Ronald B. Gillam, Ph.D. - US grants

This proposal focuses on short-term, modality-specific memory mechanisms in "specifically language impaired" (SLI) children. There have been few recent investigations of verbal memory in the speech-language pathology literature. There have, however, been recent breakthroughs in understanding the perceptual and conceptual mechanisms underlying memory in normal adults that could be beneficially applied to SLI children. The proposed research will be based on two phenomena of short-term verbal memory. First, the modality effect is the finding that the accuracy of serial verbal recall is greater when stimuli are presented to the auditory than to the visual modality (unlike stimuli in simultaneous spatial arrays, which are better recalled from the visual modality). Second, the suffix effect is the finding that the auditory modality superiority in verbal recall is much reduced when the spoken list is followed by a final spoken item that is not to be recalled. These two phenomena have been extensively investigated and appear to indicate that normal subjects make use of temporary acoustic and phonetic-memory codes that are especially well-suited for the retention of serial verbal information, although they are vulnerable to interference. Recent studies also demonstrate that the modality and suffix effects are obtained in the memory and comprehension of natural, linguistically coherent language. In the proposed experiments, modality and suffix effects will be examined in SLI children and in other groups matched for age and reading abilities, in order to determine if the use of auditory modality-specific memory or its vulnerability to interference differs in these groups. List recall procedures will be used in the early experiments. Later studies extend the findings to the comprehension of coherent language.

This proposal focuses on short-term, modality-specific memory mechanisms in "specifically language impaired" (SLI) children. There have been few recent investigations of verbal memory in the speech-language pathology literature. There have, however, been recent breakthroughs in understanding the perceptual and conceptual mechanisms underlying memory in normal adults that could be beneficially applied to SLI children. The proposed research will be based on two phenomena of short-term verbal memory. First, the modality effect is the finding that the accuracy of serial verbal recall is greater when stimuli are presented to the auditory than to the visual modality (unlike stimuli in simultaneous spatial arrays, which are better recalled from the visual modality). Second, the suffix effect is the finding that the auditory modality superiority in verbal recall is much reduced when the spoken list is followed by a final spoken item that is not to be recalled. These two phenomena have been extensively investigated and appear to indicate that normal subjects make use of temporary acoustic and phonetic-memory codes that are especially well-suited for the retention of serial verbal information, although they are vulnerable to interference. Recent studies also demonstrate that the modality and suffix effects are obtained in the memory and comprehension of natural, linguistically coherent language. In the proposed experiments, modality and suffix effects will be examined in SLI children and in other groups matched for age and reading abilities, in order to determine if the use of auditory modality-specific memory or its vulnerability to interference differs in these groups. List recall procedures will be used in the early experiments. Later studies extend the findings to the comprehension of coherent language.

DESCRIPTION (from applicant?s abstract): Nearly 7% of elementary school children present difficulties learning and using language. Unfortunately, language impairments are often long lasting and may have serious social, academic, and vocational ramifications. More than 1 million children receive language intervention in the public schools each year, and many more are seen in hospitals and other clinical settings. Recently, a computerized intervention approach known as Fast ForWord has received a great deal of attention at scientific meetings and in the popular press. The creators of this program (Michael Merzenich, Paula Tallal, and their associates) claim that Fast ForWord trains children?s brains to process speech better by gradually reducing specialized acoustic modifications of speech stimuli as children improve on the games. The creators of Fast ForWord have completed a large, one-group, pretest-postest national field trial of their program. Their results suggest that children make approximately 1 standard deviation of improvement on standardized tests after a 6-week intervention period. We propose a randomized clinical study to compare the language outcomes of Fast ForWord to two other interventions (computer assisted language intervention without acoustically modified speech and individual language intervention) and to the outcomes of a general stimulation (control) condition. Each year for 3 years, 24 children will be randomly assigned to each of the four conditions at three regional sites (Austin, Texas Dallas, Texas, and Lawrence, Kansas) for a total of 54 children in each condition at the end of the study. The treatments will be administered in special summer programs. The primary research question is which intervention results in the most improvement in the composite language score from the Oral and Written Language Scales. Secondary questions include which intervention results in the greatest gains in conversational language, which intervention results in the greatest gains 3 and 6 months after training, which intervention results in the greatest improvement in auditory perception, and which intervention is the most cost effective. Our results will have theoretical and practical value. Theoretically, our study tests the temporal processing hypothesis of language impairment. Practically, our study will describe and compare the language, communication, auditory processing and academic outcomes of three interventions. Our analyses will help clinicians and administrators choose the most effective and least expensive treatment for the children they serve.

DESCRIPTION (provided by applicant): Research on children with Specific Language Impairment (SLI) has primarily focused on their expressive deficits. Relative to our understanding of these children's expressive difficulties, their receptive problems are poorly understood. This has led to underspecified theoretical accounts of SLI, a limited array of effective interventions to treat them, and less than favorable language intervention outcomes. The results of meta-analyses, systematic reviews, and clinical trials indicate that language intervention has a smaller effect on language comprehension than on language production (Cirrin & Gillam, 2008; Law et al., 2004; Bishop, et al., 2006). Recently, Leonard (2009) has argued that an inability to process linguistic information in the input is likely to play a prominent role in the comprehension difficulties of children with SLI. Debate exits as to whether the sentence comprehension problems in children with SLI reflect a deficit in the language system (the domain-specific view) or a general deficit in cognitive processing (the domain-general view). The primary goal of this large-scale project is to systematically determine which account provides the better descriptive and explanatory model to characterize sentence comprehension in children with SLI. The proposed 5-year project will address two related specific aims in five integrated studies. 450 children will participate: 150 children with SLI (ages 9;0-11;11), 150 age/nonverbal IQ matched (CA) children, and 150 younger children matched for short-term memory and vocabulary (YMV). Aim 1 employs a psychometric approach to investigate the relationship between several cognitive processing mechanisms (controlled attention, lexical retrieval, retrieval interference, short-term memory, working memory, and processing speed) and the comprehension of noncanonical sentences (Studies 1-2) or canonical sentences (Studies 3-4). We hypothesize that the modeling results will support the domain-general account. As part of this aim, we will also examine whether similar sets of cognitive mechanisms underlie sentence comprehension in children with SLI and CA children (with different a set likely subserving noncanonical and canonical sentences). Aim 2 more specifically examines working memory retrieval in noncanonical sentence processing to determine whether children with SLI fail or are slower to reactivate a prior constituent (NP1) during noncanonical sentence processing. We hypothesize they will be slower relative to CA children, supporting the domain-general view. Overall, the results may lead to four high impact implications for the field of SLI. First, a fundamentally new theoretical understanding of the relationship between cognitive processing and sentence comprehension may emerge. Second, models of normal adult and SLI sentence comprehension may be merged into a coherent developmental framework. Third, clinical cognitive-linguistic profiles of children with SLI will be expanded. And fourth, critical insights may emerge into which cognitive mechanisms could be targeted in alternative language treatments designed to improve the language comprehension of children with SLI.