Thoughts of a secondary school SLT

Tag: Verbal working memory

  • The cognitive causes of Developmental Language Disorder (DLD): Is DLD a language-based problem?

    In a previous article, I tried to summarise the causes of DLD from a biological perspective, focusing on the genetic and environmental factors as well as the neurological differences seen in those with the condition (See my post https://secondaryschoolslt.wordpress.com/2026/01/05/what-causes-language-disorders/). We have not yet spoken about the cognitive causes, in other words, the underlying mental processes that make learning language so difficult for some.

    Over the past few decades, numerous theories have been presented in an attempt to explain how language disorders come about. They can be split broadly into two categories: linguistic accounts and domain-general accounts.

    Linguistic theories of DLD are grounded in Noam Chomsky’s work in the late 50s and 60s on the acquisition of language, which revolutionized our thinking. Before him, people believed that children were a “blank slate”, and learnt  to talk through a mixture of imitation, trial and error and behavioural conditioning. Chomsky was the first to suggest that some aspects of language may be in-built.

    He argued that despite their surface level differences, all languages share a common structural foundation, which he called “Universal Grammar” (UG), and that this is innate. He also proposed the concept of a hypothetical “Language Acquisition Device” (or LAD), present from birth, which takes the linguistic input we receive and decodes it into a working grammar for the native language. (See his books, Syntactic Structures, and Aspects of the Theory of Syntax).

    Chomsky had noticed that young children learn grammar remarkably quickly, though the input they receive is invariably “messy” and incomplete. According to the “Poverty of Stimulus” argument, the language children are exposed to in their environment is not rich enough alone for them to learn all of the complex grammatical rules of their language, meaning that some linguistic knowledge must be hard-wired.

    In “The Language Instinct”, Steven Pinker (1994) built on Chomsky’s ideas, arguing that our brains are biologically adapted for language learning, supported by specialized circuitry that has been encoded into our DNA. Theoretically, then, DLD could arise from a faulty LAD, or in more modern terms, from some abnormality in the brain regions that support language.

    Some linguistic theories based on this foundation include the Extended Optional Infinitive (EOI) account of Rice and Wexler (1995) and the Representational Deficit in Dependent Relations (RDDR) hypothesis of Van der Lely (1998), which later evolved into the Computational Grammatical Complexity (CCG) Hypothesis (2005).

    According to the EOI account, all children go through a phase around the age of 5, when they inconsistently use grammatical morphemes that specify tense and agreement, because they believe them to be optional. So they may say things like “he walk” one day and “he walks” the next. Observing the persistent difficulties those with DLD have in this area, Rice and Wexler (1995) argued that this stage of development is greatly extended in those with the condition, leading to the problems that we see.

    Van der Lely also proposed that the difficulties of those with DLD arise out of a specific grammatical deficit, and that these children see certain rules as “optional”. Her first theory (Van der Lely, H., 1998) focused on common problems with sentences where one part relies on another, involving “syntactic movement”, such as passives or questions. However, she later expanded her theory (Van der Lely, 2005), arguing that a limited capacity for handling complex grammatical rules underlies many difficulties.

    Although linguistic accounts of language disorder may feel intuitive, they have been widely critiqued. Firstly, no single language gene has ever been found to account for most cases of DLD; nor is there any physical evidence of a LAD or specific language module in the brain. Instead, neuroimaging research has revealed that a widely distributed network of brain regions are involved in language processing, as well as a range of other tasks (Abbot and Love, 2023).

    Such theories also focus overwhelmingly on grammar, and they have been unable to account for the wide range of difficulties seen in DLD, such as the development of the lexicon, as well as more general cognitive deficits so commonly associated with the condition.

    As I mention in another post, children with language disorders have difficulties that extend beyond language: many also exhibit difficulties with verbal short term memory (VSTM), verbal working memory (VWM), processing speed, auditory perception and executive functioning skills. This has led to the development of more “domain-general” accounts of DLD.

    According to such accounts, a language disorder may not fundamentally be a language problem at all: instead, a more general deficit leads to a specific problem in language over time due to the skills that underlie language acquisition. Domain-general accounts of language disorder include the Capacity Theory of Comprehension (Just and Carpenter, 1992), the Generalised Slowing Hypothesis (Kail, 1994) and the Procedural Deficit Hypothesis (Ullman and Pierpont, 2005).

    Much attention has been drawn to the VSTM and VWM deficits so frequently exhibited by children with DLD, leading to the development of several, related theories. In particular, difficulties with nonword repetition (NWR – often used to assess VSTM) are so common as to be considered an indicator of the disorder (Bishop et al, 2016). When asked to repeat back made up words, most people only begin to struggle with words containing 4 or more syllables, whereas those with language disorders find those with 3 syllables challenging (Graf Estes et al, 2007).

    Poor performance on NWR tasks has been linked to a particular difficulty with the “phonological loop” in Baddeley and Hitch’s (1974, 2000) model of working memory. This component supposedly stores speech-based and verbal information for a short time before it decays or passes into long-term memory. Gathercole and Baddeley (1990) argued that a deficit in phonological storage could play a central role in DLD, particularly in the development of the lexicon.

    If you think about it, NWR is a task that effectively roleplays the learning of new words. Those who struggle to keep accurate representations of made up words or new words “active” in their minds for long enough may struggle to learn new words, or may need more repetitions to do so.

    The Capacity Theory of Comprehension, proposed by Just and Carpenter (1992) framed the working memory abilities associated with language processing more broadly, shifting attention away from the phonological loop, towards the lesser developed central executive of Baddeley and Hitch’s model (1974, 2000). Moving beyond the traditional view of working memory as merely a storage mechanism, they argued that simultaneous storage and processing of information is central to comprehension.

    Using a reading span task, Just and Carpenter (1992) assessed the working memories of college age students, before comparing their reading comprehension scores. Reading span tasks involve activities such as reading series of unrelated sentences before being asked to recall the final words, and they were designed to test simultaneous storage and processing of information. The researchers found few performance differences between students when task demands were low, but as they increased, those with poorer working memories began to struggle more with comprehension.

    Noting the “trade-offs” that occur between storage and processing, Just and Carpenter (1992) conceived of working memory as a shared pool of resources, or “activation” that mediates both abilities. Those with better working memories have a larger pool of resources to draw from, allowing them to process challenging sentences, whilst keeping different interpretations in mind, resolving ambiguities and integrating context.

    Those with poorer working memories, on the other hand, will reach their “capacity” sooner which could affect their understanding of sentences, especially more complex sentences, since representations constructed for the first part of the sentence will be lost by the time the sentence is complete. Weismer et al (1999) achieved comparable results in a study of children with Specific Language Impairment (SLI – now called DLD), compared to their age-matched controls.

    The Generalised Slowing Account proposed by Robert Kail (1994) drew attention to another cognitive deficit commonly seen in this population: slow processing speed. Individuals with language disorders have been found to have slower response times across a range of domains (Schul et al, 2004), which could theoretically cause difficulties with language processing on a number of levels.

    Since language is ephemeral and disappears as soon as it has been spoken, those who process language more slowly may struggle to keep up with the rapid speech stream. According to this theory, it is less about students having smaller working memories, and more about the information decaying before they have had a chance to process and store it. Or, to put it another way, if your brain takes too long to process the first half of a complex sentence, then it won’t be ready for the second half.

    Other researchers still have highlighted the auditory perception and processing difficulties present in a subset of children with language disorders. Auditory perception refers to the brain’s ability to interpret, organize and distinguish different sounds. Whilst originally understood as a difficulty processing brief, rapid sounds, more recent research has indicated greater difficulty discriminating sounds that are similar in frequency (McArthur and Bishop, 2004).

    According to such theories, a difficulty with auditory perception could lead to unstable phonological representations, making it harder for children to form a stable foundation for speech and literacy. Since many grammatical morphemes in English are also brief, unstressed and high frequency, e.g. “s” and “ed”, children with DLD may miss these sounds, affecting their learning of grammar.

    These domain general accounts have also received their fair amount of criticism. For a start, some of these cognitive deficits such as auditory processing difficulties have only been found in a subset of children with the disorder. Conversely, other children with such difficulties do not go on to display disordered language. Meanwhile, theories of working memory have come under scrutiny by those who argue that such difficulties may be the result rather than the cause of language disorders.

    More generally speaking, domain-general accounts also fail to explain why children with DLD struggle specifically with language-related tasks, and less so with other tasks that rely on the same underlying skills. Attempts to remedy language disorders by treating these underlying deficits have also yielded mixed results. Finally, there has also been some debate around whether some of the tests used in developing these theories, such as the NWR test actually test what they claim to be testing, or whether other abilities are also involved.

    More recent research has pointed to a different kind of memory difficulty. According to Ullman and Pierpont’s (2005) Procedural Deficit Hypothesis (PDH), the difficulties present in DLD as well as a range of other disorders can be explained by abnormalities in the brain that make up the procedural memory system.

    As the name suggests, procedural memory involves the learning of skills through repeated practice, without conscious thought. It is how we learn to ride a bike, and we think that this is how most typically developing individuals learn their mother tongue, especially grammar. Declarative memory, on the other hand involves conscious learning of facts and knowledge, and may be involved in vocabulary learning.

    In a meta-analysis, Lum et al (2014) concluded that SLI (now referred to as DLD) is indeed associated with procedural learning deficits. Neuroimaging studies have also revealed abnormalities in regions of the brain associated with procedural learning in those affected, such as the caudate nucleus and inferior frontal gyrus (Abbott and Love, 2023).

    In a later article, Ullman and Pullman (2016) proposed further that individuals with neurodevelopmental disorders may use declarative memory systems to compensate for their procedural memory deficits. In a study of primary school children, Conti-Ramsden et al (2015) found that procedural learning ability predicted grammatical understanding in typically developing children. By contrast, declarative memory significantly predicted grammatical abilities in those with SLI (now DLD), supporting the notion of relatively intact declarative memory systems being used as a compensatory tool in those affected.

    Although this last theory could be considered a more comprehensive account of language disorders that makes connections between brain and behaviour, it too has been criticized for being too broad. If procedural deficits underlie several different neurodevelopmental disorders such as DLD, dyslexia and Autistic Spectrum Disorder (ASD), then what causes one child to end up with one problem, and a different child to end up with another?

    Ultimately, DLD is a heterogenous condition that is unlikely to have a single cause. Whilst none of these theories may be able to completely satisfy the question of how language disorders arise, perhaps each adds a bit to our understanding of DLD. However, until a clearer consensus is reached, supporting our students with their language behaviours rather than the underlying mechanisms is still the best way to help them.

    Notes

    Abbott, Noelle, and Tracy Love. “Bridging the Divide: Brain and Behaviour in Developmental Language Disorder.” Brain Sciences 13, no. 11 (2023): 1606.

    Baddeley, Alan D., and Graham J. Hitch. 1974. “Working Memory.” In The Psychology of Learning and Motivation: Advances in Research and Theory, edited by Gordon H. Bower, Vol. 8, 47–89. New York: Academic Press.

    Baddeley, A.D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, 417-423.

    Bishop DVM, Snowling MJ, Thompson PA, Greenhalgh T, CATALISE consortium (2016)

    CATALISE: A Multinational and Multidisciplinary Delphi Consensus Study. Identifying Language Impairments in Children. PLoS ONE 11(7): e0158753. doi:10.1371/journal.pone.0158753

    Conti-Ramsden, Gina, Michael T. Ullman, and J. A. G. Lum. “The Relation between Receptive Grammar and Procedural, Declarative, and Working Memory in Specific Language Impairment.” Frontiers in Psychology 6 (2015): 1090.

    Ellis Weismer, Susan, Julia L. Evans, and Jill E. Hesketh. “An Examination of Verbal Working Memory Capacity in Children With Specific Language Impairment.” Journal of Speech, Language, and Hearing Research 42, no. 5 (October 1999): 1249-1260. [1, 2]

    Gathercole, Susan E., and Alan D. Baddeley. “Phonological Memory Deficits in Language Disordered Children: Is There a Causal Connection?” Journal of Memory and Language 29, no. 3 (1990): 336-360.

    Graf Estes, Kathryn, Julia L. Evans, and Nicole M. Else-Quest. “Differences in the Nonword Repetition Performance of Children With and Without Specific Language Impairment: A Meta-Analysis.” Journal of Speech, Language, and Hearing Research 50, no. 1 (February 2007): 177–195.

    Just, Marcel A., and Patricia A. Carpenter. “A Capacity Theory of Comprehension: Individual Differences in Working Memory.” Psychological Review 99, no. 1 (1992): 122–149.

    Kail, Robert. “A Method for Studying the Generalized Slowing Hypothesis in Children With Specific Language Impairment.” Journal of Speech and Hearing Research 37, no. 4 (1994): 818–31.

    Lum, J. A. G., G. Conti-Ramsden, A. T. Morgan, and M. T. Ullman. “Procedural Learning Deficits in Specific Language Impairment (SLI): A Meta-Analysis of Serial Reaction Time Task Performance.” Cortex 51 (February 2014): 1–10.

    McArthur, Genevieve M., and Dorothy V. M. Bishop. “Which People with Specific Language Impairment Have Auditory Processing Deficits?” Cognitive Neuropsychology 21, no. 1 (2004): 79–124.

    Pinker, Steven. 1994. The Language Instinct: How the Mind Creates Language. New York: William Morrow.

    Rice, Mabel L., Kenneth Wexler, and Patricia L. Cleave. “Specific Language Impairment as a Period of Extended Optional Infinitive.” Journal of Speech and Hearing Research 38, no. 4 (August 1995): 850–63.

    Schul, R., J. Stiles, B. Wulfeck, and J. Townsend. 2004. “How ‘generalized’ is the ‘slowed processing’ in SLI? The case of visuospatial attentional orienting.” Neuropsychologia 42 (5): 661-671.

    Ullman, Michael T., and E.I. Pierpont. 2005. “Specific Language Impairment Is Not Specific to Language: The Procedural Deficit Hypothesis.” Cortex 41 (3): 399–433. [1, 2, 3]

    Ullman, Michael T., and Mariel Y. Pullman. “A Compensatory Role for Declarative Memory in Neurodevelopmental Disorders.” Neuroscience & Biobehavioral Reviews 51 (2015): 205–222.

    Van der Lely, Heather K. J. 1998. “SLI in Children: Movement, Economy, and Deficits in the Computational-Syntactic System.” Language Acquisition 7 (2–4): 161–92.

    Van der Lely, Heather K. J. 2005. “Domain-specific cognitive systems: insight from Grammatical-SLI.” Trends in Cognitive Sciences 9, no. 2 (February): 53–59.

  • Is poor working memory the cause of comprehension difficulties in older students with language disorders?

    Students with language disorders have difficulties that extend beyond language. There is an extensive research literature linking weaknesses in broader cognitive skills such as attention, processing speed, executive functioning, short term and working memory with language disorders (Leonard et al, 2007, 2013, Henry and Botting, 2017).

    Short term and working memory have received particular attention. Whilst some studies have identified weaknesses in the nonverbal component, suggestive of domain-general impairments in this population, deficits in verbal short term memory (VSTM) and verbal working memory (VWM) have been more consistently reported, and found to be 2-3 times larger (Vugs et al, 2013). What’s more, poor performance on nonword repetition tasks, widely used to test VSTM, is even considered a clinical marker for Developmental Language Disorder (DLD) (Bishop et al, 2016).

    VSTM and VWM are often used interchangeably, but VSTM might be more accurately considered a component of VWM. VSTM refers to the ability to hold information just heard in mind for a short period of time before it “decays” and has historically been thought of as a “storage space” with limited capacity. Indeed, most people are only able to keep three or four chunks of information in their heads at once (Montgomery et al, 2021).

    Meanwhile, VWM involves manipulation as well as storage of information, and is increasingly being thought of more as a “mental workspace”. Others have spoken of VWM in terms of a set of cognitive processes that include sustained attention, inhibition of irrelevant information, and the ability to switch simultaneously between maintenance of stored information and processing new information (Marton et al, 2007, Leonard et al, 2013).

    Various theories of working memory have been proposed in the past, one of the most influential being Baddeley and Hitch’s multicomponent model, made up of two passive storage systems: the “phonological loop” (or VSTM) and “visuo-spatial sketchpad”, as well as a “central executive” and “episodic buffer” (Baddeley and Hitch, 1974, Baddeley, 2000).

    According to this, the phonological loop stores speech-based and verbal information (and could be considered what we refer to as VSTM). Whilst information typically fades away after a couple of seconds, it is possible to keep it in an active state for longer through silent repetition (such as when you repeat a phone number or code to yourself). The visuospatial sketchpad stores visual information in a similar fashion, whilst the central executive acts as the control centre, dividing and switching attention between different tasks. Meanwhile, the episodic buffer binds the information together, and acts as an interface between short and long-term memory.

    VSTM and VWM are considered essential to learning in the classroom, from following lengthy instructions and understanding what’s going on in lessons, to keeping the steps of a task in mind and recalling the details of a story. Working memory has been found to be a more powerful predictor of academic achievement even than IQ (Alloway and Alloway, 2010). Students with poor VWM may appear inattentive, forgetful or careless, when really they are struggling to retain what was said.

    VWM is also considered to be intrinsically linked to comprehension of complex sentences, since certain elements have to be kept in mind, and even moved around whilst the next part is processed. Long, complex sentences require more processing time than simple sentences, and are not as easily understood (e.g. Marton et al, 2007), suggesting that increased VWM capacity is required to understand such sentences. Montgomery et al (2009) argued that comprehension of both simple and complex grammar is a mentally demanding task for school age children with and without language disorders that requires significant working memory resources.

    Other authors such as Balthazar and Scott (2023) have spoken at length about the various elements that can increase the processing “load” of a sentence. This includes the number of clauses, long distance dependencies (gaps), as well as (in English) anything that disrupts the subject-verb-object order such as passive constructions and post-modification of the noun phrase (see my post for more information).

    Take the following example sentence given by Marilyn Nippold (2010) from a science text book: “organisms that eat living corals, such as the crown-of-thorns sea star, can greatly damage reefs”. In this sentence, post-modification of the noun “organisms” with the phrase “that eat living corals, such as the crown-of-thorns sea star” results in an extended gap between the main subject and verb. This whole phrase must then be stored in VSTM (or the phonological loop according to the Baddeley model) until the reader reaches the main action, “can greatly damage” and understands what the sentence is about.

    For a student with a language disorder and limited VWM, this is likely to be challenging. Other studies have shown that adolescents with DLD do not understand complex sentences as well as their typically developing peers, despite similar performance for simple sentences (Montgomery et al, 2009). Some researchers have even gone so far as to argue that VWM difficulties, rather than poor language knowledge are the primary cause of receptive language difficulties in older students (Larson and McKinley, 2003).

    However, recent research suggests that the reality may be more complex. Indeed, there is some evidence to suggest that complex sentences are processed in different ways by those with and without language disorders, with working memory playing an unequal role in each. In a large scale study of 117 children with DLD and 117 typically developing peers aged 7-11, Montgomery et al (2021) investigated how a range of measures were connected with comprehension of simple and more complex sentences.

    They found that fluid reasoning and language knowledge residing in long-term memory (LTM) indirectly influenced comprehension of complex, non-canonical sentences in typically developing students. For those with DLD, on the other hand, controlled attention (an important facet of working memory), was more important (Montgomery et al, 2021).

    On the back of this research, they came up with a new memory model, the GEM (Gillam-Evans-Montgomery) model, where VWM serves as a conduit for fluid reasoning, controlled attention and long-term language knowledge. They argued that listeners face the challenge of a rapid incoming stream of speech in different ways (Montgomery et al, 2021).

    According to them, repeated experience with language allows most people to build up linguistic representations in long-term memory (LTM). Typically developing listeners are able to activate these patterns, some of which may take the form of multiword templates, to anticipate the types of words that are likely to come next, as well as to “chunk” the speech stream into noun phrases, verb phrases and even whole clauses. This information is then stored as chunks, reducing the demands on working memory capacity, before being reintegrated into a coherent whole (Montgomery et al, 2021).

    They hypothesized that students with DLD may have weaker, or non-existent representations of certain grammatical structures. This means that they will be unable to segment the speech stream in the same way as their peers, resulting in word by word processing which places enormous pressure on an already overstretched VWM. Accordingly, sentence processing is much more effortful for those with language disorders (Montgomery et al, 2021).

    This theory seems to be backed up by a review of the literature. Karavasilis et al (2023) found inconclusive evidence of a link between VSTM/ VWM and complex sentence comprehension in typically developing individuals. On the other hand, there was a consistent link between working memory and sentence comprehension in those with DLD. The authors concluded that, at least for children with DLD, a processing component is involved in comprehension of complex sentences.

    According to Montgomery et al (2021), the solution is not to attempt to improve students’ VWMs (which, in any case, has had limited success), but rather, to support their language representations in long-term memory.

    Notes

    Alloway, Tracy Packiam, and Ross G. Alloway. “Investigating the Predictive Roles of Working Memory and IQ in Academic Attainment.” Journal of Experimental Child Psychology 106, no. 1 (May 2010): 20–29.

    Baddeley, Alan D., and Graham J. Hitch. 1974. “Working Memory.” In The Psychology of Learning and Motivation: Advances in Research and Theory, edited by Gordon H. Bower, Vol. 8, 47–89. New York: Academic Press.

    Baddeley, A.D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, 417-423.

    Baddeley, Alan. 2003. “Working Memory: Looking Back and Looking Forward.” Nature Reviews Neuroscience 4, no. 10: 829–39.

    Balthazar, Catherine H., and Cheryl M. Scott. “Sentences Are Key.” American Journal of Speech-Language Pathology 33, no. 2 (2023): 564–579.

    Bishop DVM, Snowling MJ, Thompson PA, Greenhalgh T, CATALISE consortium (2016)

    CATALISE: A Multinational and Multidisciplinary Delphi Consensus Study. Identifying Language Impairments in Children. PLoS ONE 11(7): e0158753. doi:10.1371/journal.pone.0158753

    Haebig, Eileen, Christine Weber, Laurence B. Leonard, Patricia Deevy, and J. Bruce Tomblin. “Neural Patterns Elicited by Sentence Processing Uniquely Characterize Typical Development, SLI Recovery, and SLI Persistence.” Journal of Neurodevelopmental Disorders 9, no. 1 (2017): 22.

    Henry, L. & Botting, N. (2017). Working memory and developmental language impairments. Child Language Teaching and Therapy, 33(1), pp. 19-32.

    Karavasilis, Gavriil, K. Diakogiorgi, and D. Papadopoulou. 2023. “The Role of Working Memory in the Comprehension of Syntactically Complex Sentences in Children with and without Developmental Language Disorder: A Literature Review.” Psychology: The Journal of the Hellenic Psychological Society 28 (2): 205–222.

    Larson, V.L. and McKinley, N.L. (2003) Communication solutions for older students. Thinking Pub. 

    Leonard, Laurence B., Patricia Deevy, James W. Miller, Chrystal Rameela, Robert Schwartz, and J. Bruce Tomblin. “Speed of Processing, Working Memory, and Language Impairment in Children.” Journal of Speech, Language, and Hearing Research 50, no. 2 (April 2007): 408–428. 

    Leonard, Laurence B., Patricia Deevy, Marc E. Fey, and Shelley L. Bredin-Oja. “Sentence Comprehension in Specific Language Impairment: A Task Designed to Distinguish between Cognitive Capacity and Syntactic Complexity.” Journal of Speech, Language, and Hearing Research 56, no. 2 (April 2013): 577-589.

    Marton, Klara, Richard G. Schwartz, Lajos Farkas, and Valeriya Katsnelson. “Effect of Sentence Length and Complexity on Working Memory Performance in Hungarian Children with Specific Language Impairment: A Cross-Linguistic Comparison.” International Journal of Language & Communication Disorders 42, no. 6 (2007): 691–711.

    McCauley, Stewart M., and Morten H. Christiansen. 2015. “Individual Differences in Chunking Ability Predict On-line Sentence Processing.” In Proceedings of the 37th Annual Conference of the Cognitive Science Society, edited by D. C. Noelle et al., 1550–1555. Austin, TX: Cognitive Science Society. 

    Montgomery, James W., and Julia L. Evans. 2009. “Complex Sentence Comprehension and Working Memory in Children with Specific Language Impairment.” Journal of Speech, Language, and Hearing Research 52, no. 2 (April): 269-288.

    Montgomery, James W., Ronald B. Gillam, and Julia L. Evans. “A New Memory Perspective on the Sentence Comprehension Deficits of School-Age Children With Developmental Language Disorder: Implications for Theory, Assessment, and Intervention.” Language, Speech, and Hearing Services in Schools 52, no. 2 (April 2021): 449–466.

    Newman, Sharlene D., Evie Malaia, Roy Seo, and Hu Cheng. “The effect of individual differences in working memory capacity on sentence comprehension: an fMRI study.” Brain and Language 125, no. 3 (2013): 269-277.

    Nippold, Marilyn A. 2010. “Back to School: Why the Speech-Language Pathologist Belongs in the Classroom.” Language, Speech, and Hearing Services in Schools 41 (4): 377–378.

    Vugs, B., Cuperus, J., Hendriks, M., & Verhoeven, L. (2013). Visuospatial working memory in specific language impairment: A meta-analysis. Research in Developmental Disabilities, 34(9), 2596-2597.