JSLHR

Research Note

Lexical Decay During Online Sentence Processing in Adults With Specific Language Impairment Gerard H. Poll,a Holly S. Watkins,b and Carol A. Millerb

Purpose: Decay of memory traces is an important component of many theories of working memory, but there is conflicting evidence on whether the rate of decay differs for individuals with specific language impairment (SLI) as compared to peers with typical language. The authors tested the hypothesis that adults with SLI have a slower decay rate. Method: Twenty adults with SLI, ages 18–27 years, and 23 age-matched peers identified target words in sentences. Sentences were presented at normal and slow rates. Participants separately judged whether a picture and sentence matched in meaning as a measure of sentence processing efficiency.

Results: After controlling for sentence processing efficiency, the group with SLI was slower to detect words in sentences. Response times for the group with SLI increased less in the slow condition as compared to the group with typical language, resulting in a Group × Presentation Rate interaction. Conclusions: The Group × Presentation Rate interaction is consistent with a slower lexical decay rate for adults with SLI, but differences in the ability to manage interference could not be ruled out. The findings suggest that decay rate differences may play a role in the working memory limitations found in individuals with SLI.

I

capacity: how memoranda may be affected by time and competing information. For typical individuals, the limits of working memory have been explained by mechanisms that change over the course of a task: decay rate (Baddeley, 2000; Towse, Hitch, & Hutton, 1998), interference (Oberauer, Lewandowsky, Farrell, Jarrold, & Greaves, 2012), and changes in the focus of attention (Cowan, 2010). The first of these dynamic mechanisms, memory trace decay, is an important element in some explanations for working memory capacity limitations in individuals with typical language (Altmann & Gray, 2002; Towse et al., 1998). Evidence for the role of decay comes from studies in which participants are asked to recall a fixed amount of material across varied durations while also performing concurrent processing tasks. Such studies have found that time-based decay is a factor in working memory capacity (Barrouillet & Camos, 2001; Towse et al., 1998). Decay is also a key element of Baddeley’s (2000) working memory model, consisting of a phonological loop, a visuospatial sketchpad, and a central executive. The phonological loop is a temporary store for verbal information, which is assumed to decay within seconds unless refreshed by rehearsal. When participants are prevented from rehearsing information in the phonological loop, they show marked reductions in recall. Rehearsal can be prevented by articulatory suppression

ndividuals with specific language impairment (SLI) have unusually poor language abilities in the absence of hearing impairment, autism, cognitive impairment, or other clear causes (Leonard, 1998). Converging evidence indicates that SLI continues into adulthood (Clegg, Hollis, Mawhood, & Rutter, 2005; Johnson et al., 1999), with important consequences for employment and educational outcomes (Conti-Ramsden & Durkin, 2012). Processing-based explanations for SLI suggest that individuals with SLI have more limited working memory capacity (Ellis Weismer, Evans, & Hesketh, 1999) or slower cognitive processing speed (Miller et al., 2006) than individuals with typical language. For individuals with SLI, working memory limitations have been linked to sentence comprehension difficulties, in particular for complex sentences (Montgomery & Evans, 2009). The evidence is less clear on whether these individuals differ in the dynamic mechanisms affecting working memory

a

Elmhurst College, Elmhurst, IL The Pennsylvania State University, University Park

b

Correspondence to Gerard H. Poll: [email protected] Holly S. Watkins is now at The University of Texas at Dallas. Editor: Rhea Paul Associate Editor: Kristine Lundgren Received September 29, 2013 Revision received March 12, 2014 Accepted August 6, 2014 DOI: 10.1044/2014_JSLHR-L-13-0265

Disclosure: The authors have declared that no competing interests existed at the time of publication.

Journal of Speech, Language, and Hearing Research • Vol. 57 • 2253–2260 • December 2014 • © American Speech-Language-Hearing Association

2253

or an attentional distractor (Baddeley, 2000; Barrouillet & Camos, 2001). Other researchers have challenged the concept of time-based decay as a factor in working memory capacity (Oberauer & Lewandowsky, 2013), suggesting instead that interference is the limiting factor. During a working memory task interference results from competition of memoranda from either past trials ( proactive interference) or new incoming information (retroactive interference; Jonides et al., 2008). The argument is that individuals who are more successful at suppressing competing memoranda have larger working memory spans. Attentional distractors reduce working memory capacity by eliminating time needed to suppress interfering items and possibly by being interfering items themselves. Functional decay theory (Altmann & Gray, 2002) attempts to bridge the decay and interference perspectives. Decay can be functional when older, less relevant items fade and reduce interference. Individuals adapt the rate of decay to manage interference in different task conditions. As the presentation rate of incoming items increases, the decay rate should also increase to manage interference. The third dynamic mechanism is attention. Attentional focus makes items in working memory resistant to interference (Cowan, 2010). Under this conception of working memory, a very limited set of items can be maintained in the focus of attention. A broader set of items may be activated in long-term memory, but without attention the activation persists for only a short time. If a particular item is identified for later recall, it can be tagged, or activated in long-term memory (Tarnow, 2008). The speed of recognition, or the “retagging” of the target item, depends on how much activation has been lost as a function of time. As with decay rates and interference effects, it is not known whether the time course for the loss of activation of tagged items differs for adults with SLI. Even though differences in decay, interference, or retagging patterns could help explain working memory limitations in SLI, few studies have addressed these issues. The evidence that does exist is conflicting. Two studies have suggested that the decay rate is faster for individuals with SLI (Gillam, Cowan, & Marler, 1998; McMurray, Samelson, Lee, & Tomblin, 2010). In one of these studies, McMurray and colleagues (2010) asked adolescents with and without SLI to identify a picture that matched a word presented aloud. The authors accounted for each group’s pattern of looks to the correct pictures and foils by fitting a series of computational models. The best fitting model involved a faster decay rate for the group with SLI. In contrast, a study of picture naming in the presence of related and unrelated words suggested a slower decay rate for individuals with SLI (Seiger-Gardner & Schwartz, 2008). Words were presented before, during, and after the pictures. Children with SLI showed a lingering semantic inhibition effect for related words presented after the pictures that was not observed in children with typical language. In a separate word learning study no difference was found in decay rate for individuals with SLI as compared to peers

2254

with typical language (Alt & Spaulding, 2011). This study provided time for participants to rehearse words before being asked to recall them, which may explain the null finding on decay rate differences. The ability to manage interference may also explain working memory capacity limitations in SLI. Children with SLI are less able to suppress interfering information to complete tasks when compared to typical peers (Im-Bolter, Johnson, & Pascual-Leone, 2006; Spaulding, 2010). In listening span tasks, children with SLI have been shown not only to have more limited working memory capacities but also to make more errors that involve words from previous trials (Marton, Schwartz, Farkas, & Katsnelson, 2006). These results indicate that children with SLI are less able to manage interference, but the results may also imply a slower decay rate. Montgomery (2005) presented evidence of attentional or decay rate differences affecting children with SLI. Children with and without SLI listened to a target word and were asked to press a button as soon as they detected the target word in a sentence. The sentences were presented at fast, slow, and normal rates. After controlling for grammatical ability, the children with typical language had shorter response times than the children with SLI in the fast and normal rate conditions. The children with SLI had shorter response times in the slow rate condition. Montgomery suggested that the children with SLI were better able to control their attention in the slow condition, but the findings are also consistent with other conceptions of working memory, such as a slower lexical decay rate or slower loss of activation for tagged items (Tarnow, 2008). If the target word was more highly activated during the interval before its appearance in the sentence for children with SLI and less activated for the children with typical language, the children with SLI could respond to the target word more quickly. Although there is limited direct evidence of slower decay in SLI during sentence processing, a slow decay rate hypothesis is consistent with other findings. Compared to typically developing peers, children with SLI have been shown to have a larger priming effect in a lexical decision task (Pizzioli & Schelstraete, 2011) and greater than typical lexical activation for expected words in sentences (Neville, Coffey, Holcomb, & Tallal, 1993). Greater lexical activation levels may be consistent with extended persistence of activation (i.e., slower lexical decay rates) in individuals with SLI. Slower lexical decay may in turn be a factor in more limited working memory capacity. The evidence for children with SLI suggests that they differ from typical peers in the dynamic mechanisms affecting working memory capacity, including rate of decay, suppression of interference, or rate of change in activation for tagged items. Evidence is limited on whether adults with SLI differ from peers in these mechanisms. In sentence processing, typical adults immediately activate lexical items, and these activations rapidly decay (Ferrill, Love, Walenski, & Shapiro, 2012). Our hypothesis for this study was that adults with SLI would differ in the dynamic mechanisms affecting working memory. Our focus

Journal of Speech, Language, and Hearing Research • Vol. 57 • 2253–2260 • December 2014

was to determine whether adults with SLI exhibit a slower decay rate than their peers with typical language. To test this hypothesis, we had adults with and without SLI detect target words in sentences. Sentences were presented at both normal and slow rates. A Presentation Rate (normal, slow) × Group (SLI, typical language) interaction involving a smaller group difference in the slow condition would provide evidence of a slower lexical decay rate in adults with SLI.

Method Participants Participants for the study were 43 adults, ages 18– 27 years, recruited from postsecondary schools in central Pennsylvania (21) and from a registry of participants from an epidemiological study of SLI coordinated by the University of Iowa (22; for details on initial Iowa recruitment, see Tomblin, Zhang, Buckwalter & O’Brien, 2003). All participants were native speakers of English with no reports of hearing impairment, intellectual disability, autism, cerebral palsy, or serious neurological injury. Participants passed a pure-tone hearing screening at 25 dB HL at 500, 1000, 2000, and 4000 Hz. Their Performance Intelligence Quotient (PIQ) was 80 or above on a composite of Picture Completion, Digit Symbol Coding, and Matrix Reasoning subtests from the Wechsler Adult Intelligence Scale—Third Edition (Wechsler, 1997), using a score-combining process from Sattler and Ryan (1999). Participants were classified as SLI with a positive history of language difficulties and scores in the affected range on language ability tests. Reported language difficulties were either a prior diagnosis of SLI (14 cases from the Iowa registry) or difficulties with reading comprehension (four cases) or spoken grammar (two cases). Language ability testing was based on the discriminant function developed by Fidler, Plante, and Vance (2011) for classifying adults with reported learning disabilities as having language impairment. Scores from a spelling task, a word definitions task, and a sentence comprehension task were entered into a discriminant function. Scores resulting in a positive output from the discriminant function were required for the participant to be classified as SLI. Participants with no history of language learning difficulties and test scores resulting in a negative discriminant function output were classified as having typical language. Further details and scores by group are provided in Table 1. Twelve members of the group with SLI and 19 members of the group with typical language were females. The groups did not differ in age, t(41) = 1.70, p = .10; but did differ in years of education, t(41) = 4.83, p < .001; and PIQ, t(41) = 5.87, p < .001. Given that PIQ in children with SLI tends to decline in adolescence (Botting, 2005), we did not match or statistically control for PIQ because doing so would artificially restrict our ability to detect bona fide group differences, leading to findings not representative of the population of adults with SLI (Fidler et al., 2011).

Table 1. Participant characteristics.

Group with SLI (n = 20) Measure Age Years of education Performance IQ Modified Token Test Spelling Word Definitions

Group with typical language (n = 23)

M

SD

M

SD

22.5 13.1 97.3 69.7 3.7 7.6

2.0 1.1 8.1 17.1 2.3 3.3

21.5 14.5 113.7 91.3 11.4 13.1

1.8 0.9 10.0 5.1 1.9 1.5

Note. The performance IQ was calculated from the Picture Completion, Digit Symbol Coding, and Matrix Reasoning subtests of the Wechsler Adult Intelligence Scale—Third Edition using the approach from Sattler and Ryan (1999). The Modified Token Test (de Renzi & Faglioni, 1978; Morice & McNicol, 1985) scores are the group mean percentage correct. Spelling is the number of words spelled correctly of the 15 words presented from Fidler et al. (2011). Word Definitions is the group mean standard score for the subtest from the Clinical Evaluation of Language Fundamentals— Fourth Edition (Semel, Wiig, & Secord, 2003).

Materials We administered a truth value judgment task as a control measure for sentence processing time and to adjust for the ability of participants to anticipate target words from context clues. We administered a word detection task to measure lexical decay rate in a sentence context. The truth-value judgment task used 36 grammatical sentences from Miller, Kail, Leonard, and Tomblin (2001). The sentences had simple active (“The girl is chasing the boy”), simple passive (“The baby is being fed by the girl”), or compound subject (“The boy and the horse are washing the cow”) structures. The sentences agreed with the pictures in half of the trials. Sentences were audio recorded with a normal rate and prosody by a male speaker of Standard American English. Sentences were presented in a single pseudorandom order, with no more than three consecutive trials of the same class (matching or not matching sentence and picture) or sentence structure. The word detection task was based on Leonard, Miller, and Finneran (2009). For this study, we used 77 grammatical sentences. In each trial, participants heard a target word followed by a sentence. They were to press a response button as soon as they heard the target word in the sentence. For example, participants heard dinner. They then heard, “The hiker at Yosemite always cooks dinner over a campfire.” Experimental sentences were eight to 12 words long, and target words appeared as the fifth, sixth, or seventh word. Target words exceeded a written frequency of two occurrences in the Francis and Kučera corpus (1982). Thirty-four sentences in the slow condition were matched to 34 normal rate condition sentences on target word spoken frequency (Davies, 2009) and duration to the target word using the Match program (van Casteren & Davis, 2007). Nine sentences were presented at a normal rate but did not include the target word. These sentences were catch

Poll et al.: Lexical Decay in Specific Language Impairment

2255

trials designed to ensure that participants attended to the sentences carefully before responding. The use of a slowed presentation rate to test a group difference in decay rate was based on Purser and Jarrold (2005). Sentences for all conditions were digitally recorded in a sound-attenuated room by a male speaker of Standard American English. Using Praat (Boersma & Weenink, 2006), the sound files were manipulated to have a common average sound pressure level. Sentences in the slowed-rate condition were manipulated in Praat so that the duration was extended by 50% without altering the frequencies.

Procedure Both tasks were presented by computer using E-Prime 2.0 (Psychology Software Tools, 2009) scripts and a serial response box. The response box was placed on the side of dominant hand of the participant, and the participant’s hand rested on the box. Both tasks were completed in community settings using conventional headphones. Tasks were presented at a comfortable loudness. Participants were prompted to adjust loudness after the initial training trials for each task. Task order was counterbalanced across participants. For each task, the investigator read instructions from a script. In the truth-value judgment task, participants were presented with a line drawing. After 2 s, they heard a sentence. They then pressed a green or red button on a response box to indicate whether the sentence did or did not match the picture. Six practice trials were presented before the experimental trials. The score for the truth-value judgment task was the mean response time for valid trials. We measured response time from the onset of the sentence sound file to the participant’s button press. We excluded incorrect judgments and response times that were more than twice the participant’s mean, following procedures from Bowers, Vigliocco, Stadthagen-Gonzalez, and Vinson (1999) and Miller et al. (2001, 2006). In the word detection task participants were instructed to press the response button as soon as they heard the target word in the sentence. The task began with six practice trials. After successful completion of practice trials, experimental trials began with 2 s of silence followed by the target word. After 500 ms the sentence sound file was initiated. The trial ended with the participant response or after a 2-s wait time beyond the end of the sentence. We measured response times from the onset of the target word in the sentence to the participant button press. Response times greater than twice the participant’s mean response time and button presses prior to the target word were eliminated as outliers. For each condition, the mean response time of the remaining valid trials was the participant’s word detection score.

Results For the control task (truth-value judgment), outlier removal eliminated 2.7% of total responses. Accuracy for

2256

truth-value judgments was high (97% typical language, 96% SLI). Mean response times were 2,350 ms (SD = 297) for the group with SLI and 2,199 ms (SD = 304) for the group with typical language, with no significant group difference, F(1, 41) = 2.71, p = .102, hp2 = .062. For the word detection task, removal of outlier responses eliminated 4.8% of the responses for the group with typical language and 5.4% of responses for the group with SLI. Catch trials were also excluded from the analysis. There were two catch trial responses from the group with typical language and three from the group with SLI. The mean response times by group and by sentence presentation rate for the word detection task are displayed in Figure 1. Mean response times for the group with SLI were longer in both the normal (444 ms, SD = 124) and slow (492 ms, SD = 147) conditions compared to the group with typical language (normal: 341 ms, SD = 60; slow: 408 ms, SD = 91). The difference between the normal and slow conditions was smaller for the group with SLI. To better approximate normality, data from the word detection task were inverse transformed. These transformed response times were entered as the dependent variable in a mixed between–within-subjects analysis of covariance, with group and presentation rate as the independent variables and mean response time for the truth-value judgment task as the covariate. We added years of education as a second covariate to account for group differences. There were significant main effects of group, F(1, 39) = 6.36, p = .016, hp2 = .140, and presentation rate, F(1, 39) = 59.29, p < .001, hp2 = .603. The group with SLI had longer response times than the group with typical language, and participants had longer response times in the slow condition, after controlling for sentence processing efficiency and education. There was also a Group × Presentation Rate interaction, F(1, 39) = 4.67, p = .042 , hp2 = .102. The group difference in response time was larger in the normal rate condition than in the slow condition.

Discussion In this study we asked whether dynamic mechanisms affecting working memory capacity differ in adults with SLI compared to adults with typical language. We reasoned that response times would be related to degree of activation. In a sentence-embedded word detection task adults with SLI were slower to detect words than peers with typical language, but the response time gap between groups was smaller in the slow rate condition than in the normal rate condition. This interaction of group and presentation rate supported our hypothesis that decay rates differ in adults with SLI, and the main effect of rate was consistent with a general effect of decay on short-term memory traces. We will consider other accounts for the findings. Sentence context likely provided cues to target word placement. To reduce this effect, we controlled for sentence processing efficiency with truth-value judgment response times. This made differences in ability to use context cues a

Journal of Speech, Language, and Hearing Research • Vol. 57 • 2253–2260 • December 2014

Figure 1. Mean response time by group and by presentation rate for the word detection task. The group with typical language is the hatched line, indicating shorter response times. Error bars are 95% confidence intervals.

less likely explanation for our findings. The truth-value judgment task was presented only at a normal rate, however, so we are unable to rule out an effect of improved sentence processing efficiency in the slow rate condition for the group with SLI. The main effect of group accords with prior findings that adults with SLI are generally slower language processors (Miller & Poll, 2009). The finding that the group with SLI was less impacted by longer durations for holding the target word in mind is consistent with a slower lexical decay rate in SLI. Given the task of monitoring the sentences, participants had limited ability to rehearse or refresh the trace of the target word. Under such conditions, we would expect memory decay (Baddeley, 2000; Towse et al., 1998). A slower decay rate would result in a smaller decline in activation for the target word over time and a smaller increase in response time from the normal to slow condition. The main effect of rate is also consistent with a role for decay in working memory: The average participant response time was longer in the slow condition. Adults with SLI may be better adapted to slower rates. Children with SLI have been shown to benefit from slowed speech rates (Ellis Weismer & Hesketh, 1996; Montgomery, 2005). Faster speech rates are detrimental to language processing performance in children with SLI (Montgomery, 2005). Absent other factors, the shorter duration from encoding to recall in faster speech should help individuals with SLI if they have a faster decay rate, as suggested by some prior studies (Gillam et al., 1998; McMurray et al., 2010). Our findings consistent with a

slower lexical decay rate also are consistent with those of Seiger-Gardner and Schwartz (2008) and Montgomery (2005). Both Montgomery and the present study found that response times improved for individuals with SLI relative to peers with typical language when sentences were presented at a slowed rate. However, the adults in the present study were not helped as much by slowed speech as were the children in Montgomery’s study. Because children with SLI become faster processors as they grow older (Montgomery, 2005), adults with SLI are likely better adapted to normal speech rates than children with SLI. Our findings may also be interpreted as a difference in the ability of adults with SLI to manage interference (Oberauer et al., 2012). Attentional distractors using stimuli similar to what participants must remember produce interference. The attentional distractor in our task, monitoring the unfolding sentence, was therefore likely to have produced interference (Oberauer & Lewandowsky, 2013). The adults with SLI may have been less efficient at suppressing interference, so the slowed condition allowed them more time to suppress the interference of intervening words before the occurrence of the target. If the additional time afforded by the slow condition was more helpful to the group with SLI, this would account for the smaller group difference in response time in the slow condition. The fact that the slow rate resulted in a longer average response time for both groups of participants is less consistent with an interference account. The attentional account (Cowan, 2010) and tagging– retagging theory (Tarnow, 2008) are consistent with our

Poll et al.: Lexical Decay in Specific Language Impairment

2257

findings. If the target words were tagged items in long-term memory, Tarnow’s (2008) theory would predict that the degree of tagging (activation) declines as a function of time until the target word is encountered again. Recognition response times would increase as activation decreased. The finding that the slow rate condition resulted in generally longer response times aligns with the tagging–retagging view. If monitoring the unfolding sentence directed attention away from the target word, it also aligns with a view of working memory as the focus of attention, because items no longer in the focus of attention are subject to decay or interference (Cowan, 2010). By the tagging–retagging account, the smaller group difference in the slow rate condition suggests that activation for tagged items declines at different rates for each group. What these theoretical accounts share is that each suggests a difference for the group with SLI in a dynamic mechanism affecting verbal working memory capacity. There is considerable support for both more limited working memory capacity and slower processing as characteristics of SLI (Ellis Weismer et al., 1999; Miller et al., 2006; Montgomery & Evans, 2009). Working memory may mediate the role of slower processing in language processing (Fry & Hale, 2000; Poll et al., 2013). If an individual requires more time to process a task or linguistic item, the memory trace must be retained longer. This provides more opportunity for decay. Alternatively, the added time for processing leaves less time for suppressing interfering items. This line of reasoning makes an assumption that decay rate or interference sensitivity is consistent across language ability groups. These dynamic mechanisms may in fact differ between individuals with SLI and those with typical language, as suggested by this study and other converging findings. For example, suppression of irrelevant information is problematic in children with SLI (Im-Bolter et al., 2006; Spaulding, 2010), and such difficulties have been linked to working memory limitations (Marton, Kelmenson, & Pinkhasova, 2007).

Limitations There are limitations to this study that suggest caution before drawing conclusions for clinical practice. Our design did not allow us to clearly differentiate effects of slower decay rate from differences in interference suppression or tagging activation. A second limitation is our sentence processing efficiency measure: truth-value judgment. The task was presented only at the normal rate and involved picture interpretation in addition to auditory sentence comprehension. As a result, it was not an ideal task to rule out the use of sentence context to predict upcoming target words presented at different rates. Our study provides evidence for differences in dynamic mechanisms of working memory from one task. Given the conflicting evidence of the small number of studies on the topic, additional converging evidence is needed to draw stronger conclusions regarding the role of decay rate differences and interference in the profile of SLI. In

2258

future studies, researchers should consider ways to differentiate decay rate and interference-suppression explanations for relations between working memory and sentence processing in SLI. To arrive at clearer conclusions, researchers should also seek better controls for differences in sentence processing efficiency. Tasks must also vary in duration for retention of memoranda, to suppress rehearsal, and to maintain constant processing demands across conditions to support decay rate findings.

Clinical Implications and Conclusions Our finding of working memory limitations in SLI has important clinical implications (Boudreau & CostanzaSmith, 2011; Montgomery, Majimairaj, & O’Malley, 2008). Should additional research confirm the role of decay rate in working memory deficits in SLI, this will suggest new areas of focus for clinical practice. If decay rates are shown to be faster in SLI, for example, the implication is that representations of items held in working memory are too weak when they are needed to perform a cognitive task. In this scenario, the clinical strategy would be to focus on stronger encoding of memoranda, or on improving the maintenance of memoranda through rehearsal. However, if further evidence supports the hypothesis that lexical decay is slower in SLI, the clinical focus would logically turn to improving suppression of irrelevant information or improving resistance to interference. This study focused on whether memory traces may fade at a different rate for adults with SLI as compared to adults with typical language. We found that adults with SLI were less affected by an increased duration from encoding a target word to the recognition of that word later in a sentence. These findings support our hypothesis that adults with SLI have differences in the dynamic mechanisms of working memory and are consistent with slower decay rates for lexical items. We did not, however, eliminate interference and tagging-activation explanations for our findings. The roles of decay and interference have been debated in the research literature on individuals with typical language but have received little attention in the research on SLI. Should further studies confirm slower lexical decay in SLI, these findings would result in a revised—and more clearly specified—understanding of working memory limitations in the disorder.

Acknowledgments We thank Janet van Hell for supporting the development of this study. This study was supported by the National Institute on Deafness and Other Communication Disorders under Ruth L. Kirschstein National Research Service Award 1F31DC010960 (Gerard H. Poll, Principal Investigator). The views expressed are those of the authors and do not necessarily reflect any official position of the National Institutes of Health. This study was the basis of an honors thesis completed by the second author, who was also supported during development of this study by the Ronald McNair Scholars Program at The Pennsylvania State University. We also thank Marlea O’Brien, Connie Ferguson, Marcia St. Clair, and

Journal of Speech, Language, and Hearing Research • Vol. 57 • 2253–2260 • December 2014

Bruce Tomblin for their generous help with participant recruiting, and we thank the adults who participated in the study.

References Alt, M., & Spaulding, T. J. (2011). The effect of time on word learning: An examination of decay of the memory trace and vocal rehearsal in children with and without specific language impairment. Journal of Communication Disorders, 44, 640–654. doi:10.1016/j.jcomdis.2011.07.001 Altmann, E. M., & Gray, W. D. (2002). Forgetting to remember: The functional relationship between decay and interference. Psychological Science, 13, 27–33. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Science, 4, 417–423. doi:10.1016/S1364-6613(00)01538-2 Barrouillet, P., & Camos, V. (2001). Developmental increase in working memory span: Resource sharing or temporal decay? Journal of Memory and Language, 45, 1–20. doi:10.1006/ jmla.2001.2767 Boersma, P., & Weenink, D. (2006). Praat: Doing phonetics by computer (Version 4.4.30) [Computer software]. Amsterdam, the Netherlands: University of Amsterdam. Retrieved from www.praat.org Botting, N. (2005). Non-verbal cognitive development and language impairment. Journal of Child Psychology and Psychiatry, 46, 317–326. Boudreau, D., & Costanza-Smith, A. (2011). Assessment and treatment of working memory deficits in school-age children: The role of the speech-language pathologist. Language, Speech, and Hearing Services in Schools, 42, 152–166. doi:10.1044/01611461(2010/09-0088) Bowers, J. S., Vigliocco, G., Stadthagen-Gonzalez, H., & Vinson, D. (1999). Distinguishing language from thought: Experimental evidence that syntax is lexically rather than conceptually represented. Psychological Science, 10, 310–315. Clegg, J., Hollis, C., Mawhood, L., & Rutter, M. (2005). Developmental language disorders—A follow-up in later adult life. Cognitive, language, and psychosocial outcomes. Journal of Child Psychology and Psychiatry, 46, 128–149. Conti-Ramsden, G., & Durkin, K. (2012). Postschool educational and employment experiences of people with specific language impairment. Language, Speech, and Hearing Services in Schools, 43, 507–520. doi:10.1044/0161-1461(2012/11-0067) Cowan, N. (2010). Multiple concurrent thoughts: The meaning and developmental neuropsychology of working memory. Developmental Neuropsychology, 35, 447–474. doi:10.1080/ 875656412010494985 Davies, M. (2009). The 385+ million word corpus of Contemporary American English (1990–2008+). International Journal of Corpus Linguistics, 14, 159–190. de Renzi, E., & Faglioni, P. (1978). Normative data and screening power of a shortened version of the Token Test. Cortex, 14, 41–49. Ellis Weismer, S., Evans, J. L., & Hesketh, L. J. (1999). An examination of verbal working memory capacity in children with specific language impairment. Journal of Speech, Language, and Hearing Research, 42, 1249–1260. Ellis Weismer, S., & Hesketh, L. J. (1996). Lexical learning by children with specific language impairment: Effects of linguistic input presented at varying speaking rates. Journal of Speech and Hearing Research, 39, 177–190. Ferrill, M., Love, T., Walenski, M., & Shapiro, L.P. (2012). The time-course of lexical activation during sentence comprehension

in people with aphasia. American Journal of Speech-Language Pathology, 21, S179–S189. doi:10.1044/1058-0360(2012/ 11-0109) Fidler, L. J., Plante, E., & Vance, R. (2011). Identification of adults with developmental language impairments. American Journal of Speech-Language Pathology, 20, 2–13. doi:10.1044/ 1058-0360(2010/09-0096) (Erratum published in American Journal of Speech Language Pathology, 22, 577.) Francis, W., & Kučera, H. (1982). Frequency analysis of English usage: Lexicon and grammar. Boston, MA: Houghton Mifflin. Fry, A. F., & Hale, S. (2000). Relationships among processing speed, working memory and fluid intelligence in children. Biological Psychiatry, 54, 1–34. Gillam, R. B., Cowan, N., & Marler, J. A. (1998). Information processing by school-age children with specific language impairment: Evidence from a modality effect paradigm. Journal of Speech, Language, and Hearing Research, 41, 913–926. Im-Bolter, N., Johnson, J., & Pascual-Leone, J. (2006). Processing limitations in children with specific language impairment: The role of executive function. Child Development, 77, 1822–1841. Johnson, C. J., Beitchman, J. H., Young, A. R., Escobar, M., Atkinson, L., Wilson, B., . . . Wang, M. (1999). Fourteen-year follow-up of children with and without speech/language impairments: Speech/language stability and outcomes. Journal of Speech, Language, and Hearing Research, 42, 744–760. Jonides, J., Lewis, R. L., Nee, D. E., Lustig, C., Berman, M. G., & Moore, K. S. (2008). The mind and brain of short-term memory. Annual Review of Psychology, 59, 193–224. doi:10.1146/ annurev.psych.59.103006.093615 Leonard, L. B. (1998). Children with specific language impairment. Cambridge, MA: MIT Press. Leonard, L. B., Miller, C. A., & Finneran, D. A. (2009). Grammatical morpheme effects on sentence processing by schoolaged adolescents with specific language impairment. Language and Cognitive Processes, 24, 450–478. doi:10.1080/ 01690960802229649 Marton, K., Kelmenson, L., & Pinkhasova, M. (2007). Inhibition control and working memory capacity in children with SLI. Psychologia, 50, 110–121. Marton, K., Schwartz, R. G., Farkas, L., & Katsnelson, V. (2006). Effect of sentence length and complexity on working memory performance in Hungarian children with specific language impairment (SLI): A cross-linguistic comparison. International Journal of Language & Communication Disorders, 41, 653–673. doi:10.1080/13682820500420418 McMurray, B., Samelson, V. M., Lee, S. H., & Tomblin, J. B. (2010). Individual differences in online spoken word recognition: Implications for SLI. Journal of Memory and Language, 60, 1–39. doi:10.1016/j.cogpsych.2009.06.003 Miller, C. A., Kail, R., Leonard, L. B., & Tomblin, J. B. (2001). Speed of processing in children with specific language impairment. Journal of Speech, Language, and Hearing Research, 44, 416–433. Miller, C. A., Leonard, L. B., Kail, R., Zhang, X., Tomblin, J. B., & Francis, D. J. (2006). Response time in 14-year-olds with language impairment. Journal of Speech, Language, and Hearing Research, 49, 712–728. Miller, C. A., & Poll, G. H. (2009). Response time in adults with a history of language difficulties. Journal of Communication Disorders, 42, 365–379. doi:10.1016/j.jcomdis.2009.1004.1001 Montgomery, J. W. (2005). Effects of input rate and age on the real-time language processing of children with specific language impairment. International Journal of Language & Communication Disorders, 40, 171–188.

Poll et al.: Lexical Decay in Specific Language Impairment

2259

Montgomery, J. W., & Evans, J. L. (2009). Complex sentence comprehension and working memory in children with specific language impairment. Journal of Speech, Language, and Hearing Research, 52, 269–288. Montgomery, J. W., Majimairaj, B. M., & O’Malley, M. H. (2008). Role of working memory in typically developing children’s complex sentence comprehension. Journal of Psycholinguistic Research, 37, 331–354. doi:10.1007/s10936-008-9077-z Morice, R., & McNicol, D. (1985). The comprehension and production of complex syntax in schizophrenia. Cortex, 21, 567–580. Neville, H. J., Coffey, S. A., Holcomb, P. J., & Tallal, P. (1993). The neurobiology of sensory and language processing in language-impaired children. Journal of Cognitive Neuroscience, 5, 235–253. Oberauer, K., & Lewandowsky, S. (2013). Evidence against decay in verbal working memory. Journal of Experimental Psychology, 142, 380–411. doi:10.1037/a0029588 Oberauer, K., Lewandowsky, S., Farrell, S., Jarrold, C., & Greaves, M. (2012). Modeling working memory: An interference model of complex span. Psychonomic Bulletin and Review, 19, 779–819. doi:10.3758/s13423-012-0272-4 Pizzioli, F., & Schelstraete, M. (2011). Lexico-semantic processing in children with specific language impairment: The overactivation hypothesis. Journal of Communication Disorders, 44, 75–90. doi:10.1016/j.jcomdis.2010.07.004 Poll, G. H., Miller, C. A., Mainela-Arnold, E., Donnelly Adams, K., Misra, M., & Park, J. S. (2013). Effects of children’s working memory capacity and processing speed on their sentence imitation performance. International Journal of Language & Communication Disorders, 48, 329–342. doi:10.1111/1460-6984.12014 Psychology Software Tools. (2009). E-Prime (Version 2) [Computer software]. Pittsburgh, PA: Author. Purser, H. R. M., & Jarrold, C. (2005). Impaired verbal shortterm memory in Down syndrome reflects capacity limitation

2260

rather than atypically rapid forgetting. Journal of Experimental Child Psychology, 91, 1–23. doi:10.1016/j.jecp.2005.01.002 Sattler, J. M., & Ryan, J. J. (1999). Assessment of children: Revised and updated third edition WAIS–III supplement. La Mesa, CA: Jerome M. Sattler. Seiger-Gardner, L., & Schwartz, R. G. (2008). Lexical access in children with and without specific language impairment: A cross-modal picture–word interference study. International Journal of Language & Communication Disorders, 43, 528–551. Semel, E., Wiig, E., & Secord, W. A. (2003). Clinical Evaluation of Language Fundamentals, Fourth Edition. San Antonio, TX: The Psychological Corporation. Spaulding, T. J. (2010). Investigating mechanisms of suppression in preschool children with specific language impairment. Journal of Speech, Language, and Hearing Research, 53, 725–738. doi:10.1044/1092-4388(2009/09-0041) Tarnow, E. (2008). Response probability and response time: A straight line, the Tagging/Retagging interpretation of short term memory, an operational definition of meaningfulness and short term memory time decay and search time. Cognitive Neurodynamics, 2, 347–353. doi:10.1007/s11571-008-9056-y Tomblin, J. B., Zhang, X., Buckwalter, P., & O’Brien, M. (2003). The stability of primary language disorder: Four years after kindergarten diagnosis. Journal of Speech, Language, and Hearing Research, 46, 1283–1296. doi:10.1044/1092-4388(2003/ 100) Towse, J. N., Hitch, G. J., & Hutton, U. (1998). A reevaluation of working memory capacity in children. Journal of Memory and Language, 39, 195–217. van Casteren, M., & Davis, M. H. (2007). Match: A program to assist in matching the conditions of factorial experiments. Behavior Research Methods, 39, 973–978. Wechsler, D. (1997). Wechsler Adult Intelligence Scale—Third Edition. San Antonio, TX: The Psychological Corporation.

Journal of Speech, Language, and Hearing Research • Vol. 57 • 2253–2260 • December 2014

Copyright of Journal of Speech, Language & Hearing Research is the property of American Speech-Language-Hearing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.