Using Technology to Individualize Reading Instruction

by David Rose and Bridget Dalton, CAST, Inc.

This book chapter is reprinted with permission of the publisher. Reference:Rose, D. & Dalton, B. (2002). Using Technology to Individualize Reading Instruction. In C.C. Block, L. B. Gambrell & M. Pressley (Eds.), Improving comprehension instruction: Rethinking research, theory, and classroom practice (pp. 257-274). San Francisco: Jossey Bass Publishers.)

Modern technology is radically changing the ways in which we can study human learning and the ways in which we can foster it. In this chapter we wish to examine both of these radical changes as they relate to the future of teaching reading comprehension.

Insights from New Technologies for Studying Learning in the Brain

d

Figure 11.1: Computer images depicting brain activity measured from human subjects under two conditions: hearing words and seeing words. Colors demarcate active brain regions. Warmer colors represent higher levels of activity and cooler colors lower levels of activity. Thus, highly active brain regions appear as "hotspots" on the images.

The last decade of the twentieth century brought an explosion of knowledge in the neurosciences. Much of that explosion is due to new technologies that have revolutionized the way we can study the learning brain. In the past most of our knowledge of the human brain came from postmortem studies of individuals with brain damage. New imaging technologies (such as PET scans and MRI scans) allow us to look into active, living, learning brains without damaging them. Using computers we can view normal and abnormal patterns of activity as colorful images. Figure 11.1 contains computer images showing what we might call hotspots in the brain, regions that are using a lot of energy - burning a lot of glucose -- and so showing up in warmer colors in the images.

Learning Is Distributed in the Brain

What have we learned from these new kinds of images? First, we have learned that processing is distributed: different regions of the brain carry out different aspects of learning (just like numerous specialists carry out different elements of movie production -- "Color by Technicolor" and "Dolby Sound"). Each of these regions is specialized for a particular task, and the overall effect is rather like that of a well-organized committee or work group, where each member's special skills contribute to the group's overall success. The sum is often greater than its parts.

For example, studies have shown that when a person views an image, that person processes various features of the image - its color, shape, orientation - in different regions of the brain. Because each aspect of the image is processed separately but simultaneously by highly specialized regions, we can recognize images quickly and efficiently. The combined effect of these "specialists" working is the creation of something very complex in a short amount of time.

This has some important implications for cognitive constructions - like reading comprehension and language. Cognitive constructions are composed of many component processes that are distributed throughout the brain. Although we tend to lump them together conceptually, the brain does not.

d

Figure 11.2: Computer images depicting average brain activity measured from human subjects performing four different language tasks. Each task is associated with a unique pattern of brain activity.

All four of the images in Figure 11.2 show brain activity during the performance of language tasks, but very different regions of the brain are involved in the different instances, reflecting differences in the specific task . This suggests that the brain has no single language center but instead has many regions that contribute to normal language facility. Different regions are recruited depending on the specific demands of the language task.

What does this say about reading comprehension? First, comprehension is likely to also be a distributed process - that is, it is likely to be made up of many component processes rather than a single process called comprehension. It is also likely comprehending takes different forms and even that different comprehension tasks may elicit different patterns of processing. A study by Nichelli, and colleagues (1995), for example, showed that reading elicits different patterns of activity in the brain depending upon the reading instructions. When the researchers presented the same text (one of Aesop's fables) four times to a single individual, giving different reading instructions each time,, the brain showed different patterns of activity each time the person read text. The response of the brain was not determined by the text but by the reader's purpose in comprehending the text.

Brain studies reveal that reading comprehension, like other cognitive activities, is probably a highly differentiated and distributed activity.

Individual Differences and Learning in the Brain

Brain studies reveal several other important things about how the brain operates during learning. First, brain images show that the distribution of hot spots varies among individuals: somewhat different regions of the brain become active when different individuals perform the same task. Although the overall patterns of activity across individuals exhibit some resemblance, they show consistent and compelling individual differences.

d

Figure 11.3: An image showing differences in brain activity of individuals with and without dyslexia while reading.

Shaywitz, Shaywitz., Pugh, and CAST Research Team (1998) performed one illustrative study of individual differences in reading. They compared brain images of reading collected from individuals with and without dyslexia-type reading disabilities. The brain patterns of these two groups turned out to be very different.

Note that in Figure 11.3 the brain activity of the dyslexic reader is heavily concentrated in the frontal regions of the brain, while the brain activity of the regular reader shows a much more distributed pattern of activity in posterior portions of the brain. Later in the chapter we will examine what is happening in each of these regions. Even without this information for now, these images yield some striking insights. For one, it does not appear (as many mistakenly believe) that dyslexic individuals are "not trying". It seems more likely that they are trying quite hard but that the energy is being expended in very different ways. In this study, (Shaywitz, Shaywitz., Pugh, and CAST Research team members, 1998) and in many others, individual differences are clear.

Experience and Learning in the Brain

Another source of variance in brain images, is related to time and experience. One of the most unremarkable and yet totally surprising findings of recent brain imaging studies is that the brain changes when it learns. The changes that brain imaging makes visible are different than those that are apparent at the anatomical and chemical level. The images show a changing distribution of activity over time.

d

Figure 11.4: Computer images showing experience-dependent changes in brain activity in human subjects. The first two images show brain activity recorded from subjects as they performed a verb generation task (subjects were presented with a noun and asked to generate a verb associated with it) - either for the first time (naïve) or after several iterations (practiced). The third image shows brain activity after a variation of the verb generation task was introduced. Learning is accompanied by a redistribution and reduction in brain activity, and some of these learning effects carry over to a similar, but somewhat novel, task.

The first two images in Figure 11.4 show brain activity recorded from a subject performing a verb generation task (the researcher presented the person with a noun and asked the person to generate a verb associated with it), either for the first time (naïve) or after several iterations (practiced). The third image shows brain activity after the researcher introduced a variation of the verb generation task. Learning is accompanied by a redistribution and deduction in brain activity, and some of these learning effects carry over to a similar but somewhat novel task.

The computer images track a fairly simple language task. The researcher presented the subject with a noun and asked him or her to generate a verb in association with it. As the first image shows, this task elicits a fairly predictable pattern of brain activity- a moderate burn in the temporal region and a very large burn in the frontal region, with smatterings of activity elsewhere. What is striking is the change from the first image to the second. The second image is of the person carrying out exactly the same task after significant learning has occurred through repeated trials. The result of the learning is a huge reduction, or savings, in activity - the brain burns less glucose once it is facile with this task.

Renewed activity is observable in the third image, the result of slight changes in the task that make the individual work again. But the activity is still less than it was to begin with - indicating that there is considerable savings, or transfer, from the first set of trials to the last.

Many studies at the CAST Research Center have repeated this pattern of findings: the brain exhibits tremendous activity at the beginning of a task when it is unskilled Later, as learning occurs, brain activity is re-distributed or reduced. The unskilled brain shows a very different distribution of activity from the skilled one. At least one neuroscientist has described this early pattern as one of scaffolding - the brain recruiting multiple resources to assist in new and unstructured tasks (Petersen, van Mier, Fiez, & Raichle 1998).

To summarize, any complicated activity like reading comprehension requires a highly distributed set of processors in the brain. What parts of the brain an individual actually uses can be quite variable and depend on the specific nature of the task, on the specific nature of the individual, and on what learning has already occurred. (Other factors also make a difference in what areas of the brain we use - including brain injury, sex differences, and so on, but they are beyond the scope of this chapter.)

Thus far we have described only patterns of brain activity -- not their functional significance. The next section briefly probes three broad regions of the brain -which we call recognition, strategic, and affective networks -- contribute to cognitive tasks like reading comprehension.

Reading Comprehension and the Brain Recognition Networks

The back half of the brain's cerebral hemispheres (the posterior lobes) consist of giant networks of neurons that are specialized in one way - they receive input from the various sensory organs (eyes and ears, for example) and use it to construct meaning. These recognition networks make it possible to know that a particular pattern of input represents a cup, a dog, the sound of your grandmother's voice, or the smell of your morning coffee. In reading, these networks perform similarly to allow you to recognize the letter A, the word dog, the words of your grandmother, a paragraph about a cup of coffee, and so forth.

Damage to these posterior networks can impair the brain's ability to recognize things. At the extreme are individuals who can no longer recognize familiar people by their faces or the distinctive sounds of a classical symphony. Many aspects of reading similarly depend on intact pattern recognition: recognizing that the letters c-a-t are a unique pattern that stands for the word cat, recognizing the silent e pattern in spelling, identifying a particular pattern of words as a sonnet or a haiku, recognizing a particular arrangement of words as William Faulkner's style.

Strategic Networks

In the front half of the cerebral hemispheres (the frontal lobes) are networks that are specialized for entirely different functions. These networks are specialized to construct and transmit external patterns of output for making skillful patterns of action and for knowing how to do things - like taking steps, saying words, shooting a foul shot, reading a book, driving a car, planning a vacation, and writing a narrative. All of our skills, strategies, and plans are essentially highly patterned actions.

Damage to the frontal lobes interferes with generating successful plans and actions. At one extreme are individuals who are paralyzed (unable to move voluntarily): at the other are individuals who are able to move easily but are unable to effectively plan and coordinate their various activities - they seem disorganized, impulsive, uncoordinated.

Reading comprehension requires these frontal lobe functions just as much as functions carried out in the posterior recognition cortex. Understanding is more than simply reception or perception; good readers approach text skillfully and strategically. They monitor their own performance by making and testing predictions; they scan important pieces of text for salient information; they identify the overall structure of the text and draw inferences about meaning and motive; they investigate parts of words to hypothesize their meanings; they reread puzzling sentences. All of these reading comprehension skills of comprehension depend upon strategic networks.

Affective Networks

A third set of networks is localized primarily to the central core of the brain. These networks are critical for emotions - fear, desire, sadness, excitement, and hope. These networks specialize not in recognizing or generating patterns -but in determining whether the patterns we perceive matter to us, thus helping us to decide which actions and strategies to pursue.

Impairments in affective networks distort the importance of various aspects of the world, thereby influencing our abilities to establish priorities, select what is valuable, focus attention, and choose actions. The ability to determine accurately the patterns that really count, to differentiate the important from the unimportant, is a third integral component to human intelligence.

Without well-functioning affective networks, the individual's reading comprehension is impaired. Readers are not able to direct and sustain attention to specific aspects of text. Nor are they be able to focus appropriately on specific words or paragraphs that are important or vary their style or rate to accommodate differences in content or purpose, and so forth.

Normally, these three networks operate in concert - as a distributed system. For example, recognition networks function to identify a particular pattern of shapes, colors, and smells as a hamburger. Strategic networks create plans and actions that allow us to walk over, reach out, lift, and munch on that hamburger. And affective networks motivate us, depending on our status in Weight Watchers points, to either approach or avoid that hamburger.

Implications for Understanding Reading Comprehension

Brain research emphasizes the varied and distributed nature of comprehension - comprehension is not a single activity but many processes distributed across different functional networks that operate in parallel. We have emphasized three broad networks. All of these are essential to comprehension, and their impairment can make readers vulnerable in specific ways: students may fail because they have not learned to recognize the relevant patterns in text, because they do not have strategies for constructing meaning from text, or because they do not find reading text important enough to sustain the effort it demands.

Brain research also emphasizes that the fundamental processes underlying reading comprehension differ among individuals in important ways. Individuals do not differ in some generalized or simply quantifiable way (like a global intelligence score) but in many specific ways. Intelligence is the product of processing "committees" or networks, each element of which may be different.

Learning is marked by qualitative changes in the kinds of processors that the brain engages. Early learning activates a spatially and anatomically distinctive array of brain structures, often more expansive than those that are activated by later learning or mastery.

Do these broad generalizations about these three functional networks enlighten us as we think about comprehension? Let us close by looking again at Figure 11.3 which compares dyslexic and normal readers. It is now clear that not only is the dyslexic population displaying a pattern that is differently distributed than regular readers, the dyslexic readers are expending most of their energy within the strategic networks. What does that mean?

Certainly more research is needed to know for sure, but this pattern is similar to the beginner or unskilled student in other research. The dyslexic is reading words with the front part of the brain, effortfully and strategically. Other readers are more automatic, recognizing the patterns in the words easily, with little effort. The allocation of effort to strategically sound out words must come at the cost of allocating that same effort to strategically monitoring comprehension. Thus, like the beginning reader, the dyslexic reader requires a lot of scaffolding -from either internal or external resources - in order to adequately comprehend meaning. That scaffolding can come from skillful use of the brain's internal scaffolding processes, from a willing tutor, or from a peer who is reading collaboratively.

Until now, it has been difficult to find that scaffolding within the text itself. But, as we shall see, that is changing.

The Engaging the Text Project

Using Technology to Transform Text into Supported Reading Environments

In this section we describe how reading instruction is evolving to offer the kinds of scaffolding that readers require to overcome weaknesses in all the three networks we have discussed - each one essential to reading comprehension. We are applying recent advances in the neurosciences and reading comprehension research on strategy instruction (Palincsar & Brown, 1984, Pressley, 2000) to the design of computer-supported reading environments.

Comprehension of a particular text is the result of an interaction, or transaction, between the reader, the text, the purpose for reading, and in school, the instructional context (Lipson & Wixson, 1997). Reading is a complex cognitive process that is socially-based and constructed. It is also a thinking process, and skilled readers actively construct meaning as they read. To succeed with such a multifaceted and challenging task, learners need highly effective instruction. In the Engaging the Text Project (Dalton, Pisha, Eagleton, Coyne, & Deysher, 2001), we have been working with middle school teachers to study the effects of computer-supported strategy instruction on struggling readers' comprehension. Most of the students in the study have been identified as having learning disabilities and are typically reading three to four grade levels below placement. These students struggle to decode and may never develop the automatic word recognition essential to fluent reading and text comprehension (Ehri, 1994). Many also have difficulty reading for meaning, monitoring their comprehension, and taking action when they don't understand (Lipson & Wixson, 1997).

As a result of these struggles, they find that books and other texts that constitute the general curriculum function as barriers rather than gateways for learning. Decoding difficulties block students from access to important content, and comprehension problems block students from responding to and learning from text in meaningful ways. For many students, these difficulties contribute to low self-efficacy and a feeling that applying effort will not pay off in a positive outcome (Guthrie, 2001). Some invest their energy in compensating for their difficulties, while others disengage from literacy and other academic tasks or act out. For these students, the consequences can be severe given the climate of high stakes testing in many states. Already, results show that a significant number of students are not passing basic competency exams (Massachusetts Department of Education, 2000).

How can we make reading comprehension instruction more effective? A wealth of research evidence over the last twenty years strongly supports the teaching of reading comprehension strategies (for recent reviews of this literature, see the National Reading Panel, 2000; Pressley, 2000; and Rosenshine & Meister, 1994; for a review of this literature on students with learning disabilities, see Swanson, 1999). The most commonly used strategies are making predictions, questioning, summarizing and clarifying. Visualization and graphic organizers are also often included in strategy instruction, as well as strategies for self-monitoring and evaluation. The general consensus that "comprehension instruction can effectively motivate and teach readers to learn and to use comprehension strategies that benefit the reader" (National Reading Panel, 2000, p. 4-6) and that multiple strategy instruction carried out in natural classroom settings is more beneficial than the teaching of individual strategies

The Engaging the Text Project uses hypertext Web links to deliver a supported reading comprehension environment that includes interactivity and multimedia. Research on students' comprehension in hypertext is somewhat limited, but the results are promising (for a review, see Kamil, Intrator, & Kim, 2000). Studies with learning disabled students indicate that they benefit from supports such as vocabulary definitions (MacArthur & Haynes, 1995) and anaphoric reference (Boone & Higgins, 1993), but that students do not always have access the supports they need. The work of Anderson-Inman and her colleagues demonstrates that embedding tools and supports in hypertext can improve achievement, if the tools are pedagogically sound and an instructor teaches students to use them (Anderson-Inman et. al, 1996; Anderson-Inman & Horney, 1997; Anderson-Inman & Zeitz, 1993).

The Engaging the Text Project is applying research-based strategy instruction to digital text. It is grounded in Universal Design for Learning theory (UDL) (Meyer & Rose, 1998; 2000; Rose & Meyer, in press;). Universal design originates in the field of architecture. In universally designed architecture, architects design structures to accommodate the widest spectrum of users, including those with disabilities. An important benefit of designing for diversity is that all users tend to benefit from the results. For example, we all take advantage of the ramped curb cut in sidewalks that was originally designed to provide access for individuals in wheelchairs - whether we are pushing a stroller, riding a skateboard, or pulling a luggage cart.

Whereas universal design in architecture is concerned with physical access, UDL focuses on the need for instructional methods and materials that provide students a flexible system of supports for both access and learning. If we think about this in relation to what we know about how the brain learns and the processes involved in comprehending text, it means that we design reading experiences that flexibly support or scaffold students' diverse recognition, strategic, and affective networks.

d

Figure 11.5: Currently only a prototype CD-ROM, The Thinking Reader is a computer-supported reading environment that embeds support for decoding and comprehension strategies in digital text.

Scaffolding is central to this instructional approach and fits well with Vygotsky's (1978) concept of the zone of proximal development. Learning takes place within this zone, where challenge and support are in balance so that the learner is able to achieve success and increase mastery.

We designed a research prototype CD-ROM, the Thinking Reader, which embeds strategy instruction in digitized novels. From a UDL perspective, the Thinking Reader provides supports to scaffold students' diverse recognition, strategic, and affective networks of learning. Figure 11.5 presents a screen shot from the Thinking Reader, and the following classroom scenario illustrates how the tool is being used in the classroom to develop readers who not only read for understanding but are strategic and engaged readers.

Thinking Reader Scenario

Derek, a 6th grade student who reads on the third grade level, is seated at the computer with headphones on, reading a digital version of Hatchet (Paulsen, 1987), an award-winning novel that is required by his school district. He clicks on a read-aloud button to have the text read to him. He encounters an unknown word, "wilderness", and clicks on it to obtain a definition from the glossary. As he continues reading, the Thinking Reader occasionally prompts him to stop and think about the story, and to use one of the strategies he has learned, such as predicting, questioning, clarifying and summarizing.

Summary writing is somewhat difficult for Derek, so he clicks on the strategy hint button. A genie appears and offers one of several hints that are based on a rubric for good summary writing, such as "A good summary captures the gist, or most important information." or "Be sure to include the character and the problem in your summary" . Derek writes his summary in the response box on-screen and sends his work to be posted in his work log. He logs off and joins his class in a brief discussion of the novel. His teacher asks, "Do we need to clarify? Who can give me a summary of what just happened in the story? How do you think Brian is feeling right now? What do you predict will happen next?" As students talk they resolve a confusion about where Brian is flying for the summer and predict that the plane might crash or that something will happen when Brian meets his father.

The following week, Derek and his teacher have a miniconference to review his work log that includes all of his strategy responses. Derek has selected his best example of strategy use and identified a goal that he thinks he should work on for the next few weeks: using more descriptive words in his visualizations. He and his teacher decide he is ready to move to another level of scaffolding, one that provides less structure and will help move him along toward more independent use of strategies while he is reading. Derek's teacher takes note of the fact that Derek and many other students are not clear about the distinction between questioning and clarifying, and she decides to conduct a minilesson the next day, modeling how to use these strategies and guiding students as they practice applying them while reading.

At the end of the year, Derek and his teacher reflect on how he has changed as a reader. Derek is feeling more confident about his abilities as a reader, since the read aloud feature of the software allowed him to read the same novels that his classmates were reading, focusing on understanding, rather than decoding. His growth as a reader is also demonstrated on the end-of-year standardized reading assessment, his willingness to participate in class discussions, and his new interest in adventure stories and the work of author Gary Paulsen. This has been a successful year for Derek. It has also been a successful year for his teacher, who views the Thinking Reader as an important tool for differentiating instruction and addressing the diverse needs of her students. The software does not replace her as a teacher but extends her capacity to reach all of her students and to teach more effectively.

The results of the Engaging the Text Project suggest the promise of developing computer-supported reading environments based on an understanding of recent research in the neurosciences on how the brain learns and the extensive body of research on reading comprehension, strategy instruction and engagement.

Three Recommendations for the Future

As research in the neurosciences continues to reveal the structure of individual differences in learning, we need to continually apply those findings to understanding individual differences in reading comprehension.

As research develops new reading technologies and digital texts that allow us to individualize the teaching of reading comprehension, we need to provide teachers with those technologies.

As classrooms increasingly have access to new technologies for supporting the teaching of reading comprehension, we need to ensure that support and instruction are embedded within those technologies to assist both teachers and students in learning to use them well.

Acknowledgement: We would like to express our appreciation to the teachers and students who worked with us on the Thinking Reader and to the U. S. Department of Education, Office of Special Education Programs, who funded the research project.

References

Anderson-Inman, L. & Horney, M. (1996/97 December/January). Computer-Based Concept Mapping: Enhancing literacy with tools for visual thinking. Journal of Adolescent & Adult Literacy 40 (4), 302-306.

Anderson-Inman, L., Knox-Quinn, C., & Horney, M.A. (1996). Computer-based study strategies for students with learning disabilities: Individual differences associated with adoption level, Journal of Learning Disabilities 29, 461-484.

Anderson-Inman, L. & Zeitz, L. (1993, August/September). Computer-Based Concept Mapping: Active Studying for Active Learners. The Computing Teacher, 6-11.

Boone, R. & Higgins, K. (1993). Hypermedia basal readers: Three years of school-based research, Journal of Special Education Technology, 12, 86-106.

Dalton, B., Pisha, B., Eagleton, M, Coyne, M. & Deysher, S. (2001). Engaging the text: Computer-supported Reciprocal Teaching and Strategy Instruction. Final report to the U.S Office of Special Education Programs. Peabody, MA: Cast, Inc.

Ehri, L. (1994). Development of the ability to read words: Update. In R. Ruddell, M. Ruddell, & H. singer (Eds.), Theoretical models and processes of reading (4th ed., pp. 323-358). Newark, DE: International Reading Association.

Guthrie, J.T. (2001, March). Contexts for engagement and motivation in reading. Reading Online, 4(8). Available: http://www.readingonline.org/articles/art_index.asp?HREF=/articles/handbook/guthrie/index.html

Kamil, M.L., Intrator, S.M., & Kim, H. S. (2000). The effects of other technologies on literacy and literacy learning. In M.L. Kamil, P.B. Mosenthal, P.D. Pearson, & R. Barr (Eds.), Handbook of reading research: Volume III (pp. 771-790). Mahwah NJ: Erlbaum.

Lipson, M.Y. & Wixson, K.K. (1997) Assessment & instruction: An interactive approach. New York, NY, USA, Longman.

MacArthur, C.A., & Haynes, J.B. (1995). Student Assistant for Learning from Text (SALT): A Hypermedia Reading Aid. Journal of Learning Disabilities, 28(3), 150-159.

1999 Massachusetts Comprehensive Assessment System (MCAS) Technical Report (2000) Boston: Massachusetts Department of Education

Meyer, A. & Rose, D., (2000). Universal Design for Individual Differences, Educational Leadership 58(3), 39-43.

Meyer, A., & Rose, D. H. (1998). Learning to Read in the Computer Age. Cambridge, MA: Brookline Books. Also available online at http://www.cast.org

National Reading Panel (2000). Teaching Children to Read: An Evidence-based assessment of the scientfic research literature on reading and its implications for reading instruction. Wash, DC: national Institute of Child Health and Human Development.

Palincsar, A.S., & Brown, A.L. (1984). Reciprocal teaching of comprehension-fostering and comprehension-monitoring activities. Cognition and Instruction, 1, 117-175.

Paulsen, G. (1987). Hatchet. New York: Simon and Schuster.

Pressley, M. (2000). What should comprehension instruction be the instruction of? In M.L. Kamil, P.B. Mosenthal, P.D. Pearson, & R. Barr (Eds.), Handbook of reading research: Volume III (pp. 545-561). Mahwah NJ: Erlbaum.

Rose, D. & Meyer, A.. (in press.) Teaching Every Student in the Digital Age: Universal Design for Learning. Alexandria, VA: Association for Supervisors of Curriculum Development.

Rose, D. & Meyer, A. (2000). Universal Design for Learning. Journal of Special Education Technology, 15 (1), 67-70.

Rosenshine, B., & Meister, C. (1994). Reciprocal teaching: A review of the research. Review of Educational Research, 64(4), 479-530.

Shaywitz, S.E., Shaywitz, B.A., Pugh, K.R., et al. (1998). Functional disruption in the organization of the brain for reading in dyslexia. Proceedings of the National Academy of Sciences of the United States of America, 95, 2636-2641.

Swanson, L. (1999). Reading Research for Students with LD: A Meta-Analysis of Intervention Outcomes. Journal of Learning Disabilities, 32 (6), 504-532.

Vygotsky, L.S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.