Recent research has shown that the early lexical representations children establish in their second year of life already seem to be phonologically detailed enough to allow differentiation from very similar forms. In contrast to these findings children with specific language impairment show problems in discriminating phonologically similar word forms up to school age. In our study we investigated the question whether there would be differences in the processing of phonological details in normally developing and in children with low language performance in the second year of life. This was done by a retrospective study in which in the processing of phonological details was tested by a preferential looking experiment when the children were 19 months old. At the age of 30 months children were tested with a standardized German test of language comprehension and production (SETK2). The preferential looking data at 19 months revealed an opposite reaction pattern for the two groups: while the children scoring normally in the SETK2 increase their fixations of a pictured object only when it was named with the correct word, children with later low language performance did so only when presented with a phonologically slightly deviant mispronunciation. We suggest that this pattern does not point to a specific deficit in processing phonological information in these children but might be related to an instability of early phonological representations, and/or a generalized problem of information processing as compared to typically developing children.

This Thesis explores the development of word-initial s+plosive consonant clusters in the speech of Swedish children between the ages of 1;6 and 4;6. Development in the word-initial consonant clusters is viewed as being determined by 1) the children’s ability to articulate the target sequence of consonants, 2) the level of understanding of which acoustic features in the adult model production are significant for the signalling of the intended distinction, and 3) the children’s ability to apply established production patterns only to productions where the acquired feature agrees with the adult target, to achieve a contrast between rival output forms. This Thesis employs a method where output forms are contrasted with attempted productions of potential homonym target words. Thus, development is quantified as an increase in the manifestations of phonetic features where it agrees with the adult norm, coupled by a decrease in the same feature in output forms where it is inappropriate according to the specifications of the phonological system of the ambient language. Acoustic investigations of cues of voicing, aspiration, place of articulation and syllable onset complexity, and auditory investigations of place, manner and syllable onset complexity were conducted. The Thesis has four outcomes. One, a description of the perceptual quality of the productions in terms of place, manner, voicing and syllable onset complexity is presented. Two, a developmental sequence of stable acquisition of these features is proposed; manner is shown to be acquired first, followed by syllable onset complexity and place of articulation. Evidence is provided that the voiced/aspirated distinction is still being acquired at the end of the investigated age period. Three, the developmental use of acoustic cues of place and voicing are described. Voice Onset Time and Spectral Skewness are shown to be used by children in order to increase the likeness to the adult target in terms of voicing and place of articulation. Aspiration Amplitude is shown to be used as an auxiliary cue to Voice Onset Time. The place cues Spectral Tilt Change, F2, Spectral Mean and Spectral Variance were shown to be used in order to refine already produced consonants rather than approach the adult target model. Four, the Thesis provides evidence of periods of confusions in the output of children. With the reductions of these patterns of confusion, evidence is provided of children’s re-organisation of their internal representation of the consonant to be produced. / Denna avhandling undersöker utvecklingen av ordinitiala konsonantkluster av formen s+klusil i talet hos svenska barn mellan åldrarna 1;6 och 4;6. Utvecklingen av de ordinitiala klustren betraktas som bestämd av 1) barnets förmåga att artikulera den konsonantsekvens som utgör målet, 2) barnets förståelse för vilka akustiska särdrag i den vuxna målproduktionen som är signifikanta för att signalera en viss distinktion och 3) barnets förmåga att tillämpa ett etablerat produktionsmönster endast på de produktioner där det tillägnade draget överensstämmer med den vuxna målproduktionen, så att en kontrast uppnås mellan konkurrerande utformer. Avhandlingen tillämpar en metod där producerade utformer kontrasteras med produktioner av målord som utgör potentiella homonymer till dessa. Utvecklingen kan då kvantifieras som en ökning av antalet förekomster av ett fonetiskt drag som överensstämmer med den vuxna normen för den relevanta kontexten, kopplad till en minskning av antalet förekomster av samma drag i kontext där draget är inkorrekt givet det fonologiska systemet i det språk som tillägnas. De drag som undersöktes var de akustiska korrelaten till stämbandston, aspiration, artikulationsställe och komplexitet i stavelseansatsen, och vidare de auditiva korrelaten till artikulationsställe, artikulationssätt och komplexitet i stavelseansatsen. Fyra resultat redovisas. För det första presenteras en beskrivning av den perceptuella kvaliteten hos barnens produktioner i termer av artikulationsställe, artikulationssätt och komplexitet. För det andra föreslås en utvecklingssekvens för stabilt tillägnande av dessa drag: artikulationssätt tillägnas först, följt av komplexitet hos stavelseansatsen och artikulationsställe. Sist tillägnas distinktionen mellan tonande och aspirerad klusil, där data visar att draget inte tillägnats till fullo vid slutet av det undersökta åldersintervallet. För det tredje ger avhandlingen en beskrivning av hur barnen utvecklar sitt bruk av akustiska korrelat till artikulationsställe och ton. Voice Onset Time och snedhet i spektrum används för att närma sig det vuxna målet gällande aspiration och artikulationsställe. Aspirationsamplitud används som ett hjälpkorrelat till VOT vad gäller fonetisk aspiration. Andra korrelat för artikulationsställe, som förändring i spektral lutning, F2, spektralt medelvärde och spektral varians, används för att förfina de egna produktionerna snarare än för att få produktionen att närma sig det vuxna målet. För det fjärde ges i avhandlingen evidens för perioder då barnens produktion uppvisar förväxlingar vad avser distributionen hos vissa fonetiska drag. Minskade förekomster av sådana förväxlingar visar på en omstrukturering av barnens underliggande representation för den konsonant som ska produceras.

Neural networks underpinning working memory demonstrate sign language specific components possibly related to differences in temporary storage mechanisms. A processing approach to memory systems suggests that the organisation of memory storage is related to type of memory processing as well. In the present study, we investigated for the first time semantic, phonological and orthographic processing in working memory for sign- and speech-based language. During fMRI we administered a picture-based 2-back working memory task with Semantic, Phonological, Orthographic and Baseline conditions to 11 deaf signers and 20 hearing non-signers. Behavioural data showed poorer and slower performance for both groups in Phonological and Orthographic conditions than in the Semantic condition, in line with depth-of-processing theory. An exclusive masking procedure revealed distinct sign-specific neural networks supporting working memory components at all three levels of processing. The overall pattern of sign-specific activations may reflect a relative intermodality difference in the relationship between phonology and semantics influencing working memory storage and processing.

The present study aims to investigate the phonological development of typically developing monolingual Swedish speaking children aged three to four years. The result could be used in the standardization of a new phonological test material for clinical speech and language pathology assessments. One-hundred and thirty four children aged three and four years (73 girls and 61 boys) were assessed with the new material. The children were tested in kindergartens in three communities in southeastern Sweden. Their assessments were audio recorded, transcribed phonetically and analyzed with Percentage Phonemes Correct (PPC) and Percentage of Words Correct (PWC). An analysis of speech error patterns of substitutions of phonemes, reduction of consonant clusters and word structural deviations was conducted. The data recorded from the children were divided and analyzed in four semi-annual intervals and two annual intervals. The result shows a clear developmental trend with children in the older age group having a higher PPC and PWC than the younger children. Significant differences were found between the groups in annual intervals. Significant differences could only be demonstrated between the youngest age group and the other groups at semi-annual intervals. No significant gender differences were observed. Most of the errors of individual phonemes and consonant clusters among children in all age groups were distortions rather than substitutions, but the younger children simplified consonant clusters more often.

The current research examined whether nonword repetition (NWR) tasks, designed to measure phonological short-term memory, are also influenced by familiarity with lexical representation of a given language. In Study 1, children with and without exposure to Hebrew were administered a NWR task based on the Hebrew language structure (HNWR). On the HNWR, participants with Hebrew exposure significantly outperformed participants who had no familiarity with any Semitic language. This indicates that long-term phonological and lexical knowledge can be used to aid NWR performance. Study 2 investigated whether a NWR task based on a foreign language could minimize the lexicality effect. English speaking undergraduate students rated the less familiar HNWR task to be lower in wordlikeness than English-like NWR tasks. These findings demonstrate that regardless of language background a NWR task based on an unfamiliar language structure is a more valid measure of the phonological processing skills required for vocabulary acquisition.

The current research examined whether nonword repetition (NWR) tasks, designed to measure phonological short-term memory, are also influenced by familiarity with lexical representation of a given language. In Study 1, children with and without exposure to Hebrew were administered a NWR task based on the Hebrew language structure (HNWR). On the HNWR, participants with Hebrew exposure significantly outperformed participants who had no familiarity with any Semitic language. This indicates that long-term phonological and lexical knowledge can be used to aid NWR performance. Study 2 investigated whether a NWR task based on a foreign language could minimize the lexicality effect. English speaking undergraduate students rated the less familiar HNWR task to be lower in wordlikeness than English-like NWR tasks. These findings demonstrate that regardless of language background a NWR task based on an unfamiliar language structure is a more valid measure of the phonological processing skills required for vocabulary acquisition.

Multiple domains of deficit have been proposed to account for the apparent reading failure of children with a reading disability. Deficits in both phonological awareness and rapid automatized naming are consistently linked with the development of a reading disability in young school age children. Less research, however, has sought to connect these two reading related processes to global theories of deficit, such as temporal processing deficits, in the explanation of reading fluency difficulties. This study sought to explore the relationship between aspects of temporal processing, as indexed through measures of motor fluency and control, and measures of reading related processes, phonological awareness and rapid automatized naming, to word reading fluency. Using structural equation modeling, measures of patterned motor movement were found to be negatively and significantly related to measures of phonological awareness. Measures of oral and repetitive movement were found to be positively and significantly related to measures of patterned movement. Finally, phonological awareness was found to be a significant predictor of word reading fluency both independently and through rapid automatized naming. No direct relationship between measures of motor control and fluency and word reading fluency was found. These findings suggest that temporal processing, as indexed by measures of motor fluency and control, are moderately predictive of the facility with which a child with a reading disability can access, manipulate, and reproduce phonetically based information. Implications for the inclusion of motor based measures in the assessment of children with reading disabilities and future directions for research are discussed.

Deficiencies in mathematics skill constrain students’ educational achievement and subsequently, their employment outcomes. This study included 265 school-identified students with mild intellectual disabilities. The research questions investigated the extent to which phonological awareness, color naming speed, and vocabulary knowledge, was related to mathematics skill after controlling for grade level via regression analyses. Further, the mediating effects of expressive vocabulary on the relationship between receptive vocabulary and mathematics skill as well as the indirect effect of receptive vocabulary knowledge on mathematics skill through expressive vocabulary were examined. The findings indicated that after controlling for grade level, phonological awareness, naming speed, and vocabulary knowledge were significantly related to mathematics skill. The mediating effects of expressive vocabulary as well as the indirect effects of receptive vocabulary knowledge on mathematics skill were also significant.

Speech production requires cooperation between cognitive, linguistic and motor processes. It also requires spatial and temporal control of muscles, as well as simultaneous and coordinated activity of respiration, phonation and articulation (Cheng, Murdoch, Goozée & Scott, 2007; Yorkston, Beukelman, Strand & Bell, 1999; Raphael, Borden & Harris, 2011). Voice Onset Time (VOT) reflects the timing between phonation and articulation (Hoit-Dalgaard, Murry & Kopp, 1983). VOT is the most reliable acoustic cue for distinguishing between voiceless or voiced plosives (Auzou et al. 2000). Studies of English-speaking children with phonological impairment have shown atypical VOT-patterns (Bond & Wilson, 1980). The aim of the present study was to investigate Voice Onset Time (VOT) among Swedish children with phonological impairment and to examine if their VOT-values differ from typically developed Swedish children. Participants were 38 children aged 4;2−11;6 distributed over eight age- groups and five developmental stages of phonology. Audio recordings of minimal pairs were made at preschools, schools or at speech pathology clinics. The results indicated that children with phonological impairment produced VOT with deviant values and with a great variability. A marked acoustic difference between voiceless and voiced stop consonants was present, but not in all cases. Since the VOT-values were distributed over the group of children with phonological impairment, no developmental trend toward adult-like values that could be related to increasing age was found for either the acquisition of producing VOT or the acquisition of producing voicing lead. No differences in VOT were seen between the children in different phonological developmental stages or ages. No correlation between the degree of deviance of VOT and the proportion of Procent Phonemes Correct (PPC), age or phonological processes were found. From the results the conclusion can be drawn that children with phonological impairment have deviant VOT-values that could be caused by lack of phonological knowledge, but in particular since the variability did not decrease with increased age, have difficulties with motor execution. / Tal kräver ett samarbete mellan kognitiva, språkliga och motoriska processer. Det kräver även spatial och temporal kontroll av muskler samt samtidig och koordinerad aktivitet av andning, fonation och artikulation (Cheng, Murdoch, Goozée & Scott, 2007; Yorkston, Beukelman, Strand & Bell, 1999; Raphael, Borden & Harris, 2011). Voice Onset Time (VOT) ger en bild av koordinationen mellan fonation och artikulation (Hoit-Dalgaard, Murry & Kopp, 1983). VOT är den mest pålitliga akustiska referensen för att kunna skilja mellan tonande och tonlösa klusiler (Auzou et al., 2000). Engelsktalande barn med fonologisk språkstörning har visat atypiska VOT-mönster (Bond & Wilson, 1980). Syftet med föreliggande studie är att undersöka Voice Onset Time (VOT) hos svensktalande barn med fonologisk språkstörning och utröna huruvida deras VOT-värden skiljer sig från VOT-värden hos barn med typisk språkutveckling. I studien deltog 38 barn i åldrarna 4;2−11;6 fördelade på åtta åldersgrupper och fem fonologiska utvecklingsstadier. Inspelningar av bildbenämning av minimala par skedde på barnens förskola, skola eller logopedmottagning. Resultatet visade att barn med fonologisk språkstörning hade avvikande VOT-värden med stor variation. Det förekom både att grupper kunde och inte kunde producera akustiska skillnader mellan tonlösa och tonande klusiler. Resultaten var så spridda i barngruppen att ingen direkt utveckling mot vuxenlika värden kunde relateras till ökad ålder för varken utveckling av VOT produktion eller vuxenlik förekomst av förton. Ingen skillnad vad gäller VOT sågs mellan barn i olika fonologiska utvecklingsstadier eller åldrar. Hur mycket ett VOT-värde avvek kunde inte förklaras av ålder, språkliga processer eller hur många korrekta fonem (PPC) som producerades. Av resultaten dras slutsatsen att barn med språkstörning har avvikande VOT-värden som delvis kan hänföras till bristande fonologisk kunskap men framförallt, eftersom variabiliteten i barnens värden inte minskade med ökad ålder kan hänföras till svårigheter med det motoriska genomförandet.

A central issue in word recognition is how readers retrieve and select the right representation among others in the mental lexicon. Recently, it has been claimed that recognition of individual words is influenced by the degree to which the words possess unique vs. shared letters or sounds relative to other words, that is, whether the words have few or several neighbors. Research on so-called neighborhood density effects advances understanding of the organization and operation of the mental lexicon. Orthographic neighborhood effects have been claimed to be facilitative, but recent studies of visual word recognition have led to a revised understanding of the nature of the orthographic neighborhood density effect. Through a reexamination of orthographic and phonological neighborhood density effects, the specific objective of the present research is to understand how orthographic and phonological representations interact across two different writing systems, i.e., English (an alphabetic orthography) and Chinese (a morphosyllabic orthography). The phenomena were studied using a joint behavioral (lexical decision) and neural imaging approach (near infrared spectroscopy, or NIRS). Orthographic and phonological (more, specifically, homophone) neighborhood density were manipulated in three lexical decision experiments with English and three with Chinese readers. After different sources of facilitative inter-lexicon connections were controlled, orthographic and phonological neighborhood density effects were found to be inhibitory in both writing systems. Inhibitory neighborhood density effects were also confirmed in two NIRS experiments of English and Chinese. The present research provided a better control of lexical characteristics than was the case in previous research on neighborhood effects and found a clear and consistent pattern of neighborhood density effects. This research supports interactive-activation models of word recognition rather than parallel-distributed models, given the evidence for lateral inhibition indexed by inhibitory neighborhood density effects. As such, the present study furthers the understanding of the organization and operation of the mental lexicon.