Investigating the role of lexical stress and its relationship to phonological abilities in developmental dyslexia : Experimental evidence from Italian

Abstract

Background: Phonological deficits are among the most consistently reported impairments in Developmental Dyslexia (DD). However, their nature and underlying factors remain debated. One proposed explanation isreduced acoustic sensitivity to lexical stress, which could hinder phonological encoding and phoneme access. Yet, empirical findings remain inconsistent, likely due to methodological variability across studies. Given the frequent cognitive deficits in DD, an alternative explanation suggests that cognitive mediation may play a central role, with phonological impairments arising from deficient cognitive mechanisms that restrict access to otherwise spared phonological representations. Understanding the interplay between phonological processing and lexical stress sensitivity is crucial for clinical advancements, particularly in both diagnosis and intervention, and provides key insights into the relationship between spoken and written language. Objective: This thesis investigated whether lexical stress processing is impaired in DD, leading to disruptions in segmental processing and phoneme access, or whether phonological deficits stem from cognitive mediation deficits that restrict access to intact phonological representations. The thesis focused on the Italian language, which is more transparent and exhibits distinct acoustic cues underlying lexical stress compared to English—a language that has been more extensively examined in the dyslexia literature. This contrast offered an opportunity to evaluate the cross-linguistic generalizability of findings related to dyslexia-associated deficits in lexical stress processing and phonological processing. Method: To account for potential developmental compensation mechanisms and age-related changes associated with adulthood in dyslexia, we recruited two groups differing in their proximity to adulthood. A total of 40 adolescents with DD and 30 typically developing (TD) adolescents, along with 28 children with DD and 29 TD children, completed a phoneme monitoring task to assess implicit acoustic sensitivity to prominence in phoneme identification. Adolescents completed the task first in Study 1, while children were administered a simplified version in Study 3. Lexical stress sensitivity was further assessed in adolescents in Study 2 using a sensorimotor synchronization task, in which participants were instructed to tap in alignment with the perceived beat of spoken sentences. In contrast, children’s lexical stress sensitivity was evaluated in Study 4 through a perceptual identification task that required them to distinguish between two Italian minimal pairs differing in lexical stress along a continuum, such as papa [ˈpapa] “pope” vs. papà [paˈpa] “dad”. Both tasks were selected for their reduced metalinguistic component. Pupillometry was recorded during phoneme identification in children, as—unlike behavioural measures—it provides an implicit, objective index of online phonological access by being sensitive to both phonological and cognitive processing, thereby minimizing false negatives. Despite its potential, this cutting-edge method has rarely been applied to the study of DD. All participants underwent a cognitive assessment. Comorbid dyscalculia was retained as an inclusion criterion due to its non-linguistic nature, though it was controlled for throughout the analysis. Results: Findings from Studies 1 and 3, which employed phoneme monitoring tasks, indicated significant group-level differences in the phoneme identification performance across both cohorts, which could not be explained by inherently fragile lexical stress representations. Conversely, phonological identification differences between individuals with DD and their TD peers, regardless of age, appeared to be modulated by cognitive resources. Among TD adolescents, phonological identification rates were higher when attention-shifting ability was strong and lower when attention- shifting scores were weak. In contrast, this relationship was not observed in individuals with DD, whose performance remained consistently low, irrespective of attention-shifting ability. Additionally, cognitive resources contributed to group-level variations in children's pupillometric responses during distractor processing in the phoneme monitoring task. Studies 2 and 4 further confirmed the comparability of lexical stress representations across cohorts. Specifically, both adolescent groups demonstrated similar alignment to weak and strong syllables in the sensorimotor synchronization task of Study 2. In parallel, the categorical perception curves for lexical stress observed in Study 4 were consistent across children. Conclusion: Overall, this research thesis confirmed that phonological and possibly lexical stress representations remain intact in individuals with DD throughout late childhood and adolescence. However, access to these representations appears to be compromised, likely due to an imbalance between stronger and weaker cognitive resources—an interaction that requires further empirical validation. Rather than reflecting comprehensive phonological and lexical stress deficits, difficulties in DD seem to involve specific subdomains of phonological and prosodic competence, further corroborating the multicomponential nature of these domains, with awareness being highly interconnected with the cognitive domain. Although interconnected, segmental and prosodic domains seem to develop independently and exhibit notable cross-language differences, reaching an adult-like form only in late adulthood. Therefore, the potential contributing factors of language transparency variability, along with SES and reading exposure, should be considered and incorporated into future research to better understand their role in phonological processing in DD.