccadic movements or in the number of regressions
Types of text: recommendations
• Fonts, backgrounds and types of text have an impact on readability (Rello and Baeza-Yates, 2013; Rello et al., 2013; Rello et al., 2012)
• Changes in the presentation of text might alleviate some problems children and adults with dyslexia encounter while reading (Gregor and Newell, 2000)
• Cursive letters are hard to read (Rello and Baeza-Yates, 2013)
• Dyslexie font and Arial are recommended (British Dyslexia Association, 2012; Leeuw, 2010)
Current study and its aims
• To compare the reading proficiency of one second-grade child with language impairment (LI) and one second-grade child without LI on 4 texts that varied in fonts and linguistic complexity
• Observed variables: total reading time, total fixation duration, number of fixations, saccades and regressions
• Device used: SMI Red-M system
Questions
• Do child with LI and child with typical language development (TLD) who are NOT YET SKILLED READERS nevertheless differ in their reading performance?
• Does dyslexia-friendly font significantly improve reading performance for children with LI?
• Do types of text influence equally on these two groups?
• On which variables do these two groups differ the most?
Characteristics of texts used
• Text 1 and Text 2 different fonts
• Text 3 and Text 4 different linguistic complexity
Text 1 Font dyslexie; 358 graphemes; 13 sentences
Text 2 Font Courier New; 350 graphemes; 12 sentences
Text 3 Font Arial; 496 graphemes; 10 sentences – 3 simple and 7 compound; on average 11 words/sentence + simple, understandable and frequent words; content not toodemanding
Text 4 Font Arial; 638 graphemes; 7 sentences – all compound and complex; on average 19 words/sentence + more complex and less frequent words; content more demanding
Results
175166 178123
218201
372288
8056092458
104940
147676
0
50000
100000
150000
200000
250000
300000
350000
400000
Dyslexia Friendly Font Complex Font Simple Text Complex Text
ms
Total Reading Time
LI TLD
Graph 1. Total Reading Time of participant with LI and participant with TLD in four different conditions
933
809,7
1007,9
1390,1
502,8
390,2
576,3 570,7
0
200
400
600
800
1000
1200
1400
1600
Dyslexia Friendly Font Complex Font Simple Text Complex Text
ms
Average Fixation Duration
LI TLD
Graph 2. Average Fixation Duration of participant with LI and participant with TLD in four different conditions
180
211199
255
146
204
166
230
0
50
100
150
200
250
300
Dyslexia Friendly Font Complex Font Simple Text Complex Text
Nu
mb
er
Fixation Count
LI TLD
Graph 3. Number of Fixations of participant with LI and participant with TLD in four different conditions
150
257281
513
233249
278
400
0
100
200
300
400
500
600
Dyslexia Friendly Font Complex Font Simple Text Complex Text
Nu
mb
er
Saccade Count
LI TLD
Graph 4. Number of Saccades of participant with LI and participant with TLD in four different conditions
43
29
9
25
31
47
9
15
0
5
10
15
20
25
30
35
40
45
50
Dyslexia Friendly Font Complex Font Simple Text Complex Text
Nu
mb
er
Into-word and Out-Of-Word Regression Count
LI TLD
Graph 5. Number of Regressions of participant with LI and participant with TLD in four different conditions
28,667
11,284
3,203
4,873
13,305
18,876
3,237 3,75
0
5
10
15
20
25
30
35
Dyslexia Friendly Font Complex Font Simple Text Complex Text
Per
cen
t
Regression/Saccade Ratio
LI TLD
Graph 6. Regression/Saccade Ratio of participant with LI and participant with TLD in four different conditions
Let us look more closely …
Scan Path of participant with LI: Text 1 Scan Path of participant with TLD: Text 1
Answers to our questions
• LI and TLD differ in their reading performance
• Dyslexia-friendly font appears not to significantly improve reading performance for children with LI in the stage when reading is still not automatised
• Types of text influence more on the LI participant, especially linguistic complexity
• LI and TLD participant differ in the Total Reading Time, Number of Fixations andTotal Fixation Duration; small differences are observed in the Number of Saccadesand Number of Regressions
We should not rush into conclusions …
• Small sample – no generalisation
• No standardised texts
• Participants are still in the process of reading automatisation
• Relatively small number of variables
New questions arise …
• Which measures are best to be taken and are most informative for the reading paradigm?
• Should the SLPs conducting a research or using eyetracking in clinical settings additionally adjust the text stimuli for the participants with LI (e.g., font, text alignment, text complexity, background colours,…)?
• How to handle data?
• Adler-Grinberg, D., Stark, L. (1978). Eye movements, scanpaths, and dyslexia. Optometry & Vision Science, 55(8), 557-570.
• American Speech-Language-Hearing Association. (1993). Definitions of communication disorders and variations [Relevant Paper]. Available at: www.asha.org/policy.
• British Dyslexia Association (2012). Dyslexia style guide. Available at: http://www.bdadyslexia.org.uk.
• Eden, G. F., Stein, J. F., Wood, H. M., Wood, F. B. (1994). Differences in eye movements and reading problems in dyslexic and normal children. Vision research, 34(10), 1345-1358.
• Fischer, B. Weber, H. (1990). Saccadic reaction times in dyslexic and age-matched normal subjects. Perception, 19, 805-818.
• Gregor, P., Newell, A. F. (2000). An empirical investigation of ways in which some of the problems encountered by some dyslexics may be alleviated using computer techniques. In Proceedings of the fourth international ACM conference on Assistive technologies (pp. 85-91). ACM.
• Leeuw, R. (2010). Special font for dyslexia?. Master’s thesis. University of Twente.
• Lyon, G. R., Shaywitz, S. E., Shaywitz, B. A. (2003). A definition of dyslexia. Annals of dyslexia, 53(1), 1-14.
• Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological bulletin, 124(3), 372.
• Rello, L., Baeza-Yates, R. (2013). Good fonts for dyslexia. In Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility (p. 14). ACM.
• Rello, L., Baeza-Yates, R., Dempere-Marco, L., Saggion, H. (2013). Frequent words improve readability and short words improve understandability for people with dyslexia. In Human-Computer Interaction–INTERACT 2013 (pp. 203-219). Springer Berlin Heidelberg.
• Rello, L., Kanvinde, G., Baeza-Yates, R. (2012). Layout guidelines for web text and a web service to improve accessibility for dyslexics. In Proceedings of the international cross-disciplinary conference on web accessibility (p. 36). ACM.
• Schuele, C. M. (2004). The impact of developmental speech and language impairments on the acquisition of literacy skills. Mental retardation and developmental disabilities research reviews, 10(3), 176-183.