Nelson Cowan - US grants

This research examines preschool children's short term memory for speech. The goal is to investigate two basic processes in preschool children about which little is known: (a) the decay over time of memory for speech, and (b) the interference with memory for speech caused by subsequent input to the child. Because areas of the cortex subserving spoken language reception undergo substantial maturation in the preschool years, the related properties of memory may change. Memory for speech frequently has been explored in adults, but the techniques that have been used are not directly applicable to young children. Therefore, in the present proposal, three paradigms are developed in order to examine young children's memory for speech. One, a same-different, discrimination procedure, will be used to examine memory decay. A second paradigm makes use of rhymes to examine modalityspecific aspects of children's memory for spoken words. These experiments will examine memory decay as well as interference between words. In a third paradigm, children's ability to remember words is examined in the presence of various sentence-like interfering stimuli. Thus, the plan is to assess the most basic properties of children's memory for speech, and then to examine these properties of memory within realistic language-learning situations. The theoretical impact of the research is that, together with knowledge of brain maturation, it will contribute to an understanding of brain-behavior correspondences. Moreover, because preschool children are much less likely than adults to use retention strategies such as rehearsal, the research should provide information about basic structural as opposed to strategic properties of memory. It therefore should contribute to a general understanding of cognitive processes as well as development. The proposed research should also be of considerable practical importance, both because it would lead toward an improvement in the format of verbal instruction s to children, and because memory for speech is frequently impaired within many disorders of language.

This is a proposal to examine short-term memory for speech and its development in childhood. The research is based on the premise that an analytic understanding of memory for speech requires a distinction between passive processes that occur automatically (e.g., auditory sensory and phonological memories and their decay) and active processes that can be initiated and/or carried out only with at least some minimal allocation of attention (e.g., attentive encoding of speech; overt or covert articulation). The main goal of the research is to disentangle these types of memory faculty and to determine how each of them develops. To this end, there are three specific aims: (1) To examine developmental changes in the rate of auditory memory decay; (2) To examine developmental changes in how spoken utterances are temporarily retained if the utterances are ignored at the time of their presentation (e.g., changes in which items are retained, for how long they are retained, or with what forms of coding they are retained); and (3) To determine the contributions of passive, auditory and phonological memory decay vs. active, articulatory processes in immediate memory tasks and their development. Research addressing each of these aims is based on a different method. In the examination of auditory memory decay, subjects compare two slightly different sounds separated by a variable interval during which memory of the first sound can decay. In the examination of ignored speech, subjects are to ignore monosyllabic, spoken words while engaging in the silent processing of visual materials. Occasionally, there is a cue for the subject to interrupt the visual task and identify one or more of the most recent syllables that had been ignored. Last, the examination of passive decay and active processing involves an immediate, serial memory task in which the delay between the to-be-recalled list and the time of recall of each item is measured and related to the amount of speech memory decay that occurred. In all three procedures, there is an emphasis on testing subjects of different ages at similar levels of performance so that forgetting rates can be compared. It is clear that memory of the auditory and phonological qualities of speech are affected in important ways by a wide range of language-related disabilities, but little is known about the memory mechanisms involved or their normal development. The present research helps to fill these gaps.

DESCRIPTION: This competing continuation proposal examines how short- term memory for attributes of speech and related sounds changes during childhood. It distinguishes between passive, effortless processes (e.g., sensory and phonological memory storage) and active processes that are carried out only if sufficient attention is allocated (e.g., covert articulation and memory search). Whereas past research has focused on developmental changes primarily in a few active processes (e.g., covert verbal rehearsal), the investigator has recently observed substantial decreases in passive forgetting rates across ages, in two different procedures: one in which participants compared the pitches of two slightly different tones separated by a silent interval of variable duration, and another in which participants tried to identify words that were ignored during both their presentation and a subsequent silent interval. The investigator has also obtained a fresh perspective on changes that may occur in active processing. Specifically, in a serial recall task, the durations of silent pauses between words in the spoken responses were found to decrease with development. The results suggested that the pause changes stem from a developmental increase in the efficiency of short-term memory search. Interestingly, pause durations correlated with span but sill were independent of another correlate of span, the speaking rate. The proposed research aims to establish more firmly these newly observed changes in passive and active processing, and to determine their contribution to short-term memory development. The methods involve combinations of the experimental tasks already in use in the investigator's laboratory and variations of those that allow cross-fertilization among them. For example, two-tone comparisons will be examined with the first tone ignored during its presenation, and ignored-speech studies will examine memory for multiword lists. The goal is to determine the relations between tasks in order to identify underlying theoretical mechanisms that enter into more than one of them. Children's passive forgetting rates will be used along with both memory search and speaking rates to predict memory span much more accurately than has been possible until now. The results should lead to an improved theoretical model of normal short-term memory developmental that will be of considerable use in the interpretation and eventual treatment of learning and language acquisition difficulties in children.

Working memory (WM) can be defined as the amount of information that can be held in the mind and manipulated in order to allow the completion of a mental task such as comprehension, problem-solving, or serial recall. WM often is measured with tasks that require the storage and processing of information at the same tome. Past research shows that individual and developmental differences in WM are important for such complex cognitive tasks, but it is not yet clear what basic processes underlie WM measures and their developmental change. This revised competing continuation proposal adapts methods that we have devised recently to observe developmental changes in serval basic processing parameters, from age 6 to adulthood: the short-term memory retrial rate for recently presented information, speeded articulation rates; the persistence of automatically held information, such as sensory memory; and the capacity of S.T.M., measured in unrelated chunks of information. The observations of developmental changes in these parameters are important because many researchers have suggested that most of these basic parameters (analogous to computer hardware) stay fixed, and that is primarily knowledge and strategies (analogous to computer software) that change with development in childhood. Rather than just accumulating further support for developmental changes in the basic parameters, the planned research has three other specific aims. (1) The first aim is to understand the mechanisms of the basic parameters within information processing system. For example, one study examines whether the capacity limits are central limits in the attentional focus as expected, or instead are capacity limits in modality-specific stores ( e.g., separate capacity limits for auditory/verbal and visual/spatial stimuli). (2) the second aim is to learn how the basic parameters influence one another. For example, one study examines whether the developmental change in retrial rates could be an indirect consequence of the change in capacity limits, by imposing a memory load during the measure of retrieval rates. (3) The third aim is to learn how the basic parameters are related to more complex measures of information processing ability, as well as to one another. This will be investigated in relatively large correlational study and with models incorporating the basic parameters, complex as well as simple memory spans, and measures of aptitude and scholastic achievement. This theoretical advance also should help in understanding various learning and language disorders, which are known to be marked by poor memory span for theoretical reasons that need clarification.

DESCRIPTION (provided by applicant): This is a revision of a competing continuation / renewal. The period 7/1/2001 - 6/30/2006 continued work funded since 1984. This period resulted in 1 book, 37 journal articles, and 14 book chapters related to working memory and its development in childhood. Working memory is the collection of processes that keep limited information in an especially accessible form for use in ongoing cognitive tasks. To build on theoretical progress, I emphasize recent research on the roles of two types of processes: those requiring attention and effort, and those taking place automatically. Our research on working memory in elementary school children and adults is sharpening notions of what develops in that time range, including changes in both the scope and control of attention (e.g., Cowan, Fristoe et al., in press). Our child research also provides insight into theories of adult processing. For example, the working memory tasks that best predict mental aptitudes seem to be those for which covert verbal rehearsal cannot be used efficiently;not necessarily tasks that include separate storage and processing components as conventional wisdom suggests (Cowan, Elliott et al., 2005). Our current aims address three basic questions: (1) What is the role of working-memory storage in a form abstract enough to extend across different types of codes (i.e., a central memory)? (2) Is attention used to store information, or only to defend the stored information from interference? Finally, (3) How are capacity and retrieval speed related? Even though correlations between retrieval speed and memory ability are high, experimentally increasing this speed did not improve capacity in children (Cowan, Elliott et al., 2006). To examine the role of attention in working-memory storage, the proposed followup research involves various procedures (e.g., dual-task designs to probe which processes interfere with which types of memory;verbal-spatial association memory tests to identify an abstract, cross-code, central type of memory). Speech response timing is used with multi-word chunk teaching procedures to estimate the capacity of attention- related working memory in chunks. The research could help to clarify a wide range of cognitive disorders that involve working memory (e.g., learning and language disabilities and attention deficits). It could indicate whether normal functioning of working memory depends on attention for storage or just for processing information that is of practical import in formulating plans for clinical treatment and diagnosis.

PROJECT SUMMARY/ABSTRACT BACKGROUND: Working memory (WM) is the limited information retained in an active state for use in ongoing cognition. Improvements in WM from 6-14 years are critically important for how children do diverse cognitive functions like reasoning, problem-solving, and language. When teaching children (typical or challenged), little is known about how to take into account WM limits because we do not yet understand what factors contribute to typical WM development. Research by the P.I. under Grant R01-HD-21338 shows that accounts of WM development that have been proposed are insufficient (e.g., improvement in ignoring distraction or in item familiarity). We have shown this using methods in which the factors in question were experimentally controlled in new ways (e.g., capacity growth was observed independent of any distraction or familiarity effects). In my theoretical framework, WM capacity limits come from how many items can be retained concurrently in the focus of attention with enough detail to guide responses. In a new developmental hypothesis within that account, the number of WM items may remain fixed but WM develops in the richness of features within items and patterns noticed across items. Older children and adults would use features and patterns to minimize the need for attention to WM items. The RESEARCH GOAL is to assess this hypothesis with variants of new dual-set (e.g., visual+acoustic) recognition tasks, to identify roles of attention, patterns, and features in typically developing children and young adults. SPECIFIC AIMS are to uncover these bases of WM development in four ways. (1) We assess whether it is not just the number of items or objects in WM that increases in development, but the completeness of their feature representations. (2) Inasmuch as our previous work demonstrated the importance of general attention to retain information in WM, we explore a factor that may free up attention in more mature participants. Older children may engage in rapid pattern formation and memorization to ?off-load? materials out of the focus of attention, freeing attention for subsequent input and WM maintenance. We examine whether extra structure in the material allows younger children to off-load more like older ones do. (3) We recently found that off-loading occurs for acoustic and verbal lists (words or tones) much more than for visual arrays, reducing interference between acoustic items and other items. We will investigate the basis of these intriguing findings. Possibly, sequential lists of colored spots, like acoustic sequences, will allow better off-loading than do arrays of spots. We also will examine whether vibrotactile sequences compete with visual objects for attention, more than do acoustic items. Last, (4) we adopt experimental techniques from recent adult work to determine whether attention-dependent and attention-free mechanisms both develop similarly in terms of a) an increased numbers of items in WM, or b) increased item precision; their mechanisms may differ. IMPACT: How cognition should be engaged in educating children and addressing learning disorders depends on basic mechanisms of WM development, which we must uncover.