Saccades are ballistic eye movements which position an object on the fovea for high definition vision,1 and are vital in reading.2 The relationship between saccadic dysfunction and reading disabilities is well documented3–8 and is a common area of optometric therapeutic intervention.7 This knowledge, combined with a need for clinically applicable assessment tools, led to the development of visual-verbal tests, which have been assumed to depend on the quality of ocular motor and rapid automaticity naming (RAN) skills.9,10 Earlier visual-verbal saccade tests (such as the Pierce and New York State Optometric Association King- Devick tests11) did not control for RAN skills, and hence a poor test result could be caused by cognitive or verbal delays.12
The Developmental Eye Movement (DEM) test was designed to control for RAN ability, by incorporating a vertical subtest (Bernell Corp, Mishawaka, IN).10 The three DEM subtests are scored by completion time and errors. Subtests A and B are composed of 40 single digit numbers arranged in two vertical columns and require equally spaced small vertical saccades only. Subtest C is an array of 16 horizontal rows, each with 5 unevenly spaced digits. The requirement for horizontal saccades of varying magnitude lead to an assumption that a higher level of ocular motor co-ordination would be required.10 The DEM test ratio is given by the horizontal score (subtest C time corrected for any omission or addition errors) divided by the vertical score (time for subtests A + B) and is designed to differentiate between poor saccadic function (assumed to give a higher horizontal time and ratio) and poor RAN ability (in which all subtest speeds would be increased, but the ratio would remain normal).12
The DEM test design, and assumptions implicit in the design, has led to its widespread use as an indicator of “saccadic programming,”10 “oculomotor dysfunction,”13 or “horizontal saccadic eye movements.”14,15 However, to date there is no empirical evidence that the DEM test is actually related to eye movements. Hence, this study aims to compare performance on the DEM test with quantitative eye movement recordings, to assess its validity as a saccadic assessment tool.
Given the DEM test’s anecdotal support in optometry and education disciplines, three non-motor correlates of DEM test performance were also investigated: symptomatology, reading ability, and visual processing.
It has been shown that ocular motor dysfunction is correlated with symptoms such as skipping and re-reading lines of text16 and that reading disabilities also correlate with general asthenopic symptoms.17 A previous study found a relationship between ocular motor dysfunction symptoms and the DEM test,13 but did not control for confounding factors such as attention deficit disorder.18 Hence, this study aimed to incorporate a more detailed symptom survey (implementing the validated Convergence Insufficiency Symptom Survey (CISS)19,20) and a larger subject cohort.
Previous studies have shown a link between poor reading and visual-verbal saccade tests,14,21,22 but these have focused on young (5 and 6 years21) or older children (15 years14). As the DEM test was designed for use in children between 6 and 14 years,10 the present investigation compared reading and DEM test performance in children aged 8 to 11 years.
Visual processing was assessed using a rapid serial visual presentation (RSVP) task,23 commonly used in reading research.24–34 This involves presenting text sequentially in a fixed location eliminating the need for saccades to successive targets.33 RSVP provides a measure of reading speed without eye movements,33 thus giving an indication of visual processing and verbalization skills (including automaticity of naming9).
One hundred eighty children, aged 8 to 11 years, were enrolled from two government primary schools during a vision screening program. The schools were matched for location, size (approximately 300 children), and socio-economic influences. All participants spoke English as their primary language. The protocol was approved by the University of Melbourne human research ethics committee, and adhered to the tenets of the Declaration of Helsinki.
All participants had an initial comprehensive vision assessment by a registered optometrist (author LNA). The following exclusion criteria prevented children entering the research: unaided visual acuity worse than or equal to 6/9.5 in either eye, abnormal stereoacuity, any amblyopia or strabismus, parental report of behavioral disorders (e.g., attention deficit hyperactivity disorder), history of ocular disease or surgery or use of medications that may affect the visual system. Twenty-two children were excluded, leaving a study population of 158 children (41 8-year-olds, 37 9-year-olds, 45 10-year-olds, and 35 11-year-olds, 56.3% female). Subjects from the different schools completed different tasks, as the initial field trip focused on saccadic correlates and the second trip on reading and visual processing (see Table 1 for breakdown of subject population) Testing order (of DEM test, eye movement recordings, visual processing, and reading tasks) was randomized using Latin squares.
Eye Movement Recording
Eye movements were recorded with binocular Microguide Series 1000 Eye Movement Spectacles (Microguide, Downers Grove, IL), an infrared limbus tracker.35 System bandwidth is DC-100 Hz with linearity over ±20° and resolution <0.1°. Horizontal eye movements were sampled at 1000 Hz with 12-bit resolution. Due to technical limitations of the eye tracker, vertical saccades could not be recorded. The system was connected to a digital oscilloscope and a Toshiba X200 laptop computer (17” LCD screen, Windows Vista), which simultaneously controlled stimulus presentation and data acquisition under Matlab 2007a (The Math Works, Natick, MA).
Subjects were seated 33 cm from the computer screen (i.e., same working distance as the DEM test) and their head was kept steady using flexible foam cheek rests, which allowed free jaw movement for speaking. Before recording, monocular calibrations were performed over ±15°. The calibration stimulus was shifted for each eye, using the subject’s inter-pupillary distance, to avoid vergence errors. All recordings were binocular, and analysis was completed for the right eye in almost all subjects (n = 154). There were four subjects for whom the right eye data were contaminated and in whom the left eye was analyzed.
Eye movements were recorded during four different tasks (Table 2):
* Reflexive saccades (n = 41): This required saccades to a suddenly appearing target (presented randomly in the central 30° of vision) and represents a lower level of saccadic control, with little cognitive processing required.36
* Self-paced saccades (n = 72): This is a volitional saccade task,37 where subjects refixate repeatedly between two dots as quickly as possible.
* Although reflexive and self-paced saccades do not relate directly to DEM test demands, they provide information about saccadic accuracy, speed, and initiation. Hence, if the DEM test is directly related to dysfunction in saccade parameters, a correlation with these simple eye movement tasks would be expected.
* Reading-style eye movement tasks (numbers, n = 41 and dots, n = 63): This involved two tasks where the targets were placed in the same locations as the DEM numbers (i.e., each slide had three lines of targets, black print on white background, same dimensions as the original DEM test and the subject had to name the central line of targets). Due to inherent restrictions of the Microguide system, it was not possible to track eye movements while children read the commercially available version of the DEM test. Hence, a computer-based version of the DEM test was used, which presented the same numbers on a computer screen (the number task). A pilot study of 24 8- to 11-year-olds showed that this task was tightly correlated with the original DEM test, in both horizontal subtest time (Fig. 1A, r = 0.96, r2 = 0.92, p < 0.0001) and number of errors (Fig. 1B, r = 0.85, r2 = 0.72, p < 0.0001). This is henceforth referred to as the number task and was completed under the same conditions as the full experiment. The second reading-style task used dots in the same locations as the DEM numbers and was designed to invoke similar eye movements but removed the numerical processing demand (i.e., a cognitively simpler task). Both tasks elicit volitional saccades, which represent higher order saccade behavior.37 The hypothesis was that, because these tasks are more similar to the DEM test itself (i.e., same number of targets, spatial arrangement, and physical size), they would be more likely to show correlations with DEM test performance than the reflexive or self-paced tasks.
The main outcome measures (averaged over all saccades) obtained from the eye movement recordings were:
* Primary saccadic gain: The ratio of initial saccade amplitude to target amplitude, such that 1.0 represents a completely accurate saccade, <1.0 is an undershoot, and >1.0 is an overshoot.38
* Peak saccadic velocity: For the reflexive task, this was represented by the asymptotic peak velocity.40 For the remaining static stimuli, which elicited saccades of only a few amplitudes, the peak velocity of an 8° saccade was calculated from an interpolation of the linear regression line.
* Number of corrective saccades: If gain was far from 1.0, subjects often made additional saccades to obtain accurate foveation. The number of these was counted from the eye movement trace.
* Saccadic latency: The time from target presentation to saccade onset (defined as when eye velocity ≥30°/s).38 The latency could only be defined for dynamic visual stimuli and hence was only measured in the reflexive task.
* Inter-saccadic interval (ISI): The time between the onsets of two consecutive saccades, thus including saccade planning, initiation, and execution.39 A pure measure of saccade initiation, with limited cognitive processing, was obtained from the dot and self-paced tasks. The ISI in the number task relates to saccade initiation and execution plus numerical processing and verbalization.
DEM Testing (n = 158)
All subjects completed the DEM test according to published instructions.40
Symptomatology (n = 158)
Symptomatology was quantified using the CISS.41 Although the CISS was originally developed for detection of convergence insufficiency,19,20 it relates to near point function and includes questions that have been linked to ocular motor dysfunction.13 It has 15 questions, with five possible answers: never (0 points), not very often (1), sometimes (2), fairly often (3), and always (4). The maximum score for the CISS is 60, and a fail criterion of ≥16 is recommended.19
Reading Performance (n = 77)
Reading was assessed using the Burt reading test, which is similar to the American Wide Range Achievement Test.42 This oral test consists of 110 words of increasing difficulty which are standardized from 5.3 to 14.3 years, hence giving a raw and age-equivalent score for each subject. Reading data were collected at only one school.
Visual Processing (n = 75)
The implementation of two RSVP tasks (numbers and words) allowed assessment of visual processing and naming speed for both numerical and linguistic tasks, which relate to the DEM test and reading performance, respectively. In each trial, five targets (single digit or age-appropriate three-letter words, subtending 0.7 visual degrees vertically) were presented sequentially on the computer screen. Each was initially presented for 2000 ms. Presentation rate was then increased until the subject failed three consecutive trials (verbally named <80% of the targets in a trial correctly). The minimum presentation time required was 62.5 ms. The RSVP outcome measure was the fastest presentation speed at which the subject could correctly identify four of the five targets.
Eye position data were analyzed using Matlab 2007a (The Math Works, Natick, MA). Each outcome measure was compared with DEM test performance using linear and non-linear regression analysis (GraphPad Prism 5.0) and in all cases the best lines of fit were linear (reported here). Linear regression analysis was chosen due to the claimed dependence of the DEM test to aspects of saccade generation and the simplest form of such a relationship would be linear. Group comparisons were made using the unpaired t-test for parametric data, and Mann–Whitney U test for non-parametric data.
One hundred twenty-nine of 158 (81.65%) subjects passed the DEM test (within one standard deviation of the age-matched normative value), giving a good range of test performance scores (test ratios of between 0.947 and 1.867, Table 1). Mean DEM ratio test scores (±standard deviations) for the age groups were: 8-year-olds 1.34 ± 0.22, 9-year-olds 1.26 ± 0.15, 10-year-olds 1.18 ± 0.15, and 11-year-olds 1.13 ± 0.12. All children in this population were able to satisfactorily complete the test. As mentioned previously, children with confounding factors such as attention-deficit hyperactivity disorder18 or dyslexia43 were excluded, thus the results are from a normal, unselected school-based cohort, similar to that used in the development of the DEM test.10
Eye Movements vs. DEM Test Performance (n = between 41 and 72)
There was no significant correlation between DEM test performance (ratio or horizontal sub-score) and any of the following parameters: saccadic gain and number of corrective saccades (Table 3), peak velocity (Table 4), latency or ISI in the non-numerical tasks (Table 5). There was a significant correlation between the ISI in the number task, which involved numerical processing, and the DEM horizontal sub-test (r2 = 0.456, p < 0.0001, Table 5), showing a relationship with visual processing and verbalization demands.
Performance Factors vs. DEM Test Performance
Symptomatology (CISS) (n = 158)
There was no significant correlation between DEM test performance and CISS score (Table 6). No significant difference was found in the DEM ratio scores obtained by children that passed the CISS (score <16), compared with those who failed (score ≥16) (Mann-Whitney U test, p = 0.125, Fig. 2). After removal of one outlier (confirmed using the extreme studentized deviate test, p < 0.05), there was also no significant correlation with any single CISS question.
Reading Performance (Burt Reading Test) (n = 77)
There were significant correlations between all reading (Burt test raw and age-equivalent scores) and DEM test performance measures (Table 6). The strongest relationship was between the DEM horizontal score and the Burt test raw score (r2 = 0.515, p < 0.0001) (Fig. 3).
Visual Processing Speed (RSVP task) (n = 75)
There were significant correlations between both tasks (numbers and words) and DEM test performance (Table 6, Fig. 4). The strongest relationship was between word RSVP and the DEM horizontal score (r = 0.70, r2 = 0.49, p < 0.0001).
Does DEM Test Performance Relate to Eye Movement Skills?
These results show that DEM test performance has no significant correlations with the following saccadic parameters: gain, peak velocity, number of corrective saccades made, latency or inter-saccadic, interval except when the latter parameter was measured during the reading of numbers spaced as they are in the DEM test.
Gains ≈1 in all saccade tasks, consistent with previous studies of normal children.38,44,45 On average, subjects made less accurate saccades to reflexive targets than to stationary stimuli, possibly due to the decreased task demand; because subjects were not required to verbally respond to the reflexive stimuli, they did not need to make highly accurate saccades. The number of corrective saccades was another measure of saccadic accuracy. In all tasks (reflexive, dot and number) subjects with poor gain made more corrective saccades, but the number of saccades did not correlate with DEM test performance.
The mean reflexive asymptotic peak velocity was 536.2°/s (±81.64), which corresponds well with previous pediatric research (that reported mean velocities of between 521.2°/s and 537.4°/s in children aged 8 to 19 years38). The predicted peak velocity of an 8° saccade was significantly higher in the dot task (328.61°/s) than in the number task (306.82°/s) (unpaired t-test, p = 0.021), suggesting that tasks that require higher levels of cognitive processing result in slower saccadic speeds. However, these were two separate groups of subjects, so no direct comparisons can be made.
This study found a mean reflexive saccadic latency of 190 ms (±31 ms) in children aged 8 to 11, which agrees with previous reports that mean saccadic latency decreases from 439 ms at 3 years to 172 ms at 14 years.46
Mean ISI in the tasks of limited cognitive demand (dot and self-paced) did not correlate with DEM test performance. However, the ISI in the number task (which required numerical visual processing and verbalization) did correlate with the DEM horizontal sub-score (r2 = 0.456, p < 0.0001). These results indicate that the DEM horizontal sub-test is not related to saccade initiation or execution alone (i.e., ISI in dot and self-paced tasks) but is related to a measure of saccade initiation when numerical visual processing and verbalization are required together (i.e., ISI in the number task).
This study provides no evidence that the DEM test ratio or horizontal sub-score relate to any of the saccadic parameters measured (gain, velocity, accuracy, or initiation) as long as visual processing and verbalization are not required. These data indicate that the DEM test is not an appropriate saccadic eye movement assessment tool, but suggest a relationship exists with visual processing and verbalization skills.
Does DEM Test Performance Relate to Performance Skills?
Previous reports have linked the DEM test with reading performance.14 The present study supports this relationship, finding significant correlations between DEM performance and reading ability (Burt reading test) and visual processing speeds (RSVP tasks). However, there were no significant correlations between the DEM test and symptomatology (CISS).
The strongest reading correlation was between the horizontal DEM sub-test and the Burt raw score (r2 = 0.51). Garzia et al10 also found only a moderate correlation between DEM test ratio and reading (r2 = 0.303 in their study, r2 = 0.308 in present study). This suggests that the horizontal sub-test may be more closely related to the visual processing and cognitive demands required for reading than the test ratio or vertical sub-score. This significant correlation suggests that the DEM test (especially the horizontal sub-test) could be used clinically as an indirect form of reading assessment. In other words, a failure on the DEM test could indicate a child who is at risk of academic delay and provide supplementary evidence to support treatment of co-existing conditions (i.e., hyperopia, binocular vision anomalies). However, it is worth noting that performance on the Burt reading test only accounts for around 50% of the variance in DEM test scores.
One of the important factors in DEM test performance appears to be visual processing and naming speed, as assessed in this study by RSVP tasks. It has been acknowledged that this is an important factor in visual-verbal tests,9 but it has previously been thought that the DEM test ratio could cancel out this variable. In other words, the ratio should not relate to factors such as attention, automaticity of naming or visual processing speed, as these should affect both vertical and horizontal subtests equally.47 The results of this study suggest this is not the case. Instead, the visual processing demands of vertical and horizontal number reading appear to differ. This is evident from the linear regression slopes in Fig. 5A (number task) and 5B (word task), which differ significantly. Specifically, equal increases in RSVP processing time lead to less of an increase in the vertical sub-score than in the horizontal sub-score. Processing also accounted for less of the variance in the vertical plane than in the horizontal (approximately 38 vs. 48%).
One possibility is that the vertical sub-score is influenced by the ability to generate sequences of equally spaced volitional vertical saccades (which we could not record and which are controlled by a partly distinct neural pathway from horizontal saccades1). If this rarely called-upon ability influenced the vertical sub-score, this outcome measure might thus be less driven by visual processing time than the horizontal sub-score. Evaluation of horizontal and vertical saccades as well as processing time would be necessary to disambiguate this relationship. This does not imply, however, that the DEM might after all be a tool for ocular motor assessment, as it is specifically intended to evaluate horizontal eye movements. Further investigations should include assessment of vertical eye movements during the DEM task, to establish the role these play in test outcomes.
Contrary to previous research,13 this study showed no significant relationship between DEM test performance and symptomatology. It is likely that this is a result of the different subject populations, with previous researchers using a clinical cohort.13 Another possible factor is the symptom survey used. The CISS was chosen as it is has been shown to be a valid and reliable near point symptom assessment tool,19,20 but it does have limited questions with direct applicability to ocular motor function.
In summary, the results of this DEM test validation study show that the test is significantly related to reading performance and visual processing. The efficacy of the DEM test ratio to cancel out automaticity and cognitive factors must be questioned, as these factors differ between the vertical and the horizontal subtests. Indeed, the present results suggest the DEM subtests may provide more important information than the test ratio.
The DEM test is useful clinically for distinguishing children at risk of reading and academic delay (giving an indication of reading ability, visual processing, and verbalization speed), but the present results demonstrate it is not a valid assessment tool for horizontal saccadic eye movements, which are the predominant eye movements when reading.
The present results indicate that the DEM test is not related to horizontal saccadic accuracy, speed or initiation and so, strictly speaking, should not be referred to as an eye movement test. It was correlated with ISI when the latter was measured on a numeric processing task. This task required the visual processing and verbalization of numerals in the same way that the DEM does. Thus, this saccadic parameter was correlated with the DEM only when saccades were initiated between stimuli which required them to be read, identified, and verbalized. The absence of a significant correlation between DEM and ISI as measured on any other saccade task and the presence of a significant correlation between DEM and RSVP performance suggest that this sole DEM–saccade parameter correlation resulted from the task’s visual processing, not its ocular motor, components. Consequently, clinical protocols should be reassessed, as targeted saccadic therapy48 has to be a questionable management strategy for a DEM test failure in light of present findings. Previous research may now also need to be re-interpreted. For example, a study of migrant children stated that they had normal “eye movement co-ordination” because they performed at an age-appropriate level on the DEM test.49 The results of the present study mean that a broader outcome statement is more appropriate, referring only to DEM test performance and not including claims of ocular motor function.
This study was supported by University of Melbourne financial grants and an Australian Postgraduate Award (APA) for author LNA.
Neville A. McBrien
Department of Optometry and Vision Sciences
The University of Melbourne
Melbourne, Victoria 3010, Australia
1.Leigh RG, Zee DS. The Neurology of Eye Movements, 4th ed. New York: Oxford University Press; 2006.
2.Liversedge SP, Findlay JM. Saccadic eye movements and cognition. Trends Cogn Sci 2000;4:6–14.
3.Stromberg EL. Binocular movements of the eyes in reading. J Gen Psychol 1938;18:349–55.
4.Lefton LA, Lahey BB, Stagg DI. Eye movements in reading disabled and normal children: a study of systems and strategies. J Learn Disabil 1978;11:549–58.
5.Poynter HL, Schor C, Haynes HM, Hirsch J. Oculomotor functions in reading disability. Am J Optom Physiol Opt 1982;59:116–27.
6.Hoffman LG. Incidence of vision difficulties in children with learning disabilities. J Am Optom Assoc 1980;51:447–51.
7.Kulp MT, Schmidt PP. Effect of oculomotor and other visual skills on reading performance: a literature review. Optom Vis Sci 1996;73:283–92.
8.Sherman A. Relating vision disorders to learning disability. J Am Optom Assoc 1973;44:140–1.
9.Richman JE, Walker AJ, Garzia RP. The impact of automatic digit naming ability on a clinical test of eye movement functioning. J Am Optom Assoc 1983;54:617–22.
10.Garzia RP, Richman JE, Nicholson SB, Gaines CS. A new visual-verbal saccade test: the development eye movement test (DEM). J Am Optom Assoc 1990;61:124–35.
11.Oride MK, Marutani JK, Rouse MW, DeLand PN. Reliability study of the Pierce and King-Devick saccade tests. Am J Optom Physiol Opt 1986;63:419–24.
12.Rouse MW, Nestor EM, Parot CJ, Deland PN. A reevaluation of the Developmental Eye Movement (DEM) test’s repeatability. Optom Vis Sci 2004;81:934–8.
13.Tassinari JT, DeLand P. Developmental Eye Movement Test: reliability and symptomatology. Optometry 2005;76:387–99.
14.Powers M, Grisham D, Riles P. Saccadic tracking skills of poor readers in high school. Optometry 2008;79:228–34.
15.Kulp MT. Survey of perceptual testing among binocular vision/pediatrics faculty. J Optom Vis Dev 2001;32:93–8.
16.Rouse MW Optometric assessment of visual efficiency problems. In: Schieman MM, Rouse MW, eds. Optometric Management of Learning- Related Vision Problems. St. Louis: Mosby; 1994:263–97.
17.Grisham JD, Sheppard MM, Tran WU. Visual symptoms and reading performance. Optom Vis Sci 1993;70:384–91.
18.Farrar R, Call M, Maples WC. A comparison of the visual symptoms between ADD/ADHD and normal children. Optometry 2001;72:441–51.
19.Borsting EJ, Rouse MW, Mitchell GL, Scheiman M, Cotter SA, Cooper J, Kulp MT, London R. Validity and reliability of the revised convergence insufficiency symptom survey in children aged 9 to 18 years. Optom Vis Sci 2003;80:832–8.
20.Rouse MW, Borsting EJ, Mitchell GL, Scheiman M, Cotter SA, Cooper J, Kulp MT, London R, Wensveen J. Validity and reliability of the revised convergence insufficiency symptom survey in adults. Ophthalmic Physiol Opt 2004;24:384–90.
21.Kulp MT, Schmidt PP. The relation of clinical saccadic eye movement testing to reading in kindergartners and first graders. Optom Vis Sci 1997;74:37–42.
22.Lowther AH, Rainey BB, Goss DA, Kidd GR, Swartz T, Horner D. The Developmental Eye Movement test as a predictor of word recognition ability. J Optom Vis Dev 2001;32:9–14.
23.Forster KI. Visual perception of rapidly presented word sequences of varying complexity. Percept Psychophys 1970;8:215–21.
24.Legge GE, Mansfield JS, Chung ST. Psychophysics of reading. XX. Linking letter recognition to reading speed in central and peripheral vision. Vision Res 2001;41:725–43.
25.Kwon M, Legge GE, Dubbels BR. Developmental changes in the visual span for reading. Vision Res 2007;47:2889–900.
26.Legge GE, Ahn SJ, Klitz TS, Luebker A. Psychophysics of reading. XVI. The visual span in normal and low vision. Vision Res 1997;37:1999–2010.
27.Legge GE, Cheung SH, Yu D, Chung ST, Lee HW, Owens DP. The case for the visual span as a sensory bottleneck in reading. J Vis 2007;7:9.1–15.
28.Legge GE, Hooven TA, Klitz TS, Stephen Mansfield JS, Tjan BS. Mr. Chips 2002: new insights from an ideal-observer model of reading. Vision Res 2002;42:2219–34.
29.Arditi A, Cho J. Serifs and font legibility. Vision Res 2005;45:2926–33.
30.Gilbert LC. Speed of processing visual stimuli and its relation to reading. J Educ Psychol 1959;50:8–14.
31.Potter MC, Kroll JF, Yachzel B, Carpenter E, Sherman J. Pictures in sentences: understanding without words. J Exp Psychol Gen 1986;115:281–94.
32.Raymond JE, Shapiro KL, Arnell KM. Temporary suppression of visual processing in an RSVP task: an attentional blink? J Exp Psychol Hum Percept Perform 1992;18:849–60.
33.Rubin GS, Turano K. Reading without saccadic eye movements. Vision Res 1992;32:895–902.
34.Ward NJ, Juola JF. Reading with and without eye movements: reply to Just, Carpenter, and Woolley. J Exp Psychol Gen 1982;111:239–41.
35.Kumar A, Krol G. Binocular infrared oculography. Laryngoscope 1992;102:367–78.
36.Pierrot-Deseilligny C, Rivaud S, Gaymard B, Agid Y. Cortical control of reflexive visually-guided saccades. Brain 1991;114(Pt 3):1473–85.
37.Pierrot-Deseilligny C. Cortical control of saccades in man. Acta Neurol Belg 1991;91:63–79.
38.Salman MS, Sharpe JA, Eizenman M, Lillakas L, Westall C, To T, Dennis M, Steinbach MJ. Saccades in children. Vision Res 2006;46:1432–9.
39.Abel LA, Douglas J. Effects of age on latency and error generation in internally mediated saccades. Neurobiol Aging 2007;28:627–37.
40.Richman JE, Garzia RP. Developmental Eye Movement (DEM) Test Manual. South Bend, IN: Bernell Corporation; 1990.
41.Borsting E, Rouse MW, Deland PN, Hovett S, Kimura D, Park M, Stephens B. Association of symptoms and convergence and accommodative insufficiency in school-age children. Optometry 2003;74:25–34.
42.Ng SM. Research into Children’s Language and Reading Development. Singapore: Institute of Education; 1988.
43.Biscaldi M, Fischer B, Aiple F. Saccadic eye movements of dyslexic and normal reading children. Perception 1994;23:45–64.
44.Accardo AP, Pensiero S, Da Pozzo S, Perissutti P. Some characteristics of saccadic eye movements in children of primary school age. Doc Ophthalmol 1992;80:189–99.
45.Lasker AG, Denckla MB, Zee DS. Ocular motor behavior of children with neurofibromatosis 1. J Child Neurol 2003;18:348–55.
46.Irving EL, Steinbach MJ, Lillakas L, Babu RJ, Hutchings N. Horizontal saccade dynamics across the human life span. Invest Ophthalmol Vis Sci 2006;47:2478–84.
47.Kulp MT, Schmidt PP. Relationship between visual skills and performance on saccadic eye movement testing. Optom Vis Sci 1998;75:284–7.
48.Scheiman M, Wick B. Clinical Management of Binocular Vision: Heterophoric, Accommodative and Eye Movement Disorders, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2002.
49.Hatch SW, Pattison D, Richman JE. Eye movement dysfunction vs. language delays in migrant children. J Am Optom Assoc 1994;65:715–18.