Monday, February 29, 2016

Semantic and phonological contributions to short-term repetition and long-term cued sentence recall.

Meltzer, J.A., Rose, N.S., Deschamps, T., Leigh, R.C., Panamsky, L., Silberberg, A., Madani, N., & Links, K.S. (2016). Semantic and phonological contributions to short-term repetition and long-term cued sentence recall. Memory & Cognition, 44, 307-329.

Short-term memory(STM) involves the recall of information briefly held in mind. It is generally agreed that there is domain-specificity within STM such that recall of phonological information is impaired when a person is engaged in another phonological but not visuospatial task, and vice versa. It is equally well-accepted that phonological retention and recall is supported by activated semantic knowledge held in long-term memory (LTM). For example, better recall is observed for lists of known vs. unknown words, or sentences vs. strings of words reflecting the better retention of phonological strings with known semantic associations.

Another long-standing tenet of factors influencing encoding in LTM is the ‘level of processing’ of information (Craik & Lockhart, 1972). According to this view, better encoding (and recall) is achieved with deep vs. shallow processing of material presented. An example of shallow processing is making a superficial judgment about information such as the font a word is written or the phonological rhyme whereas deep processing involves semantic processing.

The authors of the present study argued that shifting participants towards greater reliance on semantic processing would enhance long-term recall of presented information even if the manipulation disrupted short-term recall. Participants completed a sentence repetition task as follows: A sentence was presented and during an initial short delay, the participant completed a phonological task or a non-phonological, finger tapping task. The phonological task was expected to disrupt phonological encoding and shift reliance towards semantic processing, whereas no such shift was expected for the finger tapping task. After 100 sentences were completed, participants completed a cued LTM task in which they were given pairs of two words that cued one of the sentences previously presented and asked to recall as much of the sentence as they could.

Not surprisingly, sentence recall from STM was lower when a phonological compared to nonphonological task was completed during a brief delay. Of greater interest was that the opposite pattern was observed for the LTM cued recall task: Relative to what they had remembered in the STM task, participants more accurately recalled sentences during which they had completed a phonological vs. nonphonological task at encoding. Further, sentences were either abstract or concrete with the idea that the abstract sentences would benefit more from the greater semantic processing achieved in the phonological disruption task. Indeed, the benefit to LTM recall after completing a phonological vs. nonphonological task at encoding was found to be greater for abstract than concrete sentences but only when the phonological task was demanding (i.e., counting backwards by 3 vs. repeating ‘babataka’).

These findings were taken as evidence that engaging semantic mechanisms at encoding to achieve a deeper level of processing promotes long-term retention. It would follow from this that rote immediate repetition is a poor strategy for promoting long-term learning as compared to strategies that engage learners with meaning.

Blogger: Lisa Archibald

Thursday, February 25, 2016

Structure of Phonological Ability at Age Four

Wolff, U. & Gusstafsson, J. E. (2015). Structure of phonological ability at age four. Intelligence, 53, 108-117.

Phonological awareness involves the ability to identify and manipulate speech sounds, and is an important predictor of reading abilities. There is debate as to whether phonological awareness is one-dimensional or is reflective of multiple abilities. Many studies have investigated this using phonological awareness tasks that vary in their linguistic and processing complexity. Linguistic complexity involves the size of the linguistic unit being manipulated, and can range from morphemes (meaningful units; e.g.: ball in snowball), to syllables, to phonemes (sound units; e.g.: /b/ sound in snowball). Processing complexity involves how much processing effort the tasks involves, and ranges from simple to complex. For instance, a simple task would involve judging whether two words rhyme, while a more complex task would involve deleting phonological units (e.g.: “Say spit without the /p/ sound”). Phonological awareness abilities have also been linked to a component of intelligence known as general fluid (Gf) intelligence. Little is known about the link between phonological awareness and Gf, but it is possible that Gf is linked to the processing complexity involved in certain phonological awareness tasks.

The present study used a range of tasks that varied in both their linguistic and processing complexity to examine phonological awareness in young children. Three hundred and sixty-four four-year-old Swedish children completed a battery of phonological awareness and cognitive tasks. Results from this test battery were analyzed using a structural equation modelling approach, which allowed the researchers to examine the relationship between observed task performance and underlying phonological awareness and cognitive abilities. They found that that phonological awareness is indeed one-dimensional, with variability in task performance coming from the distinct linguistic and processing demands. Additionally, Gf intelligence was related to the processing complexity of the phonological awareness task but not linguistic complexity, as expected. 

These results have important implications for designing tools to assess phonological awareness. It is important to consider both the linguistic and processing complexity of phonological awareness tasks used in assessments, and understanding that other cognitive abilities such as intelligence contribute to phonological awareness.   

Blogger: Nicolette Noonan
Nicolette is a Ph.D. student working with Drs. Lisa Archibald and Marc Joanisse

Monday, February 22, 2016

Rapid naming and phonemic awareness in children with reading disabilities and/or specific language impairment: differentiating processes?

De Groot, B. J. A., Van den Bos, K. P., Van der Meulen, B.F., Minnaert, A. E. M. G. (2015). Rapid naming and phonemic awareness in children with reading disabilities and/or specific language impairment: differentiating processes? Journal of Speech, Language, and Hearing Research, 58. 1538-1548.

Specific language impairment (SLI) and reading disability (RD) have higher rates of comorbidity than expected, and it remains unclear how the two disorders are related to one another. The current study examined whether decoding and recognition processes during reading can be differentiated in children with SLI with or without a comorbid RD. Phonemic awareness (PA), the ability to identify and manipulate speech sounds, is often considered to be a core deficit of RD, although studies have also observed PA deficits in children with SLI. Rapid automatized naming (RAN) involves naming arrays of objects, colours, letters, or numbers, and has been shown to be impaired in RD.
In the present study, groups of children between 8 and 13 years old with SLI, RD, and comorbid SLI and RD completed RAN and PA tasks to examine whether performance on both tasks would differ between groups. The RAN tasks included a letter array and a number array, and the PA tasks consisted of an elision task (i.e. say “string” without the “r”) and a substitution task (switch the first sounds in the words “red fish”). Performance was analyzed using a principal components analysis, which revealed two components explaining 85% of the variance: a RAN factor and a PA factor. The analyses revealed that the SLI-only group had low PA scores, the RD-only group had low scores on both RAN and PA, and the comorbid SLI and RD group had the most severely impaired performance on all the tasks. A similar pattern of results was demonstrated with a multivariate analysis of variance.

These findings are consistent with past research demonstrating RAN and PA impairments in children with RD.  Although some literature suggests that PA impairments may not be severe in children with SLI older than 9 years, the severe PA impairment observed in the present study’s SLI group may be associated with the large short-term memory load in the PA substitution task. The authors suggest that their data best fits a model of RD and SLI in which the two disorders are similar but distinct phonologically-linked disorders. The authors emphasize the importance of distinguishing between RD-only, SLI-only, and comorbid groups in research and in clinical practice, given that additive impairment effects were observed when SLI and RD occurred comorbidly.

Blogger: Alex Cross; Alex is a student in the combined speech language pathology/Ph.D. program supervised by Drs. Lisa Archibald and Marc Joanisse