View in EDS HTML Full Text PDF Full Text

Pushing the Speed of Assistive Technologies for Reading

Saved in:
Bibliographic Details
Title: Pushing the Speed of Assistive Technologies for Reading
Language: English
Authors: Schneps, Matthew H. (ORCID 0000-0002-5105-3909), Chen, Chen, Pomplun, Marc, Wang, Jiahui, Crosby, Anne D., Kent, Kevin
Source: Mind, Brain, and Education. Feb 2019 13(1):14-29.
Availability: Wiley-Blackwell. 350 Main Street, Malden, MA 02148. Tel: 800-835-6770; Tel: 781-388-8598; Fax: 781-388-8232; e-mail: cs-journals@wiley.com; Web site: http://www.wiley.com/WileyCDA
Peer Reviewed: Y
Page Count: 16
Publication Date: 2019
Document Type: Journal Articles
Reports - Research
Education Level: Higher Education
Postsecondary Education
Descriptors: Assistive Technology, Technology Uses in Education, Dyslexia, Reading Rate, Reading Comprehension, College Students, Handheld Devices, Visual Stimuli, Auditory Stimuli, Teaching Methods, Program Effectiveness
DOI: 10.1111/mbe.12180
ISSN: 1751-2271
Abstract: People who are practiced in using text-to-speech can drive listening speeds to surprisingly high limits. Here, we investigate the extent to which people who are otherwise untrained, with and without dyslexia, can increase their reading speed when forcibly accelerated visual or auditory presentations are used in isolation or in tandem. The experiment examined the reading speed and comprehension of 43 college students using three methods enabled by software on a handheld device: forcibly accelerated visual augmentation, auditory text-to-speech, and a combination of the two. We found that both typical and impaired readers attained the highest reading speed using the combined method, controlling for comprehension. Importantly, those with dyslexia using the combined methods reached the equivalent reading speed of typical readers using paper, visual, or auditory methods, with no loss in comprehension. Findings here suggest that in future evolutions--using technologies available today--parallel neurological pathways for language processing can be exploited to optimize reading for those impaired.
Abstractor: As Provided
Entry Date: 2019
Accession Number: EJ1206761
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwGQrZTSwCs0eYgtjqqRMFEVAAAA4jCB3wYJKoZIhvcNAQcGoIHRMIHOAgEAMIHIBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDMHH0nknJLxd4cEq7QIBEICBmqtgObkyrfgiqemPq21d_vd5AON5nL-fFa16DJ4GPfDeUZdr-hKmQ06BqxXFPb6ctBzldflykWvxU4PRnPmme9RnhOpVNjeftzPvppf78jzaRvzUqFM7CxmQsVjS_1T2_dqz__FJjahNuqaP-vknArIJh9ck5ZmyqY4rWzxYkrYcOvgZ-aAXk6JJqxleVVGZbaFwa0WvsOxIbr8=
Text:
  Availability: 1
  Value: <anid>AN0134909981;[309x]01feb.19;2019Feb27.06:22;v2.2.500</anid> <title id="AN0134909981-1">Pushing the Speed of Assistive Technologies for Reading </title> <p>People who are practiced in using text‐to‐speech can drive listening speeds to surprisingly high limits. Here, we investigate the extent to which people who are otherwise untrained, with and without dyslexia, can increase their reading speed when forcibly accelerated visual or auditory presentations are used in isolation or in tandem. The experiment examined the reading speed and comprehension of 43 college students using three methods enabled by software on a handheld device: forcibly accelerated visual augmentation, auditory text‐to‐speech, and a combination of the two. We found that both typical and impaired readers attained the highest reading speed using the combined method, controlling for comprehension. Importantly, those with dyslexia using the combined methods reached the equivalent reading speed of typical readers using paper, visual, or auditory methods, with no loss in comprehension. Findings here suggest that in future evolutions—using technologies available today—parallel neurological pathways for language processing can be exploited to optimize reading for those impaired.</p> <p>Reading is a struggle for many people, whether due to impairment in auditory‐phonological or visual‐orthographic processing (Shaywitz & Shaywitz, [<reflink idref="bib56" id="ref1">56</reflink>]). The focus of dyslexia research has been on reading acquisition and decoding, as this is critically important in those who are first learning to read. It is generally thought that poor experience with reading leads to a spiraling decline in children with dyslexia, and that with early intervention and practice the negative effects of this decline can be reduced. However, as such programs of support begin to take hold, it is expected that the numbers of students with dyslexia entering higher education will continue to rise. And yet little is known about how to support those with dyslexia who continue to have difficulties with reading in higher education, where the volume of material needing to be read is substantial, posing concurrent demands for both speed and comprehension far exceeding those in the primary grades, addressed via therapy. Here, we investigate whether current neurological frameworks for dyslexia can provide insights into the design of assistive technologies that would enable advanced students with reading impairment to read at levels of speed and comprehension that are comparable to their peers who are unimpaired, who read using normal visual methods on paper.</p> <hd id="AN0134909981-2">Theories of Reading Impairment</hd> <p>Impairments in phonological processing are predominantly attributed to be the core mechanism of dyslexia (Olson, Forsberg, Wise, & Rack, [<reflink idref="bib40" id="ref2">40</reflink>]; Vellutino, Fletcher, Snowling, & Scanlon, [<reflink idref="bib68" id="ref3">68</reflink>]). Phonological processing deficits are believed to result in difficulties in phonemic or letter‐sound decoding (Blau, van Atteveldt, Ekkebus, Goebel, & Blomert, [<reflink idref="bib3" id="ref4">3</reflink>]), which in turn affect word identification performance and subsequent reading comprehension (Blachman, [<reflink idref="bib2" id="ref5">2</reflink>]; Snowling, [<reflink idref="bib59" id="ref6">59</reflink>]; Stanovich, [<reflink idref="bib61" id="ref7">61</reflink>]; Vellutino, Scanlon, Small, & Tanzman, [<reflink idref="bib69" id="ref8">69</reflink>]; Vellutino, Scanlon, & Tanzman, [<reflink idref="bib70" id="ref9">70</reflink>]). Convergent reports have shown that people with dyslexia perform below average in phonological and auditory sensitivity tasks (Bradley & Bryant, [<reflink idref="bib8" id="ref10">8</reflink>]; Fletcher et al., [<reflink idref="bib23" id="ref11">23</reflink>]; Katz, [<reflink idref="bib30" id="ref12">30</reflink>]; Thomson & Goswami, 2010). Moreover, poor performance in such tasks at a young age can effectively predict future reading difficulties (Bradley & Bryant, [<reflink idref="bib9" id="ref13">9</reflink>]; Torgesen, Wagner, & Rashotte, [<reflink idref="bib66" id="ref14">66</reflink>]).</p> <p>Increasing numbers of studies suggest that this phonological processing deficit is a secondary impairment of more fundamental auditory parameters. Among these are rapid auditory processing—problems in processing brief, rapidly varying auditory cues (Boets, Wouters, Van Wieringen, & Ghesquiere, [<reflink idref="bib4" id="ref15">4</reflink>]; McArthur & Bishop, [<reflink idref="bib35" id="ref16">35</reflink>]); amplitude of envelope onset—impairment in rise time perception constraining the segmentation of syllables to smaller components (Pasquini, Corriveau, & Goswami, [<reflink idref="bib41" id="ref17">41</reflink>]; Richardson, Thomson, Scott, & Goswami, [<reflink idref="bib47" id="ref18">47</reflink>]; Rocheron, Lorenzi, Füllgrabe, & Dumont, [<reflink idref="bib49" id="ref19">49</reflink>]; Thomson, Fryer, Maltby, & Goswami, [<reflink idref="bib62" id="ref20">62</reflink>]; Thomson & Goswami, [<reflink idref="bib63" id="ref21">63</reflink>]; Thomson & Goswami, [<reflink idref="bib64" id="ref22">64</reflink>]); and temporal sampling—perceptual difficulties with syllables, rhymes, and phonemes due to impaired neuroelectric oscillations that encode incoming information at specific frequencies—delta for prosodic perception and theta for syllabic perception (Goswami, [<reflink idref="bib27" id="ref23">27</reflink>]).</p> <p>Recent research has shown that literacy skill is not only associated with enhancement in auditory activation but also in visual response (Dehaene et al., [<reflink idref="bib17" id="ref24">17</reflink>]). Successful graphemic parsing, such as rapid serial letter scanning, relies on grapho‐phonological and lexico‐semantic processing, which activates the early visual analysis and the visual word form system from the posterior parietal cortex (Facoetti et al., [<reflink idref="bib22" id="ref25">22</reflink>]; Jobard, Crivello, & Tzourio‐Mazoyer, [<reflink idref="bib28" id="ref26">28</reflink>]; McCandliss, Cohen, & Dehaene, [<reflink idref="bib37" id="ref27">37</reflink>]; Vidyasagar & Pammer, [<reflink idref="bib71" id="ref28">71</reflink>]; Warrington & Shallice, [<reflink idref="bib73" id="ref29">73</reflink>]). Visual attention deficits are often reported to be comorbid with deficits in phonological skills (Borsting et al., [<reflink idref="bib5" id="ref30">5</reflink>]; Cestnick, [<reflink idref="bib12" id="ref31">12</reflink>]; Cestnick & Coltheart, [<reflink idref="bib13" id="ref32">13</reflink>]; Eden & Zeffiro, [<reflink idref="bib19" id="ref33">19</reflink>]; Shaywitz & Shaywitz, [<reflink idref="bib57" id="ref34">57</reflink>]; Vellutino et al., [<reflink idref="bib68" id="ref35">68</reflink>]). Indeed, studies have shown that poor readers experience cross‐modal association difficulties (Jones, Branigan, Parra, & Logie, [<reflink idref="bib29" id="ref36">29</reflink>]) and have wider processing speed differences between visual and auditory modalities. The mismatch in processing time results in visual and auditory desynchronization among poor readers compared to impaired readers (Breznitz, [<reflink idref="bib10" id="ref37">10</reflink>]; Sela, [<reflink idref="bib55" id="ref38">55</reflink>]).</p> <p>Synchronization theory suggests that accurate and rapid word decoding builds upon successful content and time integration between auditory, visual, and motor control. Temporal sampling theory also proposes that difficulty in low‐frequency neuroelectric oscillations (theta and delta range 1.5–10 Hz) will result in impaired rhythmic entrainment, a process of synchronization of internal rhythm to an external rhythm (Patel, [<reflink idref="bib42" id="ref39">42</reflink>]; Repp, [<reflink idref="bib45" id="ref40">45</reflink>]), that manifests as impaired auditory–visual integration (Goswami, [<reflink idref="bib27" id="ref41">27</reflink>]). Such theories of auditory–visual integration behind dyslexia gave rise to the question whether people with reading impairments can make effective use of concurrent visual and auditory stimuli in applications of assistive technology.</p> <hd id="AN0134909981-3">Assistive Technologies for Reading</hd> <p>With regard to various theories of reading difficulty, a number of approaches for intervention have been proposed, including training of phonological awareness and rhythmic processing. These have been found to effectively improve word identification and reading performance in people with reading difficulty (Bradley & Bryant, [<reflink idref="bib9" id="ref42">9</reflink>]; Fox & Routh, [<reflink idref="bib24" id="ref43">24</reflink>]; Thomson, Leong, & Goswami, [<reflink idref="bib65" id="ref44">65</reflink>]). However, for various reasons, not all individuals are able to benefit from the available therapeutic approaches, and so alternatives are needed for those in higher education who would be at a disadvantage because of the vast amounts of materials required to be read. From the perspective of the universal design for learning, no one reading method can be thought to be superior to all others, in that reading technology must adapt to the neurology of individual readers (see Rose & Strangman, [<reflink idref="bib50" id="ref45">50</reflink>]). Therefore, it is legitimate to consider the use of accommodating tools to assist those who struggle to read despite the availability of therapies. In such cases, a common method for addressing the auditory‐phonological deficits inherent in reading difficulty is to use text‐to‐speech technology (TTS) (Chiang & Liu, [<reflink idref="bib15" id="ref46">15</reflink>]; Elkind, [<reflink idref="bib20" id="ref47">20</reflink>]; Elkind, Cohen, & Murray, [<reflink idref="bib21" id="ref48">21</reflink>]). Long used in this way as an assistive technology, TTS has in recent years come into common usage, in large part because of its utility in situations where visual displays are impractical, such as in smart phones, automobiles, and other mobile computing devices. Thus, TTS technology has evolved tremendously in recent years, and today high‐quality synthetic voice engines capable of rendering naturalistic speech are widely available.</p> <p>In typical applications, TTS is driven at rates comparable to the natural rate of normal speech, which is typically 130–190 words per minute (wpm) (Reynolds & Givens, [<reflink idref="bib46" id="ref49">46</reflink>]). However, this is at least several factors slower than the speeds typically attained in normal reading. As a result, the use of TTS by struggling readers places these students at a distinct disadvantage for applications important in higher education, where the volume of material is large, time is short, and reading must be done. Compressed natural speech can be accelerated to rates of about 600 wpm without substantial loss in intelligibility (Chodorow, [<reflink idref="bib16" id="ref50">16</reflink>]; Dupoux & Green, [<reflink idref="bib18" id="ref51">18</reflink>]; Mehler et al., [<reflink idref="bib38" id="ref52">38</reflink>]; Sebastián‐Gallés, Dupoux, Costa, & Mehler, [<reflink idref="bib54" id="ref53">54</reflink>]; Vagharchakian, Dehaene‐Lambertz, Pallier, & Dehaene, [<reflink idref="bib67" id="ref54">67</reflink>]), far exceeding typical speeds for normal reading. In principle, it is therefore possible to address the slowness of TTS by compressing its presentation rate, but it remains unknown whether this can be effective for students with reading disabilities, who are subject to processing deficits described by temporal sampling theory.</p> <p>If reading parity cannot be achieved through accelerated TTS, assistive technologies aimed at addressing its visual aspects may be useful. One approach that has been tried is to address inter‐letter crowding, an effect reported to be heightened in dyslexia (Bouma & Legein, [<reflink idref="bib7" id="ref55">7</reflink>]; Spinelli, De Luca, Judica, & Zoccolotti, [<reflink idref="bib60" id="ref56">60</reflink>]). Crowding normally acts to impair letter identification in the periphery when letters are crowded by their neighbors in a word. Increased sensitivity to crowding will reduce reading speed by limiting the number of letters able to be perceived at a glance (Pelli et al., [<reflink idref="bib43" id="ref57">43</reflink>]). This phenomenon was found to have distinct explanatory power in single word reading speed (Bosse, Tainturier, & Valdois, [<reflink idref="bib6" id="ref58">6</reflink>]) and passage reading comprehension (Chen, Schneps, Masyn, & Thomson, [<reflink idref="bib14" id="ref59">14</reflink>]) in dyslexia. Heightened sensitivity to crowding may also slow sentence reading by reducing the parafoveal preview benefit in dyslexia. Normally, readers obtain a benefit from a parafoveal preview of the next word to be fixated in a sentence (Sheridan & Reichle, [<reflink idref="bib58" id="ref60">58</reflink>]). Heightened sensitivity to crowding, which impairs letter recognition in the periphery, may additionally slow reading by hindering this priming effect (Frömer et al., [<reflink idref="bib25" id="ref61">25</reflink>]). Increasing the letter spacing is reported to effectively improve reading in dyslexia (Gori & Facoetti, [<reflink idref="bib26" id="ref62">26</reflink>]; McCandliss, [<reflink idref="bib36" id="ref63">36</reflink>]; Zorzi et al., [<reflink idref="bib74" id="ref64">74</reflink>]), presumably by reducing demands for crowding in the parafovea.</p> <p>While single word reading is important during reading acquisition, at later stages (where students may already have strong vocabulary and large numbers of sight words) other factors can dominate. The rapid reading of sentences places strong demands on visual attention, in order to accurately control the gaze as it is propelled in sequence from one word to the next. It has been reported that people with dyslexia exhibit sluggish attention shifting (Lallier et al., [<reflink idref="bib33" id="ref65">33</reflink>]; Roach & Hogben, [<reflink idref="bib48" id="ref66">48</reflink>]), and this would act to impede the advancement of the gaze during reading. This can be addressed by using format manipulations or text masking to help direct the gaze and thus reduce demands on visual attention. Regarding the former, the use of narrow linewidths (two to three words per line) was shown to produce benefits in high school students with dyslexia (Schneps, Thomson, Chen, Sonnert, & Pomplun, [<reflink idref="bib51" id="ref67">51</reflink>]; Schneps, Thomson, Sonnert, et al., [<reflink idref="bib52" id="ref68">52</reflink>]), to speed reading by dramatically reducing incidence of regressions. Text masking is another visual approach that has been demonstrated to be effective in speed reading and has been used as a therapy to train gaze advancement in people with dyslexia (Breznitz et al., [<reflink idref="bib11" id="ref69">11</reflink>]). Yet another visual approach is the use of rapid serial visual presentation, a technique long used in reading research to eliminate the need for gaze tracking entirely, and thus increase speed reading.</p> <hd id="AN0134909981-4">Use of Combined Visual and Auditory Methods</hd> <p>Given the educational needs of older readers in higher education for both accuracy and speed, and given the myriad phenomena governing speed and comprehension in those reading impaired, we ask whether assistive technologies that concurrently address visual and auditory factors in dyslexia can allow students with reading impairment to achieve parity with those unimpaired. According to temporal sampling theory, the concurrent use of visual and auditory inputs would benefit those with reading impairment, by providing access to alternate input streams for language processing in instances where one input becomes confounded during reading.</p> <p>In this article, we investigate whether the concurrent use of rapidly accelerated TTS and visual presentation, when used as an adjunct to support enhanced visual formatting of text, can improve reading in people with or without dyslexia. In a paradigm that examined reading comprehension and speed, college students read using a small handheld device in each of three conditions using auditory, visual, and combined modalities. When displaying text on the device, we used a text formatting found to benefit those with dyslexia. This includes shortened line widths to reduce demands on attention, and extra spacing between lines and letters to reduce effects of crowding. Furthermore, to control for sluggish attention shifting, word‐by‐word masking is used to force the gaze forward in the text. The speed of the masking is locked to the speed of the TTS. Our experiments directly compare the use of TTS alone, use of accelerated visual presentation alone, and concurrent use of the two methods. Response is compared in a sample of those impaired and unimpaired, examining reading speed and comprehension.</p> <hd id="AN0134909981-5">METHODS</hd> <p></p> <hd id="AN0134909981-6">Participants</hd> <p>This project was reviewed and approved by the University of Massachusetts Boston (UMB) Institutional Review Board (IRB). All participants signed informed consent agreements prior to participation in accordance with procedures for human subject research stipulated by this IRB review, and received compensation for their time in the study.</p> <p>Participants were adults, consisting primarily of college students attending UMB, a public inner‐city institution with a diverse population of students, including many from economically disadvantaged households. An aim in recruitment was to create a sample representing a broad range of reading abilities, including those who may have been impaired by age, dyslexia, or other factors. Given this, the recruitment questionnaire included a forced choice item asking potential participants to rate themselves as <emph>strong or struggling</emph> readers, and the resulting sample was balanced according to their response. After recruitment, people were then reassigned to groups (TR and DYS) based on their response to tests of reading (see below). Participants with known perceptual issues (other than corrective lenses) were excluded, as were those who had histories of neurological impairments other than dyslexia or attention deficit hyperactivity disorder (ADHD). Also excluded were those whose primary language was not English (since age 5).</p> <hd id="AN0134909981-7">Determining Reading Profile</hd> <p>All participants undertook a 1.5‐hr session evaluating their baseline reading ability and neurological profile. The current literacy and phonological awareness profile of each participant was measured prior to the experiment using the test of word reading efficiency (TOWRE; Torgesen et al., [<reflink idref="bib66" id="ref70">66</reflink>]), and three subtests of the Comprehensive Test of Phonological Processing (CTOPP; Wagner, Torgesen, Rashotte, & Firm, [<reflink idref="bib72" id="ref71">72</reflink>]), including rapid letter naming and nonword repetition. If standard scores fell 1.5 standard deviations below the mean on at least two out of five of these tests, we took this to be consistent with a dyslexia diagnosis and assigned the participant to the DYS group. Otherwise, participants were considered typical readers and assigned to the TR group. Using the Adult ADHD Self‐Report Scale (ASRS‐V1.1) Symptom Checklist (Kessler et al., [<reflink idref="bib31" id="ref72">31</reflink>]; Kessler et al., [<reflink idref="bib32" id="ref73">32</reflink>]), 8 of 21 normal readers and 4 of 22 impaired readers were identified as having comorbid ADHD. ADHD was not considered in making group assignments. A comparative visual search task (Pomplun et al., [<reflink idref="bib44" id="ref74">44</reflink>]) and a visual attention span test (Bosse et al., [<reflink idref="bib6" id="ref75">6</reflink>]), consisting of a six‐letter global report, were additionally used to characterize attention as further described below but not used in making group assignments. The observed characteristics of the group are summarized in Table 1.</p> <p>Demographics and Reading Profiles of Participants</p> <p> <ephtml> <table><thead valign="bottom"><tr><th /><th align="center">Normal readers (TR)</th><th align="center">Impaired readers (DYS)</th></tr></thead><tbody valign="top"><tr><td>Gender</td><td align="center">15 females, 6 males</td><td align="center">11 females, 11 males</td></tr><tr><td>Age</td><td align="center">25.50 (9.54)</td><td align="center">26.33 (9.09)</td></tr><tr><td>Wear glasses</td><td align="center">10 Yes, 11 No</td><td align="center">8 Yes, 14 No</td></tr><tr><td>Sight</td><td align="center">1 farsighted, 9 nearsighted, 8 normal vision</td><td align="center">2 farsighted, 6 nearsighted, 10 normal vision</td></tr><tr><td>ADHD</td><td align="center">8 identified as ADHD</td><td align="center">4 identified as ADHD</td></tr><tr><td>Standardized scores</td></tr><tr><td>Elision</td><td>9.28 (2.61)</td><td>8.00 (2.81)</td></tr><tr><td>Rapid digit naming<xref ref-type="fn" rid="mbe12180-note-0002" /></td><td>12.19 (2.13)</td><td>9.00 (2.39)</td></tr><tr><td>Rapid letter naming</td><td>14.00 (16.78)</td><td>7.27 (1.88)</td></tr><tr><td>Phoneme decoding efficiency<xref ref-type="fn" rid="mbe12180-note-0002" /></td><td>107.62 (9.34)</td><td>95.95 (9.63)</td></tr><tr><td>Sight word efficiency<xref ref-type="fn" rid="mbe12180-note-0002" /></td><td>108.52 (6.42)</td><td>93.13 (25.69)</td></tr><tr><td>Memory for digits</td><td>11.33 (2.26)</td><td>10.86 (2.33)</td></tr><tr><td>Visual attention span<xref ref-type="fn" rid="mbe12180-note-0002" /></td><td>3.77 (0.81)</td><td>3.18 (0.54)</td></tr><tr><td>Visual search – accuracy</td><td>0.83 (0.11)</td><td>0.85 (0.08)</td></tr><tr><td>Visual search – response time<xref ref-type="fn" rid="mbe12180-note-0002" /></td><td>4.81 (1.49)</td><td>6.40 (2.09)</td></tr></tbody></table> </ephtml> </p> <ulist> <item>121800001 <emph>Note</emph>. ADHD = attention deficit hyperactivity disorder.</item> <item>121800002 <emph>*p</emph> <.05<emph>. **p</emph> <.01<emph>. ***p</emph> <.001.</item> </ulist> <hd id="AN0134909981-8">Research Design</hd> <p>Baseline reading speeds and comprehension on paper (PAPER) were ascertained during the session used to evaluate their neurological profiles. In all other conditions (VISUAL, AUDIO, and COMBINED) reading speed and comprehension were determined in a session reserved for testing using the device. In this case, a balanced design was used, wherein all participants read using the device in all methods, to control for presentation order and difficulty of materials. The PAPER condition was only used as a baseline condition because procedures here differed significantly from the other three using the device. In all cases, including PAPER, participants read similar test materials to measure reading comprehension and speed.</p> <p>The study was carried out in three sessions totaling 4 hours. Neurological and reading profiles of the participants were evaluated in Session 1, and functionalities related to visual attention were observed in Session 2. The primary research question was investigated in Session 3. In this session, reading speed and comprehension using accelerated methods on a device were observed in each of three reading modalities (VISUAL, AUDIO, and COMBINED) in a design balanced for order and difficulty.</p> <hd id="AN0134909981-9">Material</hd> <p>Each of the passages to be read on paper was placed on a table or held by the participant at the preferred reading distance selected by the participant (generally about 32 cm distance). The reading took place under normal classroom lighting conditions. The reading passages were printed on standard white 8.5 × 11‐inch paper sheets. The text was single‐spaced (4.6 mm per line) and linewidths averaged 13.1 words per line. Each block of 220 words was printed normally using a Georgia 12‐pt serif font, and occupied an area of about 15.5 cm × 9.5 cm on the page. Right‐ragged margins were used, and the text was left justified (see Figure 1). The top margin was 1‐inch, and side margins were 1.25‐inch. The text filled roughly one third of the sheet, leaving a broad margin at the bottom.</p> <p>Graph: Sample passage used to determine baseline reading on PAPER. All text passages used in this study were 220 words in length, adapted from the book "The Immortal Life of Henrietta Lacks" by Rebecca Skloot.</p> <p>Graph: image_n/mbe12180-fig-0001.png</p> <hd id="AN0134909981-10">Adjusting for Observer Characteristics</hd> <p>The speed of each participant's baseline reading on paper was assessed by one of three observers using a stopwatch. Given that it is interesting to compare measurements of reading on paper with those performed using a device (where reading rates were determined by a machine—see below), it was important to understand how the stopwatch measurements by the testers differed from those made by an ideal observer.</p> <p>To estimate this, a post hoc experiment was performed. Here, the three observers used a stopwatch to time proxy participants silent reading on paper while their eye movements were recorded using a video camera. Following procedures used during the experiment, the observers would start the stopwatch, saying the word "start" to alert the participant to turn over the paper and begin reading. Reference to the video indicated that the stopwatch procedure added a reaction time latency of 1.37 s (<emph>SD</emph> = 0.46), 1.88 s (<emph>SD</emph> = 0.53), and 2.01 s (<emph>SD</emph> = 0.52) on average, respectively, for each of the three observers, determined by 48 trials for each. These reaction time latency estimates were then subtracted in the analysis to adjust the stopwatch timings for all participants, so that the corrected timings more closely approximated those that would have been recorded had eye tracking been used during the paper reading. (This correction produced nearly identical results when model estimations using adjusted and unadjusted PAPER speeds were compared—see Table 3.)</p> <p>Text presented on other modalities (VISUAL, AUDIO, and COMBINED) was forcibly accelerated at speeds driven by its text‐to‐speech engine. Forcible acceleration was not applied in the PAPER baseline measurement because this was not possible using traditional presentations of text used for paper.</p> <hd id="AN0134909981-11">Reading on Device</hd> <p>During Session 3, reading comprehension and speed were observed using a handheld device in each of three modalities: VISUAL (reading only), AUDIO (listening only), and COMBINED (both). In all cases, the texts were forcibly accelerated so that the content was presented at fixed speeds driven by settings on the device. In the VISUAL condition, participants read silently using a display paradigm that forcibly erased the words, while in the AUDIO condition participants read by listening to TTS. The COMBINED condition used synchronized presentations of accelerated text and TTS.</p> <p>A "best" reading speed in each condition was subjectively determined by each participant through a process of trial and error. Here, participants were instructed to find the highest device speed that allowed for 95% confidence in comprehension. The "best" comprehensible speed (SPEED) was recorded from the device settings used by the participants to set the TTS engine (see more below). Participants' actual comprehension at this speed was then measured using test passages and multiple‐choice instruments (see below). Comprehension scores and speeds were analyzed in relation to method and dyslexia to investigate the hypothesis that the COMBINED method was superior to all others in both those with and without dyslexia.</p> <p>In all cases, reading was performed on an Apple iPod Touch (fourth generation) running iOS 6.1.6. The device was placed in airplane mode to disable BT and WiFi searches. The screen measured 8.89 cm diagonally. The resolution was 960‐by‐640 pixel at 128 pixels/cm. Screen brightness was set to maximum, corresponding to a screen luminance with a black level of approximately 3.5 cd/m<sups>2</sups> and a white level of 450 cd/m<sups>2</sups>. Ambient lighting was under typical classroom conditions. Participants sat at a table and either held the device in their hand or placed it on a table, at their option, at a comfortable distance for reading (typically about 35 cm).</p> <p>Participants used identical device configurations to read in each of the three conditions (VISUAL, AUDIO, and COMBINED) regardless of the assigned modality. In the COMBINED condition, participants read visually on the device while listening to TTS via headphones. In the AUDIO mode, the device was covered with a piece of paper, and while in the VISUAL mode, participants read without wearing the headphones.</p> <hd id="AN0134909981-12">Visual Display</hd> <p>Text content was presented using a custom build of Voice Dream Reader by Winston Chen (Voice Dream LLC, Arlington, MA). The display format used is shown in Figure 2 in the main text. The margins, font size, and other factors were adjusted to present approximately two to three words per line, as used in prior studies (Schneps, Thomson, Chen, et al., [<reflink idref="bib51" id="ref76">51</reflink>]; Schneps, Thomson, Sonnert, et al., [<reflink idref="bib52" id="ref77">52</reflink>]). The text was black and on a white background. An Avenir New (medium) font was used at an internal setting of 25 pt, corresponding to a character x‐height of 24 pixels. To reduce inter‐letter crowding, kerning was adjusted to increase inter‐letter spacing by 100%, corresponding to a letter‐to‐letter separation of about 37 pixels. To similarly reduce interline crowding, line‐to‐line spacing was set to 103 pixels. The resulting display is illustrated in Figure 2.</p> <p>Graph: Visual display illustrating text formatting used in the VISUAL and COMBINED conditions. Letter and line spacing is increased, and linewidths shortened. During reading, software masked words from left to right at speeds controlled by the text‐to‐speech technology (TTS) engine used in the device. The mask (here suggested by a grayed‐out region) was not visible and the words were erased. The display scrolled to the top when the end of the page was reached.</p> <p>Graph: image_n/mbe12180-fig-0002.png</p> <p>Visual reading (used in the VISUAL and COMBINED conditions) was performed using word‐by‐word masking to forcibly accelerate the rate of reading, following methods similar to Reading Acceleration Program (RAP) used by (Breznitz et al., [<reflink idref="bib11" id="ref78">11</reflink>]). Whereas these prior studies used letter‐by‐letter masking to compensate for word length effects, given that the goal of the present study was to pace reading in synchrony with TTS word‐by‐word masking was used to drive reading speed at a rate appropriate for spoken language. The speed setting used in reading was thus set by the speed setting of the TTS engine (see below), whether or not the TTS that drove the reading speed was audible to the participant. When the masking point reached the start of the last line displayed, the device rapidly refreshed the page to display the last line at the top of the screen, followed by the subsequent text in the passage, and began masking from the top of the screen.</p> <hd id="AN0134909981-13">Text‐to‐Speech</hd> <p>Auditory reading via TTS (used in the AUDIO and COMBINED conditions) was accomplished using a high‐quality U.S. English voice ("Will") made by Acapela Group Babel Technologies SA (Belgium). Given that ordinary reading typically outpaces speech, we chose to synchronize word masking and TTS so that reading forcibly preceded the utterance of a word in TTS in the "combined condition": in other words, each word was spoken only after it was erased from view. The speed of the TTS could be set from 50 wpm to 700 wpm, and there was no apparent pitch‐shifting over this compression range. Audio volume was adjusted by the participants to suit their preference.</p> <p>Empirical measurement showed that the software parameters used to set the TTS tended to overestimate the actual speed of the voice. For example, a speed setting indicated as 700 wpm on the device resulted in an actual speed of TTS (as measured by stopwatch) that was only about 550 wpm. To address this, a correction formula was empirically determined for each of the passages. This was done by recording the device screen using a video camera as the device played each of the passages using TTS. These recordings were analyzed to precisely measure the actual speed of the TTS reading as produced by the device (as opposed to the speed quoted by the device manufacturer). Given that the TTS engine was designed to mimic a natural voice, it was necessary to perform these measurements for each passage used in order to correct for word length effects, wherein passages composed of short words were read more quickly than those containing longer words. Measuring the correction factor at 12 speed settings covering the range used in the experiment (150–700 wpm) it was found that the correction factor was well described by a linear formula. (Passage A: <emph>y</emph> = 0.7748<emph>x</emph> − 5.6210; Passage B: <emph>y</emph> = 0.8325<emph>x</emph> − 6.1226; Passage C: <emph>y</emph> = 0.8511<emph>x</emph> − 1.5079; Passage D: <emph>y</emph> = 0.7548<emph>x</emph> − 10.0590; Passage E: <emph>y</emph> = 0.7904<emph>x</emph> − 8.6128; Passage F: <emph>y</emph> = 0.7957<emph>x</emph> − 1.5724; Passage G: <emph>y</emph> = 0.7269<emph>x</emph> − 1.8417; Passage H: <emph>y</emph> = 0.8179<emph>x</emph> − 4.6659; Passage I: <emph>y</emph> = 0.7477<emph>x</emph> −5.0188; Passage J: <emph>y</emph> = 0.8113<emph>x</emph> − 6.0984; Passage K: <emph>y</emph> = 0.7738<emph>x</emph> − 0.8949; Passage L: <emph>y</emph> = 0.7610<emph>x</emph> − 2.4454; Passage M: <emph>y</emph> = 0.7938<emph>x</emph> − 6.8161; Passage N: <emph>y</emph> = 0.7750<emph>x</emph> − 8.6631; Passage O: <emph>y</emph> = 0.7870<emph>x</emph> − 6.9727; Passage P: <emph>y</emph> = 0.7729<emph>x</emph>+0.4343). The resulting correction factors were then applied to the participant's reading speeds accordingly, so that the TTS reading speeds that were recorded using the device settings could be inter‐compared with reading speeds measured by stopwatch, on paper.</p> <hd id="AN0134909981-14">Procedures</hd> <p></p> <hd id="AN0134909981-15">Balanced Research Design</hd> <p>During recruitment, each participant was asked to categorize themselves as a "struggling reader" or a "strong reader" (2AFC). This subjective category, which was objectively reevaluated during Session 2 of the experiment, was used to guide inclusion to balance the number of struggling and strong readers in the sample. This category was then used to randomly assign participants to one of three experimental groups (A, B, C). A Graeco‐Latin Square (Table 2) was used to balance the study for order of method (VISUAL, AUDIO, COMBINED) and test difficulty (test forms A, B, and C).</p> <p>Research Design</p> <p> <ephtml> <table><thead valign="bottom"><tr><th align="left" /><th>Block 1</th><th>Block 2</th><th>Block 3</th></tr></thead><tbody valign="top"><tr><td>Group A</td><td>Visual—Form A</td><td>Form B</td><td>Combined—Form C</td></tr><tr><td>Group B</td><td>Combined—Form B</td><td>Visual—Form C</td><td>Form A</td></tr><tr><td>Group C</td><td>Form C</td><td>Combined—Form A</td><td>Visual—Form B</td></tr></tbody></table> </ephtml> </p> <hd id="AN0134909981-16">Initial Practice</hd> <p>Participants used the device in the COMBINED mode (at an initial speed setting of 270 wpm) to read a prepared passage that served as an introduction to the protocol. This passage ran for approximately 30 min and explained the participant's role in the research, the research procedures, and additionally provided practice in manipulating the device and the settings in each of the three conditions. This procedure ensured that all participants received the same instructions, minimizing the impact of intervention by test administrators. During this introductory session, participants were encouraged to adjust the speed of the novel device during reading to push their reading speed as high as possible without compromising comprehension. Here they were guided by the suggestion that they achieve a speed that was as high as practical, but still enabled comprehension at roughly 95% accuracy. Few participants had questions once this introduction was completed, and participants were then able to carry out the remainder of the protocol largely on their own, following directions supplied by a live computer survey form that collected responses to questions and tracked their progress. Test administrators monitored the participants throughout to ensure fidelity and maintained a paper log form to track the participants' progress in the protocol.</p> <hd id="AN0134909981-17">Measuring Speed and Comprehension</hd> <p>The following two‐step procedure was carried out for each of the three conditions, so that all participants read using all methods and materials, in the order specified by their group assignment (see Table 2).</p> <p></p> <ulist> <item> <emph>Selecting the speed</emph>. Following the practice, participants used one of the three pre‐assigned reading methods (in the order specified by their group assignment as per Table 2) to read a text file designed to help them subjectively arrive at a "best speed" for that method, defined as the fastest speed they felt enabled them to comprehend with 95% confidence. Participants determined this speed by iteratively increasing or decreasing the speed setting of the TTS engine that drove the text, starting from an initial setting of 180 wpm, through a series of 20 trials. Each trial consisted of a unique 3‐sentence passage from the Skloot text. Following each trial, participants used the method to read a brief instruction that served to remind them of the goals of the exercise, encouraging them to adjust the speed up or down using recommended speed increments that were made iteratively smaller on each trial. This process generally took about 20 min, and throughout this time participants continued adjusting the speed until they judged that the speed could no longer be increased without a significant cost to comprehension. They were then asked to record this "best speed" setting and set their device to this speed for the comprehension test that followed.</item> <p></p> <item> <emph>Measuring reading comprehension</emph>. Comprehension at "best speed" was ascertained using multiple‐choice instruments, similar to those used to determine the reading speed on paper. These consisted of four unique 220‐word passages from the Skloot book. Each passage was read using the assigned method at "best speed." The reading duration was forced by this speed setting that drove the TTS engine, and therefore the reading did not need to be otherwise timed. Immediately following reading of each passage, participants used a survey form on a computer to respond to four multiple‐choice questions pertaining to the content just read. Each question had four possible answers, and responses were recorded via key press on the computer. The same online survey form was used to guide and track the participants' progress in the protocol.</item> </ulist> <p>The protocol ended with the participants responding to a questionnaire about their experiences using each of the methods.</p> <hd id="AN0134909981-18">RESULTS</hd> <p></p> <hd id="AN0134909981-19">Reading on the Device Is Generally Superior to Paper</hd> <p>An initial inspection of the data (Table 3) reveals clear effects of both method and dyslexia on SPEED. However, only minor differences are evident in COMPREHENSHION, as is expected given that participants were instructed to select a reading speed that preserved ∼95% comprehension.</p> <p>Raw Data Aggregated by Method and Group</p> <p> <ephtml> <table><thead valign="bottom"><tr><th /><th align="center">Normal readers (TR)</th><th align="center">Impaired readers (DYS)</th></tr></thead><tbody valign="top"><tr><td>Reading SPEED with each method</td></tr><tr><td>Paper (adjusted for tester latency)<xref ref-type="fn" rid="mbe12180-note-0004" /></td><td>286.00 (104.09)</td><td>203.53 (61.22)</td></tr><tr><td>Paper (unadjusted for tester latency)<xref ref-type="fn" rid="mbe12180-note-0004" /></td><td>274.46 (96.70)</td><td>197.43 (57.82)</td></tr><tr><td>Visual<xref ref-type="fn" rid="mbe12180-note-0004" /></td><td>311.47 (93.25)</td><td>257.74 (94.68)</td></tr><tr><td>Audio<xref ref-type="fn" rid="mbe12180-note-0004" /></td><td>335.25 (84.74)</td><td>290.26 (73.20)</td></tr><tr><td>Combined<xref ref-type="fn" rid="mbe12180-note-0004" /></td><td>363.00 (86.72)</td><td>302.41 (91.03)</td></tr><tr><td>Reading COMPREHENSION with each method</td></tr><tr><td>Paper</td><td>11.10 (2.17)</td><td>11.54 (2.34)</td></tr><tr><td>Visual</td><td>12.14 (1.79)</td><td>12.40 (2.21)</td></tr><tr><td>Audio</td><td>11.14 (2.28)</td><td>11.09 (2.22)</td></tr><tr><td>Combined</td><td>11.14 (2.26)</td><td>10.14 (2.26)</td></tr></tbody></table> </ephtml> </p> <ulist> <item>121800003 <emph>Notes</emph>. Standard deviation is shown in parentheses.</item> <item>121800004 *<emph>p</emph> <.05. **<emph>p</emph> <.01. ***<emph>p</emph> <.001.</item> </ulist> <p>The data was analyzed to investigate the hypothesis that the use of computer‐accelerated methods for reading, using concurrent visual and auditory input streams, would enable text to be processed more effectively and efficiently than other methods alone. The analysis considered the dependent variables of SPEED (the corrected "best speed" in wpm, as described above) and COMPREHENSION (a percentage score from 0 to 1 on the 16‐item reading content test), in terms of reading method (PAPER, VISUAL, AUDIO, and COMBINED) and dyslexia status (TR and DYS) as independent variables.</p> <p>It is noteworthy that the speed and comprehension are likely to be associated (and negatively) with each other. Typically, as one reads more carefully, the comprehension increases and the speed decreases. For this reason, we should not consider speed and comprehension in two separate regression models; instead we used a path model that simultaneously considered speed and outcome as the outcome variables, and reading methods, dyslexia status, and other covariates (e.g., age, gender) as independent variables (in other literature this method is also termed as multivariate regression because the regression model contains more than one outcome variable). The advantage of using a path model is not only that it is concise, but more importantly that it allows for the covariation between the residuals of the outcome variables. Using this approach, we can make the claim that the effect from any independent variable to any one of the outcome variables is controlled for the other outcome variable (e.g., the effect from dyslexia to speed is controlled for comprehension, and the effect from dyslexia to comprehension is controlled for speed). Figure 3 shows the diagram of the path model. The oneheaded arrow can be understood as regression, and the double headed arrow indicates that the residuals of the two variables are allowed to covary.</p> <p>Graph: Path diagram for the analytic model. Note. The VISUAL condition is the reference condition, and therefore it is not shown in the diagram. Here, we did not illustrate all of the controlled variables (e.g., age, gender, elision score, and so on). They are nevertheless included in the model used.</p> <p>Graph: image_n/mbe12180-fig-0003.png</p> <p>The model specification is similar to regression models, except that two multiple regression models are jointly estimated, and the covariance matrix of the predictor variables (Φ) and the covariance matrix of the residuals of outcome variables (Ψ) are modeled simultaneously. The notation of the model is as follows:</p> <p> <ephtml> <math display="block" overflow="scroll" altimg="urn:x-wiley:17512271:media:mbe12180:mbe12180-math-0001" xmlns="http://www.w3.org/1998/Math/MathML"><mrow>speed=β11Audio+β12Combined+β13Dyslexia+β14Baseline+ϵ1comprh=β21Audio+β22Combined+β23Dyslexia+β24Baseline+ϵ2</mrow></math> </ephtml> </p> <p> <ephtml> <math display="block" overflow="scroll" altimg="urn:x-wiley:17512271:media:mbe12180:mbe12180-math-0002" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>Φ</mi><mo>=</mo><mfenced open="[" close="]">ϕaudioΔΔ⋮⋱Δϕau−ba⋯ϕbaseline</mfenced><mspace width="0.25em" /><mtext>and</mtext><mspace width="0.25em" /><mi>Ψ</mi><mo>=</mo><mfenced open="[" close="]">ψspeedΔψsp−comψcompr.</mfenced></mrow></math> </ephtml> </p> <p>Due to space limitations, we did not illustrate all of the independent variables in Figure 3. They are nevertheless included in our path model. The complete list of independent variables includes AUDIO, COMBINED, (VISUAL as the reference dummy variable, therefore not shown in the figure and equation), dyslexia status, baseline reading comprehension (measured on PAPER), age, gender, elision score, visual span, memory for digit, visual search, and rapid letter naming. We chose visual as the reference dummy variable for reading methods, and therefore the coefficients of the path from audio to speed should be understood as the estimated difference in reading speed between audio and visual methods (just like the interpretation of dummy variables in typical regression analysis). We used the Wald test (that the distribution is chi‐squared) in the post‐hoc analysis to compare the difference between audio and combined conditions.</p> <hd id="AN0134909981-20">Speed</hd> <p>Figure 4 shows the SPEED fitted by this model, illustrated assuming a prototypical baseline COMPREHENSION value of 0.69 (a score of 11/16). Table 4 show the result based on the path analysis. The predicted speed using the AUDIO method is only marginally higher than the speed using VISUAL after controlling for all other variables. The predicted speed using the COMBINED method is significantly higher than VISUAL. Yet, the post‐hoc analysis shows the speed for AUDIO and COMBINED are not statistically significant (χ<sups>2</sups> = 1.25, <emph>df</emph> = 1, <emph>p</emph> =.263). There is a significant dyslexia effect, meaning participants with dyslexia read more slowly than typical readers. We tested for an interaction effect between dyslexia status and reading method but did not find any significant interaction effect. Interestingly, the post‐hoc test shows that the SPEED of those in the DYS group who read using the COMBINED method cannot be distinguished from those in the TR group who read on PAPER (χ<sups>2</sups> = 0.51, <emph>df</emph> = 1, <emph>p</emph> =.48), nor from those in the TR group who read using the VISUAL method (χ<sups>2</sups> = 0.64, <emph>df</emph> = 1, <emph>p</emph> =.42). Moreover, participants with dyslexia using the COMBINED method can reach a reading speed that is equivalent to typical readers using AUDIO (χ<sups>2</sups> = 0.63, <emph>df</emph> = 1, <emph>p</emph> =.427).</p> <p>Graph: Model‐predicted speed by method assuming a prototypical baseline comprehension value of 0.69 (gray line indicates typical readers; black line those with dyslexia). The error bars indicate the predicted interval for the model fitted values. Thus, assuming comprehension is the same in all methods, then the COMBINED method produces the highest speed.</p> <p>Graph: image_n/mbe12180-fig-0004.png</p> <p>Result of the Path Model</p> <p> <ephtml> <table><thead valign="bottom"><tr><th>Outcome</th><th>Predictor</th><th align="center">Coefficient</th><th align="center"><italic>SE</italic></th><th align="center"><italic>p</italic> Value</th></tr></thead><tbody valign="top"><tr><td>Speed</td><td>Intercept</td><td>378.51</td><td>65.83</td><td><0.001</td></tr><tr><td><bold>Audio</bold></td><td><bold>28.20</bold></td><td><bold>17.04</bold></td><td><bold>0.097</bold></td></tr><tr><td><bold>Combined</bold></td><td><bold>47.38</bold></td><td><bold>17.15</bold></td><td><bold>0.006</bold></td></tr><tr><td><bold>Dyslexia</bold></td><td><bold>−38.42</bold></td><td><bold>16.99</bold></td><td><bold>0.024</bold></td></tr><tr><td>Baseline<xref ref-type="fn" rid="mbe12180-note-0006" /></td><td>69.97</td><td>52.04</td><td>0.179</td></tr><tr><td>Age</td><td>−2.32</td><td>0.79</td><td>0.003</td></tr><tr><td>Female</td><td>7.85</td><td>16.26</td><td>0.629</td></tr><tr><td>Elision</td><td>−5.76</td><td>3.02</td><td>0.056</td></tr><tr><td>Visual span</td><td>15.59</td><td>12.12</td><td>0.198</td></tr><tr><td>Memory digit</td><td>−5.45</td><td>3.77</td><td>0.148</td></tr><tr><td>Visual search</td><td>−2.65</td><td>4.63</td><td>0.567</td></tr><tr><td>RLN<xref ref-type="fn" rid="mbe12180-note-0007" /></td><td>0.39</td><td>0.61</td><td>0.524</td></tr><tr><td>Comprehension</td><td>Intercept</td><td>6.55</td><td>1.65</td><td><0.001</td></tr><tr><td><bold>Audio</bold></td><td><bold>−1.21</bold></td><td><bold>0.43</bold></td><td><bold>0.005</bold></td></tr><tr><td><bold>Combined</bold></td><td><bold>−1.63</bold></td><td><bold>0.43</bold></td><td><bold><0.001</bold></td></tr><tr><td><bold>Dyslexia</bold></td><td><bold>1.21</bold></td><td><bold>0.43</bold></td><td><bold>0.776</bold></td></tr><tr><td>Baseline<xref ref-type="fn" rid="mbe12180-note-0006" /></td><td>4.86</td><td>1.31</td><td><0.001</td></tr><tr><td>Age</td><td>−0.03</td><td>0.02</td><td>0.114</td></tr><tr><td>Female</td><td>0.29</td><td>0.41</td><td>0.472</td></tr><tr><td>Elision</td><td>0.07</td><td>0.07</td><td>0.302</td></tr><tr><td>Visual span</td><td>0.16</td><td>0.30</td><td>0.598</td></tr><tr><td>Memory digit</td><td>0.16</td><td>0.95</td><td>0.098</td></tr><tr><td>Visual search</td><td>−0.03</td><td>0.12</td><td>0.798</td></tr><tr><td>RLN</td><td>0.01</td><td>0.02</td><td>0.538</td></tr><tr><td>Cov (Spd Cmp)<xref ref-type="fn" rid="mbe12180-note-0008" /></td><td /><td>−16.32</td><td>13.79</td><td>0.237</td></tr></tbody></table> </ephtml> </p> <ulist> <item>121800005 <emph>Note</emph>. Bold rows are key predictors of interest.</item> <item>121800006 Baseline is the adjusted PAPER condition.</item> <item>121800007 RLN stands for rapid letter naming.</item> <item>121800008 Cov stands for the covariation between the residuals of speed and comprehension.</item> </ulist> <p>As is evident in Figure 4, comparisons of the TR and DYS groups reveal a consistent SPEED latency in those with dyslexia across all methods and modalities, with comprehension being controlled to be constant. The speed profiles across all reading methods parallel one another, with the DYS group showing a roughly constant 80 wpm latency compared with the TR group in all conditions.</p> <hd id="AN0134909981-21">Comprehension</hd> <p>The AUDIO and COMBINED methods have lower reading comprehension scores than VISUAL, after controlling for other variables (especially the speed), as shown in Figure 5. The difference is only on average one and a half items; nevertheless, it is statistically significant. There is no difference in comprehension between dyslexic and typical readers, indicating both groups were reading for adequate comprehension (see Table 4).</p> <p>Graph: Model‐predicted comprehension by method assuming a prototypical speed value of 290 words per minute with each of the methods (gray line indicates typical readers; black line those with dyslexia). Thus, assuming the speed is the same in all methods, then the VISUAL method produces the highest comprehension.</p> <p>Graph: image_n/mbe12180-fig-0005.png</p> <p>Importantly, when comprehension is considered, DYS readers using the VISUAL method significantly outperformed those in the TR group using PAPER (χ<sups>2</sups> = 3.96, <emph>df</emph> = 1, <emph>p</emph> =.04). Therefore, if used as an assistive technology, the VISUAL method can serve to equalize reading disparities across groups. The TR group also benefited from use of this method. Those in the TR group using the VISUAL method reached similar levels of reading comprehension to those in the DYS group, and outperformed their own PAPER condition (χ<sups>2</sups> = 3.66, <emph>df</emph> = 1, <emph>p</emph> =.06) and marginally outperformed their own AUDIO condition (χ<sups>2</sups> = 2.84, <emph>df</emph> = 1, <emph>p</emph> =.09), but did not outperform their own score in the COMBINED condition (caution should be taken when comparing between PAPER and device as we discuss in the Limitations section below). Lastly, as discussed in the Material section above, parameter estimates in the PAPER condition were only slightly affected by the adjustment used for testers' latency, and therefore we used the adjusted measure in the final model.</p> <hd id="AN0134909981-22">DISCUSSION</hd> <p></p> <hd id="AN0134909981-23">Rapid Cross‐Modal Error Checking Substitutes for Slow Visual Re‐Inspection</hd> <p>Ordinarily, when reading on paper, if a reader suspects an error in interpretation, a regressive saccade is issued to re‐inspect the word previously read, and this acts to promote comprehension (Schotter, Tran, & Rayner, [<reflink idref="bib53" id="ref79">53</reflink>]). Given the evanescent nature of sound, re‐inspection of words is not possible in the AUDIO method, and there is no analogue to regression in this case. While it is possible to use regression in the VISUAL method, this is difficult because regression requires that words be read ahead of the obscuring mask, and this becomes increasingly difficult as the speed of presentation is increased. Intuitively, we would presume that reading using the VISUAL method would have produced lower comprehension because regression was prohibited. However, in our case, we showed that the VISUAL method yielded a superior comprehension outcome. Prior studies showed that regressions were dramatically reduced in dyslexia, and comprehension improved, when narrowed line formats were used, as is the case in the VISUAL method used here (Schneps, Thomson, Sonnert, et al., [<reflink idref="bib52" id="ref80">52</reflink>]). These findings suggest that regression in dyslexia is not solely due to comprehension difficulty, but is also the result of visual difficulty discerning crowded text in the parafovea, that in turn hinders comprehension. Thus, use of narrowed lines, or masking, constrains attention to improve comprehension (Moores, Tsouknida, & Romani, [<reflink idref="bib39" id="ref81">39</reflink>]; Schneps, Thomson, Chen, et al., [<reflink idref="bib51" id="ref82">51</reflink>]; Schneps, Thomson, Sonnert, et al., [<reflink idref="bib52" id="ref83">52</reflink>]).</p> <p>Using the COMBINED method, however, it is possible for readers to make cross‐modal comparisons of the words perceived in the input stream. Such comparisons can conceivably be used to resolve ambiguities in the content, and do so soon after the words are perceived, but prior to the stage where the meaning of these words is integrated into the sentence.</p> <p>In the context of this explanation, we propose that reading efficiency is increased using the COMBINED methodology via cross‐modal comparison that serves to improve error detection and recognition accuracy. It also serves to synchronize the internal rhythm to an external rhythm (see Figure 6). This process is assumed to occur soon after perception, after the sensory information is stored in a modality‐specific buffer, but prior to engagement of higher language processing leading to sentence integration (Vagharchakian et al., [<reflink idref="bib67" id="ref84">67</reflink>]).</p> <p>Graph: Schematic model to account for reading facilitation in the COMBINED method. After sensory inputs are buffered, cross‐modal comparison of visual and auditory inputs is proposed to occur on the level of the selective retrieval network to improve error detection and word discrimination, and this during speed‐reading while maintaining comprehension.</p> <p>Graph: image_n/mbe12180-fig-0006.png</p> <p>Reference to this model provides a potential explanation for the increasing cascade of reading speeds observed in the TR group, wherein COMBINED is the fastest, AUDIO is next, and VISUAL follows that. All three are faster than the baseline reading speed on PAPER. When speed is pressed to the limits of comprehension, COMBINED is superior to AUDIO because it enables rapid cross‐modal error checking and correction not afforded using TTS alone (and impossible on paper). However, each of these are more effective than VISUAL because TTS reduces the burden of phonological decoding, when reading at speed. Given that it typically takes about 500 ms to re‐inspect a word on PAPER, the COMBINED method can be substantially more efficient than PAPER because error‐checking is performed without interrupting the flow of reading. (However, given that the methodology used in the PAPER condition was different from other conditions, results pertaining to PAPER cannot be reliably compared to the other conditions.) COMBINED incurs a small but significant cost to comprehension, perhaps because here attention is divided among the modalities. Another possible explanation is that, although TTS does an excellent job of simulating speech, pronunciations are not always perfect. Thus, it is possible that unexpected enunciations occasionally disturb the rhythm of cross‐modal synchronization during reading, reducing comprehension in the COMBINED mode. In fact, we speculate that, because accurate cross‐modal synchronization seems crucial for success in comprehension using the COMBINED mode, any factor that disturbs the synchronization will hamper comprehension. When small nuances in comprehension are important, the VISUAL method is superior overall</p> <hd id="AN0134909981-24">Augmented Reading as an Assistive Technology</hd> <p>Importantly, our findings show that people with dyslexia who use computer‐enhanced methods for reading performed at levels comparable to typical readers baseline reading speeds using paper. This suggests an immediate application as an assistive technology able to help level the playing field and eliminate reading disparities between those with and without dyslexia. In this case, the choice of reading mode depends on the goal. If the goal is reading for maximum efficiency, where speed is important, and a balance between speed and comprehension is desired, the COMBINED method is likely best. Our findings indicate that performance in the DYS group using the COMBINED method cannot be distinguished from the baseline reading speed on paper of those in the TR, even after controlling for comprehension. More importantly, performance in the DYS group using the COMBINED method cannot be distinguished from the reading speed of the TR group using the VISUAL or AUDIO methods, as well. Therefore, the COMBINED method can constitute an effective assistive technology for people with reading impairments. On the other hand, if the goal of the assistive technology is to maximize comprehension irrespective of speed, we find that using the VISUAL method is most advantageous for those in the DYS group. Using the VISUAL method for comprehension, those in the DYS group outperformed the baseline reading speed of those in the TR group reading on paper. Therefore, in this case the VISUAL method is preferred.</p> <hd id="AN0134909981-25">Limitations of the Current Study</hd> <p>Findings here indicate that reading on the device was superior to reading on PAPER, contradicting prior studies (e.g., Ackerman & Goldsmith, [<reflink idref="bib1" id="ref85">1</reflink>]; Mangen, Walgermo, & Brønnick, [<reflink idref="bib34" id="ref86">34</reflink>]). However, caution should be taken in interpreting the present study, because while participants were urged to push their reading speed to the maximum when reading using the device, they were not urged to do so when reading on paper. Thus, the PAPER condition only serves to provide a baseline measurement of the participant's default reading speed and comprehension and cannot be reliably compared with the device. Future replications of this study should address this limitation, to ensure the PAPER condition can be directly compared with the others. Future studies should also take into consideration additional factors such as the effects of varying the level of difficulty of the text, the nature of the content, and other factors such as the effects of reading fatigue.</p> <p>The study here uses a small sample to detect an effect that is only moderately strong. Therefore, the experiment is under powered, and at risk of a Type II error that fails to reject the null and miss a detection of the treatment effects (such as between AUDIO and COMBINED). Furthermore, given that the analysis makes three major comparisons in the statistical models, this may raise the possibility of Type I error that mistakenly rejects the null hypothesis. Traditionally this is corrected using a Bonferroni correction, which in this case would drop the <emph>p</emph>‐value for significance from.05 to between.01 and.02. A cautious interpretation of the data would apply a Bonferroni correction and focus on the interpretation of the parameters with <emph>p</emph>‐values smaller than.01. In this case <emph>p</emph>‐values in the range.02 to.05 would be considered marginal. However, in so doing, the risk of Type II error is further increased. Clearly, a larger sample is necessary to resolve this in future studies, and we believe doing so will lead to detections of a decisive treatment effect exhibiting smaller <emph>p</emph>‐values, or other subtle interaction effects missed in this study.</p> <hd id="AN0134909981-26">CONCLUSIONS</hd> <p>The traditional methods used for reading on paper were invented to serve ancient engineering constraints that are fast becoming irrelevant in this age of computers. Given options for reading available today, various combinations of active visual displays, augmented text, and text‐to‐speech are highly efficient, and may be more efficient in some applications than the traditional paper‐based reading methods. Such directions are being encouraged through the universal design for learning, to maximize the capabilities of human neurology (Rose & Strangman, [<reflink idref="bib50" id="ref87">50</reflink>]). This study shows that when accelerated methods are used to push the speed of reading, both those with and without impairments are able to read at speeds that are higher than their baseline reading speeds on paper, without sacrificing comprehension. Furthermore, though text‐to‐speech has long been advocated as an assistive technology for those with reading impairments, our study shows that by taking advantage of cross‐modal capabilities of the brain (to simultaneously combine visual and auditory reading methodologies), the efficiency of reading can be pressed to higher levels compared with other methods. Thus, though reading on paper is still preferred by many, this study suggests that as technology introduces new methods for reading, methods that are more efficient, effective, and inclusive may eventually replace those centuries old and widely in use.</p> <hd id="AN0134909981-27">Acknowledgements</hd> <p>This work was supported by funds made possible through support to University of Massachusetts Boston by Boon Philanthropy (Henry Sinclair Sherrill) and Charles and Ellen LaFollette. In‐kind support was provided by Winston Chen, who provided modifications to Voice Dream Software, and iPods used in the experiment. We thank Arif Altun, Sophia Arnall, Tiziana Ballester, John Khan, and Diheng Zhang for their assistance with these experiments.</p> <ref id="AN0134909981-28"> <title> REFERENCES </title> <blist> <bibl id="bib1" idref="ref85" type="bt">1</bibl> <bibtext> Ackerman, R., & Goldsmith, M. (2011). Metacognitive regulation of text learning: On screen versus on paper. Journal of Experimental Psychology: Applied, 17 (1), 18 – 32.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref5" type="bt">2</bibl> <bibtext> Blachman, B. A. (2000). Phonological awareness. In M. L. Kamil, P. B. Mosenthal, P. D. Pearson & R. Barr (Eds.), Handbook of reading research (Vol. 3, pp. 483 – 502). New York, NY : Longman.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref4" type="bt">3</bibl> <bibtext> Blau, V., van Atteveldt, N., Ekkebus, M., Goebel, R., & Blomert, L. (2009). Reduced neural integration of letters and speech sounds links phonological and reading deficits in adult dyslexia. Current Biology, 19, 503 – 508.</bibtext> </blist> <blist> <bibl id="bib4" idref="ref15" type="bt">4</bibl> <bibtext> Boets, B., Wouters, J., Van Wieringen, A., & Ghesquiere, P. (2007). Auditory processing, speech perception and phonological ability in pre‐school children at high risk for dyslexia: A longitudinal study of the auditory temporal processing theory. Neuropsychologia, 45, 1608 – 1620.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref30" type="bt">5</bibl> <bibtext> Borsting, E., Ridder, W. H., III, Dudeck, K., Kelley, C., Matsui, L., & Motoyama, J. (1996). The presence of a magnocellular defect depends on the type of dyslexia. Vision Research, 36, 1047 – 1053.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref58" type="bt">6</bibl> <bibtext> Bosse, M.‐L., Tainturier, M.‐J., & Valdois, S. (2007). Developmental dyslexia: The visual attention span deficit hypothesis. Cognition, 104, 198 – 230. https://doi.org/10.1016/j.cognition.2006.05.009</bibtext> </blist> <blist> <bibl id="bib7" idref="ref55" type="bt">7</bibl> <bibtext> Bouma, H., & Legein, C. P. (1977). Foveal and parafoveal recognition of letters and words by dyslexics and by average readers. Neuropsychologia, 15 (1), 69 – 80.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref10" type="bt">8</bibl> <bibtext> Bradley, L., & Bryant, P. E. (1978). Difficulties in auditory organisation as a possible cause of reading backwardness. Nature, 271 (5647), 746 – 747.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref13" type="bt">9</bibl> <bibtext> Bradley, L., & Bryant, P. E. (1983). Categorizing sounds and learning to read—A causal connection. Nature, 301 (5899), 419 – 421.</bibtext> </blist> <blist> <bibtext> Breznitz, Z. (2006) Fluency in reading: Synchronization of processes. Mahwah, NJ : Lawrence Erlbaum.</bibtext> </blist> <blist> <bibtext> Breznitz, Z., Shaul, S., Horowitz‐Kraus, T., Sela, I., Nevat, M., & Karni, A. (2013). Enhanced reading by training with imposed time constraint in typical and dyslexic adults. Nature Communications, 4, 1486. https://doi.org/10.1038/ncomms2488</bibtext> </blist> <blist> <bibtext> Cestnick, L. (2001). Cross‐modality temporal processing deficits in developmental phonological dyslexics. Brain and Cognition, 46, 319 – 325.</bibtext> </blist> <blist> <bibtext> Cestnick, L., & Coltheart, M. (1999). The relationship between language‐processing and visual‐processing deficits in developmental dyslexia. Cognition, 71, 231 – 255.</bibtext> </blist> <blist> <bibtext> Chen, C., Schneps, M. H., Masyn, K. E., & Thomson, J. M. (2016). The effects of visual attention span and phonological decoding in reading comprehension in dyslexia: A path analysis. Dyslexia, 22, 322 – 344.</bibtext> </blist> <blist> <bibtext> Chiang, H.‐Y., & Liu, C.‐H. (2011). Evaluation of the benefits of assistive reading software: Perceptions of high school students with learning disabilities. Assistive Technology, 23 (4), 199 – 204.</bibtext> </blist> <blist> <bibtext> Chodorow, M. S. (1977) Time compressed speech and the study of lexical and syntactic processing. In W. E. Cooper & E. C. T. Walker (Eds.), Sentence processing: Psycholinguistic studies presented to Merrill Garrett (pp. 87–111). Hillsdale, NJ: Lawrence Erlbaum.</bibtext> </blist> <blist> <bibtext> Dehaene, S., Pegado, F., Braga, L. W., Ventura, P., Nunes Filho, G., Jobert, A., ... Cohen, L. (2010). How learning to read changes the cortical networks for vision and language. Science, 330 (6009), 1359 – 1364.</bibtext> </blist> <blist> <bibtext> Dupoux, E., & Green, K. (1997). Perceptual adjustment to highly compressed speech: Effects of talker and rate changes. Journal of Experimental Psychology: Human Perception and Performance, 23, 914 – 927. https://doi.org/10.1037/0096‐1523.23.3.914</bibtext> </blist> <blist> <bibtext> Eden, G. F., & Zeffiro, T. A. (1998). Neural systems affected in developmental dyslexia revealed by functional neuroimaging. Neuron, 21, 279 – 282.</bibtext> </blist> <blist> <bibtext> Elkind, J. (1998). Computer reading machines for poor readers. Perspectives, 24 (2), 4 – 6.</bibtext> </blist> <blist> <bibtext> Elkind, J., Cohen, K., & Murray, C. (2008). Using computer‐based readers to improve reading comprehension of students with dyslexia. Annals of Dyslexia, 43 (1), 1 – 22. https://doi.org/10.1007/bf02928184</bibtext> </blist> <blist> <bibtext> Facoetti, A., Trussardi, A. N., Ruffino, M., Lorusso, M. L., Cattaneo, C., Galli, R., ... Zorzi, M. (2010). Multisensory spatial attention deficits are predictive of phonological decoding skills in developmental dyslexia. Journal of Cognitive Neuroscience, 22, 1011 – 1025.</bibtext> </blist> <blist> <bibtext> Fletcher, J. M., Shaywitz, S. E., Shankweiler, D. P., Katz, L., Liberman, I. Y., Stuebing, K. K., ... Shaywitz, B. A. (1994). Cognitive profiles of reading disability: Comparisons of discrepancy and low achievement definitions. Journal of Educational Psychology, 86 (1), 6 – 23.</bibtext> </blist> <blist> <bibtext> Fox, B., & Routh, D. K. (1976). Phonemic analysis and synthesis as word‐attack skills. Journal of Educational Psychology, 68 (1), 70 – 74.</bibtext> </blist> <blist> <bibtext> Frömer, R., Dimigen, O., Niefind, F., Krause, N., Kliegl, R., & Sommer, W. (2015). Are individual differences in reading speed related to extrafoveal visual acuity and crowding? PLoS One, 10 (3), e0121986 – e0121918. https://doi.org/10.1371/journal.pone.0121986</bibtext> </blist> <blist> <bibtext> Gori, S., & Facoetti, A. (2015). How the visual aspects can be crucial in reading acquisition: The intriguing case of crowding and developmental dyslexia. Journal of Vision, 15 (1), 8.</bibtext> </blist> <blist> <bibtext> Goswami, U. (2011). A temporal sampling framework for developmental dyslexia. Trends in Cognitive Sciences, 15 (1), 3 – 10.</bibtext> </blist> <blist> <bibtext> Jobard, G., Crivello, F., & Tzourio‐Mazoyer, N. (2003). Evaluation of the dual route theory of reading: A meta‐analysis of 35 neuroimaging studies. NeuroImage, 20, 693 – 712.</bibtext> </blist> <blist> <bibtext> Jones, M. W., Branigan, H. P., Parra, M. A., & Logie, R. H. (2013). Cross‐modal binding in developmental dyslexia. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 1807 – 1822.</bibtext> </blist> <blist> <bibtext> Katz, R. B. (1986). Phonological deficiencies in children with reading disability: Evidence from an object‐naming task. Cognition, 22, 225 – 257.</bibtext> </blist> <blist> <bibtext> Kessler, R. C., Adler, L., Ames, M., Demler, O., Faraone, S., Hiripi, E., ... Walters, E. E. (2005). The World Health Organization Adult ADHD Self‐Report Scale (ASRS): A short screening scale for use in the general population. Psychological Medicine, 35, 245 – 256.</bibtext> </blist> <blist> <bibtext> Kessler, R. C., Adler, L. A., Gruber, M. J., Sarawate, C. A., Spencer, T., & Van Brunt, D. L. (2007). Validity of the World Health Organization Adult ADHD Self‐Report Scale (ASRS) screener in a representative sample of health plan members. International Journal of Methods in Psychiatric Research, 16 (2), 52 – 65. https://doi.org/10.1002/mpr.208</bibtext> </blist> <blist> <bibtext> Lallier, M., Tainturier, M. J., Dering, B., Donnadieu, S., Valdois, S., & Thierry, G. (2010). Behavioral and ERP evidence for amodal sluggish attentional shifting in developmental dyslexia. Neuropsychologia, 48, 4125 – 4135.</bibtext> </blist> <blist> <bibtext> Mangen, A., Walgermo, B. R., & Brønnick, K. (2013). Reading linear texts on paper versus computer screen: Effects on reading comprehension. International Journal of Educational Research, 58, 61 – 68.</bibtext> </blist> <blist> <bibtext> McArthur, G. M., & Bishop, D. V. M. (2001). Auditory perceptual processing in people with reading and oral language impairments: Current issues and recommendations. Dyslexia, 7 (3), 150 – 170.</bibtext> </blist> <blist> <bibtext> McCandliss, B. D. (2012). Helping dyslexic children attend to letters within visual word forms. Proceedings of the National Academy of Sciences, 109, 11064 – 11065.</bibtext> </blist> <blist> <bibtext> McCandliss, B. D., Cohen, L., & Dehaene, S. (2003). The visual word form area: Expertise for reading in the fusiform gyrus. Trends in Cognitive Sciences, 7 (7), 293 – 299.</bibtext> </blist> <blist> <bibtext> Mehler, J., Sebastian, N., Altmann, G., Dupoux, E., Christophe, A., & Pallier, C. (1993). Understanding compressed sentences: The role of rhythm and meaning. Annals of the New York Academy of Sciences, 682, 272 – 282.</bibtext> </blist> <blist> <bibtext> Moores, E., Tsouknida, E., & Romani, C. (2015). Adults with dyslexia can use cues to orient and constrain attention but have a smaller and weaker attention spotlight. Vision Research, 111, 55 – 65.</bibtext> </blist> <blist> <bibtext> Olson, R., Forsberg, H., Wise, B., & Rack, J. (1994). Measurement of word recognition, orthographic, and phonological skills. In G. R. Lyon (Ed.), Frames of reference for the assessment of learning disabilities: New views on measurement issues (pp. 243–277). Baltimore, MD: Paul H Brookes.</bibtext> </blist> <blist> <bibtext> Pasquini, E. S., Corriveau, K. H., & Goswami, U. (2007). Auditory processing of amplitude envelope rise time in adults diagnosed with developmental dyslexia. Scientific Studies of Reading, 11 (3), 259 – 286.</bibtext> </blist> <blist> <bibtext> Patel, A. D. (2014). Can nonlinguistic musical training change the way the brain processes speech? The expanded OPERA hypothesis. Hearing Research, 308, 98 – 108.</bibtext> </blist> <blist> <bibtext> Pelli, D. G., Tillman, K. A., Freeman, J., Su, M., Berger, T. D., & Majaj, N. J. (2007). Crowding and eccentricity determine reading rate. Journal of Vision, 7 (2), 20.1 – 20.36. https://doi.org/10.1167/7.2.20</bibtext> </blist> <blist> <bibtext> Pomplun, M., Sichelschmidt, L., Wagner, K., Clermont, T., Rickheit, G., & Ritter, H. (2001). Comparative visual search: A difference that makes a difference. Cognitive Science, 25 (1), 3 – 36.</bibtext> </blist> <blist> <bibtext> Repp, B. H. (2005). Sensorimotor synchronization: A review of the tapping literature. Psychonomic Bulletin and Review, 12, 969 – 992.</bibtext> </blist> <blist> <bibtext> Reynolds, M. E., & Givens, J. (2001). Presentation rate in comprehension of natural and synthesized speech. Perceptual and Motor Skills, 92, 958 – 968. https://doi.org/10.2466/pms.2001.92.3c.958</bibtext> </blist> <blist> <bibtext> Richardson, U., Thomson, J. M., Scott, S. K., & Goswami, U. (2004). Auditory processing skills and phonological representation in dyslexic children. Dyslexia, 10 (3), 215 – 233.</bibtext> </blist> <blist> <bibtext> Roach, N. W., & Hogben, J. H. (2007). Impaired filtering of behaviourally irrelevant visual information in dyslexia. Brain, 130, 771 – 785.</bibtext> </blist> <blist> <bibtext> Rocheron, I., Lorenzi, C., Füllgrabe, C., & Dumont, A. (2002). Temporal envelope perception in dyslexic children. Neuroreport, 13, 1683 – 1687.</bibtext> </blist> <blist> <bibtext> Rose, D. H., & Strangman, N. (2007). Universal design for learning: Meeting the challenge of individual learning differences through a neurocognitive perspective. Universal Access in the Information Society, 5 (4), 381 – 391. https://doi.org/10.1007/s10209‐006‐0062‐8</bibtext> </blist> <blist> <bibtext> Schneps, M. H., Thomson, J. M., Chen, C., Sonnert, G., & Pomplun, M. (2013). E‐readers are more effective than paper for some with dyslexia. PLoS One, 8 (9), e75634. https://doi.org/10.1371/journal.pone.0075634.t003</bibtext> </blist> <blist> <bibtext> Schneps, M. H., Thomson, J. M., Sonnert, G., Pomplun, M., Chen, C., & Heffner‐Wong, A. (2013). Shorter lines facilitate reading in those who struggle. PLoS One, 8 (8), e71161. https://doi.org/10.1371/journal.pone.0071161</bibtext> </blist> <blist> <bibtext> Schotter, E. R., Tran, R., & Rayner, K. (2014). Don't believe what you read (only once): Comprehension is supported by regressions during reading. Psychological Science, 25, 1218 – 1226. https://doi.org/10.1177/0956797614531148</bibtext> </blist> <blist> <bibtext> Sebastián‐Gallés, N., Dupoux, E., Costa, A., & Mehler, J. (2000). Adaptation to time‐compressed speech: Phonological determinants. Perception and Psychophysics, 62, 834 – 842.</bibtext> </blist> <blist> <bibtext> Sela, I. (2014). Visual and auditory synchronization deficits among dyslexic readers as compared to non‐impaired readers: A cross‐correlation algorithm analysis. Frontiers in Human Neuroscience, 8, 364.</bibtext> </blist> <blist> <bibtext> Shaywitz, S., & Shaywitz, B. (2005). Dyslexia (specific reading disability). Biological Psychiatry, 57, 1301 – 1309.</bibtext> </blist> <blist> <bibtext> Shaywitz, S. E., & Shaywitz, B. A. (2008). Paying attention to reading: The neurobiology of reading and dyslexia. Development and Psychopathology, 20, 1329 – 1349.</bibtext> </blist> <blist> <bibtext> Sheridan, H., & Reichle, E. D. (2016). An analysis of the time course of lexical processing during reading. Cognitive Science, 40, 522 – 553. https://doi.org/10.1111/cogs.12246</bibtext> </blist> <blist> <bibtext> Snowling, M. J. (2000). Dyslexia. Oxford, UK : Blackwell.</bibtext> </blist> <blist> <bibtext> Spinelli, D., De Luca, M., Judica, A., & Zoccolotti, P. (2002). Crowding effects on word identification in developmental dyslexia. Cortex, 38 (2), 179 – 200.</bibtext> </blist> <blist> <bibtext> Stanovich, K. E. (1991). Word recognition: Changing perspectives. In R. Barr, M. L. Kamil, P. B. Mosenthal, & P. D. Pearson (Eds.), Handbook of reading research (Vol. 2, pp. 418 – 452). Hillsdale, NJ : Lawrence Erlbaum.</bibtext> </blist> <blist> <bibtext> Thomson, J. M., Fryer, B., Maltby, J., & Goswami, U. (2006). Auditory and motor rhythm awareness in adults with dyslexia. Journal of Research in Reading, 29, 334 – 348.</bibtext> </blist> <blist> <bibtext> Thomson, J. M., & Goswami, U. (2008). Rhythmic processing in children with developmental dyslexia: Auditory and motor rhythms link to reading and spelling. Journal of Physiology—Paris, 102 (1–3), 120 – 129.</bibtext> </blist> <blist> <bibtext> Thomson, J. M., & Goswami, U. (2010). Learning novel phonological representations in developmental dyslexia: Associations with basic auditory processing of rise time and phonological awareness. Reading and Writing, 23, 453 – 473.</bibtext> </blist> <blist> <bibtext> Thomson, J. M., Leong, V., & Goswami, U. (2013). Auditory processing interventions and developmental dyslexia: A comparison of phonemic and rhythmic approaches. Reading and Writing, 26, 139 – 161.</bibtext> </blist> <blist> <bibtext> Torgesen, J., Wagner, R., & Rashotte, C. (1999). Test of Word Reading Efficiency (TOWRE). Austin, TX : PRO‐ED.</bibtext> </blist> <blist> <bibtext> Vagharchakian, L., Dehaene‐Lambertz, G., Pallier, C., & Dehaene, S. (Eds.). (2012). A temporal bottleneck in the language comprehension network. Journal of Neuroscience, 32, 9089 – 9102. https://doi.org/10.1523/JNEUROSCI.5685‐11.2012</bibtext> </blist> <blist> <bibtext> Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Specific reading disability (dyslexia): What have we learned in the past four decades? Journal of Child Psychology and Psychiatry, 45, 2 – 40.</bibtext> </blist> <blist> <bibtext> Vellutino, F. R., Scanlon, D. M., Small, S. G., & Tanzman, M. S. (1991). The linguistic basis of reading ability: Converting written to oral language. Text, 11 (1), 97 – 166.</bibtext> </blist> <blist> <bibtext> Vellutino, F. R., Scanlon, D. M., & Tanzman, M. S. (1994). Components of reading ability: Issues and problems in operationalizing word identification, phonological coding, and orthographic coding. In G.R. Lyon (Ed.), Frames of reference for the assessment of learning disabilities: New views on measurement issues (pp. 279–324). Baltimore, MD: Paul H. Brookes.</bibtext> </blist> <blist> <bibtext> Vidyasagar, T. R., & Pammer, K. (2010). Dyslexia: A deficit in visuo‐spatial attention, not in phonological processing. Trends in Cognitive Sciences, 14 (2), 57 – 63.</bibtext> </blist> <blist> <bibtext> Wagner, R. K., Torgesen, J. K., Rashotte, C. A., & Firm, P.‐E. (1999). Comprehensive test of phonological processing: CTOPP. Austin. TX: PRO‐ED.</bibtext> </blist> <blist> <bibtext> Warrington, E. K., & Shallice, T. I. M. (1980). Word‐form dyslexia. Brain, 103 (1), 99 – 112.</bibtext> </blist> <blist> <bibtext> Zorzi, M., Barbiero, C., Facoetti, A., Lonciari, I., Carrozzi, M., Montico, M., ... Ziegler, J. C. (2012). Extra‐large letter spacing improves reading in dyslexia. Proceedings of the National Academy of Sciences, 109, 11455 – 11459.</bibtext> </blist> </ref> <aug> <p>By Matthew H. Schneps; Chen Chen; Marc Pomplun; Jiahui Wang; Anne D. Crosby and Kevin Kent</p> <p>Reported by Author; Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib56" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib40" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib68" firstref="ref3"></nolink> <nolink nlid="nl4" bibid="bib59" firstref="ref6"></nolink> <nolink nlid="nl5" bibid="bib61" firstref="ref7"></nolink> <nolink nlid="nl6" bibid="bib69" firstref="ref8"></nolink> <nolink nlid="nl7" bibid="bib70" firstref="ref9"></nolink> <nolink nlid="nl8" bibid="bib23" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib30" firstref="ref12"></nolink> <nolink nlid="nl10" bibid="bib66" firstref="ref14"></nolink> <nolink nlid="nl11" bibid="bib35" firstref="ref16"></nolink> <nolink nlid="nl12" bibid="bib41" firstref="ref17"></nolink> <nolink nlid="nl13" bibid="bib47" firstref="ref18"></nolink> <nolink nlid="nl14" bibid="bib49" firstref="ref19"></nolink> <nolink nlid="nl15" bibid="bib62" firstref="ref20"></nolink> <nolink nlid="nl16" bibid="bib63" firstref="ref21"></nolink> <nolink nlid="nl17" bibid="bib64" firstref="ref22"></nolink> <nolink nlid="nl18" bibid="bib27" firstref="ref23"></nolink> <nolink nlid="nl19" bibid="bib17" firstref="ref24"></nolink> <nolink nlid="nl20" bibid="bib22" firstref="ref25"></nolink> <nolink nlid="nl21" bibid="bib28" firstref="ref26"></nolink> <nolink nlid="nl22" bibid="bib37" firstref="ref27"></nolink> <nolink nlid="nl23" bibid="bib71" firstref="ref28"></nolink> <nolink nlid="nl24" bibid="bib73" firstref="ref29"></nolink> <nolink nlid="nl25" bibid="bib12" firstref="ref31"></nolink> <nolink nlid="nl26" bibid="bib13" firstref="ref32"></nolink> <nolink nlid="nl27" bibid="bib19" firstref="ref33"></nolink> <nolink nlid="nl28" bibid="bib57" firstref="ref34"></nolink> <nolink nlid="nl29" bibid="bib29" firstref="ref36"></nolink> <nolink nlid="nl30" bibid="bib10" firstref="ref37"></nolink> <nolink nlid="nl31" bibid="bib55" firstref="ref38"></nolink> <nolink nlid="nl32" bibid="bib42" firstref="ref39"></nolink> <nolink nlid="nl33" bibid="bib45" firstref="ref40"></nolink> <nolink nlid="nl34" bibid="bib24" firstref="ref43"></nolink> <nolink nlid="nl35" bibid="bib65" firstref="ref44"></nolink> <nolink nlid="nl36" bibid="bib50" firstref="ref45"></nolink> <nolink nlid="nl37" bibid="bib15" firstref="ref46"></nolink> <nolink nlid="nl38" bibid="bib20" firstref="ref47"></nolink> <nolink nlid="nl39" bibid="bib21" firstref="ref48"></nolink> <nolink nlid="nl40" bibid="bib46" firstref="ref49"></nolink> <nolink nlid="nl41" bibid="bib16" firstref="ref50"></nolink> <nolink nlid="nl42" bibid="bib18" firstref="ref51"></nolink> <nolink nlid="nl43" bibid="bib38" firstref="ref52"></nolink> <nolink nlid="nl44" bibid="bib54" firstref="ref53"></nolink> <nolink nlid="nl45" bibid="bib67" firstref="ref54"></nolink> <nolink nlid="nl46" bibid="bib60" firstref="ref56"></nolink> <nolink nlid="nl47" bibid="bib43" firstref="ref57"></nolink> <nolink nlid="nl48" bibid="bib14" firstref="ref59"></nolink> <nolink nlid="nl49" bibid="bib58" firstref="ref60"></nolink> <nolink nlid="nl50" bibid="bib25" firstref="ref61"></nolink> <nolink nlid="nl51" bibid="bib26" firstref="ref62"></nolink> <nolink nlid="nl52" bibid="bib36" firstref="ref63"></nolink> <nolink nlid="nl53" bibid="bib74" firstref="ref64"></nolink> <nolink nlid="nl54" bibid="bib33" firstref="ref65"></nolink> <nolink nlid="nl55" bibid="bib48" firstref="ref66"></nolink> <nolink nlid="nl56" bibid="bib51" firstref="ref67"></nolink> <nolink nlid="nl57" bibid="bib52" firstref="ref68"></nolink> <nolink nlid="nl58" bibid="bib11" firstref="ref69"></nolink> <nolink nlid="nl59" bibid="bib72" firstref="ref71"></nolink> <nolink nlid="nl60" bibid="bib31" firstref="ref72"></nolink> <nolink nlid="nl61" bibid="bib32" firstref="ref73"></nolink> <nolink nlid="nl62" bibid="bib44" firstref="ref74"></nolink> <nolink nlid="nl63" bibid="bib53" firstref="ref79"></nolink> <nolink nlid="nl64" bibid="bib39" firstref="ref81"></nolink> <nolink nlid="nl65" bibid="bib34" firstref="ref86"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1206761
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Pushing the Speed of Assistive Technologies for Reading
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Schneps%2C+Matthew+H%2E%22">Schneps, Matthew H.</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-5105-3909">0000-0002-5105-3909</externalLink>)<br /><searchLink fieldCode="AR" term="%22Chen%2C+Chen%22">Chen, Chen</searchLink><br /><searchLink fieldCode="AR" term="%22Pomplun%2C+Marc%22">Pomplun, Marc</searchLink><br /><searchLink fieldCode="AR" term="%22Wang%2C+Jiahui%22">Wang, Jiahui</searchLink><br /><searchLink fieldCode="AR" term="%22Crosby%2C+Anne+D%2E%22">Crosby, Anne D.</searchLink><br /><searchLink fieldCode="AR" term="%22Kent%2C+Kevin%22">Kent, Kevin</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Mind%2C+Brain%2C+and+Education%22"><i>Mind, Brain, and Education</i></searchLink>. Feb 2019 13(1):14-29.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Wiley-Blackwell. 350 Main Street, Malden, MA 02148. Tel: 800-835-6770; Tel: 781-388-8598; Fax: 781-388-8232; e-mail: cs-journals@wiley.com; Web site: http://www.wiley.com/WileyCDA
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 16
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2019
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Higher+Education%22">Higher Education</searchLink><br /><searchLink fieldCode="EL" term="%22Postsecondary+Education%22">Postsecondary Education</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Assistive+Technology%22">Assistive Technology</searchLink><br /><searchLink fieldCode="DE" term="%22Technology+Uses+in+Education%22">Technology Uses in Education</searchLink><br /><searchLink fieldCode="DE" term="%22Dyslexia%22">Dyslexia</searchLink><br /><searchLink fieldCode="DE" term="%22Reading+Rate%22">Reading Rate</searchLink><br /><searchLink fieldCode="DE" term="%22Reading+Comprehension%22">Reading Comprehension</searchLink><br /><searchLink fieldCode="DE" term="%22College+Students%22">College Students</searchLink><br /><searchLink fieldCode="DE" term="%22Handheld+Devices%22">Handheld Devices</searchLink><br /><searchLink fieldCode="DE" term="%22Visual+Stimuli%22">Visual Stimuli</searchLink><br /><searchLink fieldCode="DE" term="%22Auditory+Stimuli%22">Auditory Stimuli</searchLink><br /><searchLink fieldCode="DE" term="%22Teaching+Methods%22">Teaching Methods</searchLink><br /><searchLink fieldCode="DE" term="%22Program+Effectiveness%22">Program Effectiveness</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1111/mbe.12180
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1751-2271
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: People who are practiced in using text-to-speech can drive listening speeds to surprisingly high limits. Here, we investigate the extent to which people who are otherwise untrained, with and without dyslexia, can increase their reading speed when forcibly accelerated visual or auditory presentations are used in isolation or in tandem. The experiment examined the reading speed and comprehension of 43 college students using three methods enabled by software on a handheld device: forcibly accelerated visual augmentation, auditory text-to-speech, and a combination of the two. We found that both typical and impaired readers attained the highest reading speed using the combined method, controlling for comprehension. Importantly, those with dyslexia using the combined methods reached the equivalent reading speed of typical readers using paper, visual, or auditory methods, with no loss in comprehension. Findings here suggest that in future evolutions--using technologies available today--parallel neurological pathways for language processing can be exploited to optimize reading for those impaired.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2019
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1206761
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1206761
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1111/mbe.12180
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 16
        StartPage: 14
    Subjects:
      – SubjectFull: Assistive Technology
        Type: general
      – SubjectFull: Technology Uses in Education
        Type: general
      – SubjectFull: Dyslexia
        Type: general
      – SubjectFull: Reading Rate
        Type: general
      – SubjectFull: Reading Comprehension
        Type: general
      – SubjectFull: College Students
        Type: general
      – SubjectFull: Handheld Devices
        Type: general
      – SubjectFull: Visual Stimuli
        Type: general
      – SubjectFull: Auditory Stimuli
        Type: general
      – SubjectFull: Teaching Methods
        Type: general
      – SubjectFull: Program Effectiveness
        Type: general
    Titles:
      – TitleFull: Pushing the Speed of Assistive Technologies for Reading
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Schneps, Matthew H.
      – PersonEntity:
          Name:
            NameFull: Chen, Chen
      – PersonEntity:
          Name:
            NameFull: Pomplun, Marc
      – PersonEntity:
          Name:
            NameFull: Wang, Jiahui
      – PersonEntity:
          Name:
            NameFull: Crosby, Anne D.
      – PersonEntity:
          Name:
            NameFull: Kent, Kevin
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 02
              Type: published
              Y: 2019
          Identifiers:
            – Type: issn-print
              Value: 1751-2271
          Numbering:
            – Type: volume
              Value: 13
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Mind, Brain, and Education
              Type: main
ResultId 1