Colorado Family Reading Study. Our first attempt to investigate the etiology of a learning disability was the Colorado Family Reading Study, initiated in 1973 by a grant from the Spencer Foundation. The primary objectives of that study were as follows: (1) to construct a short battery of tests that differentiates children with a diagnosed reading disability from matched controls; (2) to assess possible deficits in parents and siblings of affected children; and (3) to study the transmission of reading disability in families of affected children. Children with reading difficulties were ascertained by referral from local school districts and matched to control children on the basis of age, gender, grade, school, and home neighborhood.

Between October 1, 1973, and July 30, 1976, 133 reading-disabled children, their parents and siblings (7.5 to 18 years of age), and members of 125 control families were administered an extensive test battery. On measures of both reading performance and symbol-processing speed, siblings and parents of probands manifested substantial and highly significant deficits (DeFries, Singer, Foch, & Lewitter, 1978). During the next several years these family data were subjected to various genetic analyses including tests for sex-influenced polygenic inheritance and both X-chromosome and autosomal major-gene effects (for reviews, see DeFries & Decker, 1982; DeFries, Vogler, & LaBuda, 1986). Subsequently, results obtained from a complex segregation analysis of these data by Center co-investigators (Pennington, Gilger, Pauls, Smith, Smith, & DeFries, 1991) provided evidence for a dominant major-gene effect with sex-dependent penetrance.

Colorado Reading Project. In order to obtain data from reading-disabled and control children on a broader array of measures, a program project application (Differential Diagnosis in Reading Disability, 5 P01 HD11681) was funded on July 1, 1979. During the 01-03 project periods, a test battery that included measures of cognitive abilities, reading and language processes, and electrophysiological activity was developed and administered to a sample of 140 reading-disabled children and 140 matched controls. Because approximately half of these children were previously tested in the Colorado Family Reading Study, the resulting data set facilitated longitudinal analyses, as well as risk and typological analyses (e.g., DeFries, 1988; DeFries & Baker, 1983; Vogler, DeFries, & Decker, 1985; Vogler, Baker, Decker, DeFries, & Huizinga, 1989).

A twin study of reading disability was initiated during project periods 04-06. During this period, the test battery was administered to 30 identical and 30 fraternal twin pairs in which at least one member of each pair had reading difficulties, to an equal number of control twins, to 50 nontwin reading-disabled and 50 nontwin control children who were tested during project periods 01-03, and to the parents and siblings of these children. Results obtained from analyses of these data were reported by Decker and Vandenberg (1985), DeFries (1985), Olson (1985), and Shucard, Cummings, Gay, Lairsmith, and Welanko (1985) at an NICHD conference on "Biobehavioral Measures of Dyslexia."

During project periods 07-09, we proposed to enlarge the sample to include additional twin pairs and members of nuclear families in which one or more children were reading disabled. However, during this period we developed a new multiple regression analysis of twin data that is highly versatile and statistically more powerful than alternative methods (DeFries & Fulker, 1985; 1988). Thus, additional resources were devoted to the recruitment and testing of twins during subsequent project periods.

Because of its special relevance to the specific aims of the program project, a new component (Linkage Analysis) was added during budget periods 07-09. During budget periods 07-12, Drs. Shelley Smith and Bruce Pennington selected program project families who manifested apparent autosomal dominance for reading disability and then attempted to replicate their previous finding of linkage to chromosome 15 (Smith, Kimberling, Pennington, & Lubs, 1983) using genetic markers that included DNA restriction fragment length polymorphisms. In addition, a new component directed by Dr. Pennington (Epidemiology of Immunological Differences) was added during budget periods 10-12.

Because our application for the Colorado Learning Disabilities Research Center was funded for five years beginning September 30, 1990, we did not apply for competitive renewal of the program project. A final report that summarized the progress obtained in the four research components of the program project grant (Psychometric Assessment/Twin-Family Study, J. C. DeFries; Reading and Language Processes, R. K. Olson; Linkage Analysis, S. D. Smith, W. J. Kimberling, P. S. Ing, and B. F. Pennington; and Epidemiology of Immunological Differences, B. F. Pennington) during project periods 10-12, including a one-year, no cost extension, was submitted to NICHD on April 30, 1993. As indicated below, applications for competitive renewal of the Colorado Learning Disabilities Research Center were subsequently funded for the budget periods of 04/01/96 - 11/30/2000 and 12/01/2000 – 11/30/2005.

Learning Disabilities Research Center. A major goal of the co-investigators of the Colorado Learning Disabilities Research Center is to assess the genetic and environmental etiologies of reading disabilities, ADHD and their comorbidity, as well as their covariation with measures of reading, language and perceptual processes, executive functions, and other psychopathology. To accomplish this goal, staff of an Administrative Core Unit ascertain and schedule twin pairs and siblings that are tested in individual research projects. During budget periods 01-05, test batteries that included measures of cognitive abilities (Research Project I, DeFries and Fulker), reading and language processes (Research Project II, Olson), and ADHD and executive functions (Research Project III, Pennington) were administered to 100 pairs of identical twins and 100 pairs of fraternal twins (both same-sex and opposite-sex) in which at least one member of each pair had a substantial deficit in reading and/or mathematics, to a comparison group of 50 pairs of identical twins and 50 pairs of fraternal twins with a school history of no significant deficit in either reading or mathematics, and to the siblings of all twins. During budget periods 6-10, these test batteries were administered to a sample of 50 pairs of identical twins and 50 pairs of fraternal twins in which at least one member of each pair has reading deficits, to an independent sample of 50 pairs of identical twins and 50 pairs of fraternal twins in which at least one member of each pair manifests symptoms of ADHD, and to a comparison sample of 50 pairs of identical twins and 50 pairs of fraternal twins with no school history of learning disabilities or ADHD. In order to map quantitative trait loci that may influence learning disabilities, blood samples or buccal swabs were obtained from families of all twin pairs and analyzed by staff of Research Project IV (Smith).

A new computer-assisted intervention project was added to the Center during the 04/01/96 - 11/30/2000 budget period to assess possible subtype-by-treatment interactions and to test the general efficacy of different computer-assisted intervention methods (previous Research Project V, Olson and Wise). The methods and results of this project are briefly reviewed here to clarify why the project was not continued during the 12/01/2000 - 11/30/2005 budget period, and why a new version of the intervention program is proposed for the next budget period to study response to computer-assisted instruction in the schools for young children at risk for reading disability, and in the homes of twins with identified reading disabilities in grades 2-4.

Children who were in the lower 10% of their reading class in grades 2-5 were administered computer-assisted instruction in the Boulder schools (Olson, Wise, Ring, & Johnson, 1997; Wise, Ring, Sessions, & Olson, 1997; Wise, Ring, & Olson, 1999; Wise, Ring, & Olson, 2000). In the largest study, Wise et al. (2000) randomly assigned 200 2nd-5th grade children with reading disabilities to one of two training conditions for 50-60 half hour training sessions during the spring semester. The first condition combined small group- and computer-based exercises in phonological awareness and decoding plus accurate reading of interesting stories on the computer with speech support for difficult words. The second treatment group spent all of its computer training time accurately reading stories with speech support, with no explicit phonological training. Phonological training was significantly more effective for the development of phoneme awareness and phonological decoding skills for all grade levels, and was superior for growth in non-speeded word recognition at the end of training in grades 2 and 3, but it did not exceed the benefits of accurate reading of stories on the computer or growth in fluent word recognition at any grade. For poor readers in grades 4 and 5, practice in reading stories with computer support yielded greater gains in fluent word recognition compared to the phonological condition. Thus, Wise et al (2000) provided the first published demonstration of a treatment by grade level interaction with poor readers in a single study, although an analysis of individual studies at different grade levels suggested the same pattern (National Reading Panel, April 13, 2000). While it is clear that phonological training has significant benefits for early reading development, sustaining these benefits through later reading development and achieving transfer from improved phonological skills to better reading, particularly reading fluency in older poor readers, are elusive goals (Lyon, Fletcher, Fuchs, & Chhabra, in press).

Although the specific and lasting benefits of phonological training were not as strong as had been hoped, it was clear that both computer training conditions yielded strong gains in word reading accuracy and fluency compared to regular reading classroom controls (Wise et al., 1999). In fact, effect sizes relative to the classroom control group and standard-score gains were comparable to those found for intensive individually administered interventions without computer support (reviewed by Torgesen, 2002). These results might have supported an application for continued study of the programs under the CLDRC renewal for the current budget period, but there were several constraints that limited their use with children at risk below the 2nd grade, as described in the application for Project V, and they required the costly presence of a trainer for small groups of 3 or 4 children. With separate program development funding over the past 4 years and advances in computer technology, the programs have been adapted for younger children at risk for reading disability beginning at the end of kindergarten, and they can be used in the regular classroom without the support of a trainer. In the present CLDRC renewal application, we propose to use these programs for school-based studies of response to instruction for children at risk beginning in kindergarten, and for home-based studies of twins in the 2nd-4th grades with diagnosed reading disabilities (new Project V, Wise PI., Byrne & Cole Co-Is).

The previous CLDRC application for the current 2000 - 2005 budget period included two new projects that were responsive to that RFA. They included a major expansion of Project II, with a subcontract to Janice Keenan at the University of Denver to focus on reading and listening comprehension, and Project V with a subcontract to Brian Byrne to support a preschool longitudinal twin study in Australia. The increased emphasis on reading comprehension in our current Project II was based on evidence that while deficits in phonological decoding and word recognition were important in most reading disabilities, there were additional constraints on the ultimate goal of reading, comprehension. This emphasis on comprehension is continued and expanded to include fluency and writing in the renewal application for Project II. The increased emphasis on longitudinal research beginning in preschool (current Project V) was in response to concerns that we had limited knowledge of the genetic and environmental influences on reading-related development during this early critical period. Important preliminary results from the current Project V study in Australia and parallel studies in Colorado and Scandinavia have been published or are in press (Byrne et al., 2002; Byrne et al., in press; Samuelsson et al., in press). Byrne and Olson plan to continue their longitudinal research through separate funding of the Australian study and the continuation of a parallel R01-supported study in Colorado, for reasons described below.

The new RFA for competitive renewal of the Centers stressed the importance of exploring response to instruction as a basis for the early diagnosis and prevention of reading disabilities. We agree that this is a valuable approach, and we are excited to have the opportunity to explore the genetic and environmental influences on response to instruction in the proposed Project V. The cost of the new Project V has required a shift of resources from the current Project V that funded the preschool longitudinal twin study in Australia. Brian Byrne has applied for funding from the Australian Research Council to continue data collection there, and the parallel study in Colorado (Olson, PI) is supported by a separate R01 grant. A renewal application for the R01 is currently under review. In view of the uniqueness and significance of this international (also in Scandinavia) longitudinal behavior genetic study of prereading and early reading development and the progress we have made, we are confident that our competitive renewal applications will be funded. Regardless of the outcome of these reviews, the twins in the Colorado longitudinal study will be available for the study of response to instruction in the new Project V, and Byrne will continue his valuable association with the Center as co-PI of that project.

In conclusion, the CLDRC projects have evolved across the Center funding cycles to be responsive to emerging knowledge and changing emphases in the Center RFAs. With the present Center application, we look forward to developing new knowledge about the etiology of disabilities in reading and listening comprehension and reading fluency, writing, comorbidity of reading disability and ADHD, and response to instruction, all within a behavioral- and molecular-genetic design. In addition, Center investigators plan to continue their longitudinal behavior-genetic studies of early and later reading development through separately funded studies (see Other Support for PIs Byrne, Olson, and Wadsworth).