To the best of our knowledge, this is the first study comparing oculomotor patterns in dyslexic and non-dyslexic students during a phonological lexical decision task. The most important findings are: 1) dyslexics made fewer single fixations than controls, while controls made more single than double fixations; 2) dyslexics made more fixations in the consonant strings condition than in the pseudowords condition than controls; 3) both groups presented similar durations of single as well as double fixations. Lastly, with respect to the type of stimulus, the oculomotor pattern was similar in the two groups of participants as we reported: 4) more single fixations in the word than in the pseudohomophone and symbol conditions; 5) more frequent single than double fixations in the word and the pseudoword conditions; 6) more single saccades in the symbol condition compared to the word and pseudoword conditions.
These findings are discussed below.
The main difference between dyslexic and control participants found in this study was that dyslexics made fewer frequent single than double fixations when compared to the controls. These findings are in accordance with the visuo-attentional deficit in dyslexics suggested by Bosse et al. 9; the lower number of letters processed in parallel and the shorter visual attentional span (as shown in Table 1) could lead to the abnormal oculomotor pattern reported in dyslexics during reading, in line also with the findings reported in dyslexic children during a reading task as well as during a visual search task 18. In addition, the absence of Condition by Group interaction highlights difficulties in all stimuli independently of their linguistic information. What strengthens the interpretation in favor of a visuo-attentional deficit in dyslexics is the occurrence of several single and double fixations, which was significantly larger in consonant strings as compared to pseudowords. This result could be explained by the fact that dyslexics may have difficulties at an early level of visual processing including perceptual processing.
Table 1
Assessment of reading and other cognitive functions of participants. Mean value (± standard error of the mean) for the different tests run for the two groups of participants (control readers, dyslexic readers). P values are also reported.
| Controls Ν = 18 | Dyslexics Ν = 14 | P values |
Age (years) | 21.1 ± 0.5 | 21.2 ± 0.6 | 0.895 |
No-meaning text reading (words correctly read/minute) | 143 ± 5.9 | 102 ± 4.3 | < .001 |
No-meaning text reading (Reading efficiency Score, CTL) | 430 ± 18 | 305 ± 13 | < .001 |
Text reading (words correctly read/minute) | 205 ± 5.8 | 148 ± 7.8 | < .001 |
Regular word reading (score/20) | 19.3 ± 0.2 | 18.7 ± 0.3 | 0.066 |
Regular word reading (time in seconds) | 11.0 ± 0.6 | 19.9 ± 1.9 | < .001 |
Irregular word reading (score/20) | 18.8 ± 0.4 | 17.6 ± 0.5 | 0.055 |
Irregular word reading (time in seconds) | 10.6 ± 0.6 | 18.1 ± 1.7 | < .001 |
Pseudoword reading (score/ 20) | 18.6 ± 0.3 | 17.0 ± 0.6 | 0.015 |
Pseudoword reading (time in seconds) | 16.2 ± 1.4 | 33.2 ± 2.5 | < .001 |
Initial phoneme deletion (score/ 10) | 8.9 ± 0.3 | 7.9 ± 0.7 | 0.151 |
Initial phoneme deletion (time in seconds) | 38.8 ± 2.8 | 50.3 ± 2.4 | 0.005 |
Spoonerisms (score/ 20) | 18.7 ± 0.3 | 14.1 ± 1.2 | < .001 |
Spoonerisms (time in seconds) | 89.2 ± 8.5 | 216 ± 31 | < .001 |
Non-word repetition (score/ 20) | 19.0 ± 0.1 | 18.6 ± 0.3 | 0.220 |
Non-word repetition (time in seconds) | 68.8 ± 2.2 | 78.6 ± 3.2 | 0.014 |
Rapid automatized naming (RAN) letter (score/ 50) | 49.6 ± 0.1 | 47.7 ± 1.7 | 0.211 |
Rapid automatized naming (RAN) letter (time in seconds) | 16.1 ± 0.7 | 22.4 ± 1.5 | < .001 |
Visuo-attentional span (score/ 100) | 93.1 ± 1.4 | 75.9 ± 4.2 | < .001 |
Similarities subtest WAIS IV | 10.8 ± 0.5 | 10.1 ± 0.6 | 0.390 |
Matrices subtest WAIS IV | 10.5 ± 0.6 | 10.1 ± 0.3 | 0.614 |
However, we cannot exclude the existence of a phonological deficit. Besides, our behavioral data indicated a significant difference between dyslexic and control participants in tests measuring phonological awareness (i.e., initial phoneme deletion, spoonerisms; see also Table 1). The fact that dyslexics made fewer single fixations (and more double fixations) may indicate that their reading abilities are deficient, and they use the sublexical grapheme-to-phoneme conversion procedure more, whatever the type of stimulus to be read, in comparison to control readers who may use the lexical route more, since they fixate stimuli only once 34,36.
Similar results were found in dyslexic children by De Luca et al. 24 who reported more frequent fixations during reading short pseudowords and longer stimuli (words and pseudowords) in dyslexics as compared to controls, and by Hutzler et al. 27 who found a higher number of fixations in dyslexics as compared to normal children when reading pseudowords.
Furthermore, a similar result was found by Denis-Noël et al. 34 in dyslexic students; they reported longer fixation durations during the reading of inconsistent words, for which phonological processing is more demanding. Note, however, that in the present study fixation durations were not different in the two groups of participants; this could be due to the type of stimuli used, which were quite easily processed and short. We suggest that the lower occurrence of single fixations in dyslexics as compared to controls may indicate a delayed activation of the phonological code, expressed by the need to make more fixations in order to read words and word-like stimuli. Together the above-mentioned results could support the phonological deficit hypothesis.
The absence of a Group by Condition interaction with stimuli requiring a grapheme-to-phoneme conversion cannot lead to a straightforward indication in favor of the phonological deficit, but may indicate that the different eye movement patterns found in dyslexics may be attributed to both phonological and to visuo-attentional deficits. Furthermore, the observation that dyslexics made fewer single fixations irrespective of whether the stimulus contained lexical or sublexical information, strengthens the hypothesis of the presence of both phonological and visuo-attentional difficulties in dyslexics. Unfortunately, we are not able to discriminate whether this abnormal oculomotor pattern is the origin or the cause of dyslexia 37. This could be highlighted by an interaction with the type of stimulus, that could better indicate whether a deficit has a phonological or a visuo-attentional cause.
With respect to the duration of double fixations and the number and amplitude of saccades, we found a similar oculomotor pattern in both groups. More precisely, in the case of double fixations, both groups showed longer durations during the first rather than during the second fixation; this finding is in line with Rayner 38 who suggested that during the first fixation the reader processes the visual features of the word, such as its length, letter shape, and spacing, which can take longer than subsequent fixations. The absence of any difference in oculomotor pattern between the two groups of subjects could also be attributed to the use of short stimuli in our experiment. Earlier studies examining eye movements when reading short and long words 24 reported similar findings for short words while for longer stimuli more fixations were needed 24,25.
In our study, we found more frequent single fixations in the word than in the pseudohomophone and symbol condition. This result can be partially explained by the fact that reading words is acquired through the direct route of reading 39, whereas reading pseudohomophones, which are stimuli with no orthographic representation, need the application of grapheme-to-phoneme conversion rules. In the case of the comparison between the word and symbol, the lower occurrence of single fixations reported in symbols could be due to the fact that this type of stimulus was quite simple to distinguish with respect to words.
Surprisingly, our data did not show any significant group difference in terms of the occurrence of fixations and in fixation duration when reading stimuli that required a grapheme-to-phoneme conversion (pseudowords and pseudohomophones). A similar result was reported by De Luca et al. 24 in children when short words and pseudowords were presented. In the present study, this can most likely be attributed to the fact that the stimuli used were short and were high-frequency concrete nouns and the dyslexics tested were university students who had completed several years of remediation.