By Marshall Raskind, Ph.D.
What's new in the world of research related to children with learning and attention difficulties? In this summary of current peer-reviewed research, Marshall Raskind, Ph.D., shares his expert perspective in practical terms for parents like you.
The use of optical character recognition (OCR) systems combined with speech synthesis (computer-generated speech) has become increasingly accepted as a means of compensating for reading disabilities. These OCR systems, or reading machines, convert printed text to spoken language so the user can hear and see written words. These technologies are now marketed internationally (for example, WYNN®, Kurzweil 3000®), commonly found in assistive technology centers serving individuals with learning disabilities, frequently exhibited at LD conferences, generally considered in assistive technology evaluations for students with LD, and regularly discussed in publications on LD and assistive technology.
As OCR systems continue to gain popularity as a compensatory tool for children with reading difficulties, it is important for parents to know whether scientific studies support their use. Furthermore, parents need to be aware that selecting specific technologies for their children is dependent on the individual child, the task to be performed, and the setting in which it is to be used. Hopefully, this article will shed light on these issues by reviewing research on the use OCR combined with synthetic speech for persons with reading disabilities.
OCR systems are generally desktop computers combined with full-page scanners. Users scan in printed documents (e.g., pages from books) in much the same way a copier is used. The printed text is automatically changed to electronic text that is then read aloud by a built-in speech synthesizer. The text is displayed on the computer monitor while the system reads the words aloud. OCR systems often include features that allow the user to "customize" the system for individual preferences including speech rate, pitch, volume, simultaneous highlighting of spoken text, font size/style, as well as background and text color. These systems may also include additional features such as study, writing, and Internet tools.
Fortunately, the acceptance of OCR as a viable assistive technology for LD is based on both a strong theoretical framework, as well as several studies directly investigating the technology's efficacy in compensating for the reading difficulties experienced by individuals with LD. The idea that converting text to speech may help reading comprehension is suggested from research in reading disabilities. Numerous studies have shown that students with reading disabilities have a particularly difficult time with word recognition, especially phonological processing (associating the sounds of language with letters or letter combinations) (1, 2, 3, 4, 5, 6, 7, 8).
Difficulties with phonological processing may, in turn, affect reading comprehension. So much effort may be spent on decoding individual words that there may be little mental energy left for comprehension (9, 10, 11, 12). Although phonological awareness may be poor, there is also evidence that individuals with reading disabilities often exhibit no apparent difficulties in understanding spoken language (13, 14). Considering that persons with LD have difficulty with decoding print, yet may not have difficulty with oral language, it is not difficult to see how OCR systems that convert printed text to the spoken word might enhance reading comprehension.
Perhaps the strongest support for the idea that OCR combined with speech synthesis can help compensate for reading problems comes from those studies that directly investigated the technology. (It is important to emphasize that these studies were not remedial interventions designed to investigate whether the use of the technology alleviated reading deficits. Rather, these studies were compensatory, aimed at determining whether reading performance was enhanced when using the technology, as compared to unaided reading performance.) For example, in a study of middle school students with dyslexia, Elkind, Cohen & Murray (15) found that most of the children enhanced their reading comprehension scores while using an OCR system. In a subsequent study of adults with reading disabilities, Elkind, Black & Murray (16) showed enhanced performance in reading speed and endurance when using OCR as compared to reading unaided.
Similarly, Higgins & Raskind (17), in a study of postsecondary students with LD, found that severely disabled readers improved reading comprehension scores when using OCR. In addition, they found an inverse correlation between silent reading without assistance and reading with an OCR system, such that, the greater the severity of the reading disability, the more the technology elevated reading comprehension scores. However, there is a flip side to this finding. The technology actually appeared to interfere with the reading comprehension of some individuals with a less severe reading disability. The researchers speculated that the reading of every word aloud by means of the speech synthesizer may actually have interfered with comprehension by overly taxing working memory in those readers whose deficits were not as severe. A similar finding was reported by Elkind, Cohen & Murray (15). This is a very important finding as it emphasizes that a technology that may be very helpful to one person, may be of little benefit or, in fact, impede performance in another.
The most recent research in this area comes from a recently published study conducted by my colleague, Eleanor Higgins, and me (18). We researched a handheld OCR device for persons with reading difficulties which was introduced into the marketplace a few years ago. This device (Quicktionary Reading Pen II®) combines miniaturized OCR with synthetic speech and a liquid crystal display (LCD) in a battery-operated, handheld unit. The device allows the user to scan printed text either a word or line at a time. Scanned words appear on the screen within one to three seconds and are read aloud by a built-in speech synthesizer. Similar to the larger desktop systems, speech rate, volume, and speed may be adjusted.
We were interested in whether reading comprehension scores of children with LD would improve when using the technology as compared to reading without the technology. As previously suggested, OCR and speech synthesis may enable students with reading disabilities to bypass their phonological difficulties by hearing the printed word, and which may in turn enhance text comprehension. In the event that comprehension scores improved, we were also interested in determining whether the interference effect found in the previous studies would be present, since this handheld unit would be used to read aloud only single words rather than connected text.
Results of the study indicated that students with reading disabilities ages 10 to 18 performed significantly better in reading comprehension tasks when using the device as compared to reading without it. Furthermore, unlike previous studies, this research did not indicate an interference effect for the readers with less severe deficits. This result is probably due to the fact that the reading pen user only scanned difficult words on an as-needed basis, unlike the desktop unit user, to whom entire passages of text were read aloud, whether or not the user needed help with every word.
There is no intention to suggest that handheld units are superior to desktop systems, or vice-versa, but, rather, that each technology must be considered relative to the needs of the specific individual, task, and setting. For example, there are individuals who may need almost every word read to them aloud, an operation that is easier to perform on a desktop unit. Or, a person might use a more portable OCR device if he needed to use the technology in multiple settings. Other factors such as cost, compatibility, ease of use, technical support, and reliability should also be considered when selecting any assistive technology.
© 2008 GreatSchools Inc. All rights reserved. Originally created by Schwab Learning, formerly a program of the Charles and Helen Schwab Foundation.
Reviewed February 2010