Exploring Memory Compensation in Dyslexia: Strengths and Weaknesses in Memory Patterns Among Children and Adolescents

Declarative memory is a fundamental cognitive system responsible for learning and storing explicit knowledge, including facts (semantic memory) and personal experiences (episodic memory). Traditionally, it has been associated with consciously accessible information. However, emerging evidence suggests that it can also acquire and retain implicit knowledge across various cognitive domains [1, 2]. Given its flexibility and robustness, declarative memory plays a crucial role in learning processes and may serve as a compensatory mechanism in individuals with neurodevelopmental disorders such as dyslexia [2, 3].

Dyslexia is a specific neurodevelopmental disorder characterized by persistent difficulties in word reading, decoding, and spelling, despite adequate cognitive abilities and educational instruction. It has a multifactorial causal basis, involving neurobiological, psychological, and environmental factors. Research has indicated structural and functional brain differences in individuals with dyslexia, particularly in the left perisylvian and occipitotemporal regions; however, no single neurological marker has been identified [4‐6].

Research findings indicate that declarative memory remains largely intact in individuals with dyslexia, and may, in some cases, serve as a compensatory mechanism for phonological deficits [2]. One area in which declarative memory appears unaffected is nonverbal learning. Studies have shown that individuals with dyslexia exhibit typical learning of nonverbal visual information, suggesting that this aspect of declarative memory remains fully functional [3]. This distinction in preserved versus impaired memory processes resonates with fuzzy-trace theory [7, 8], which posits two types of memory traces: verbatim (surface-level, literal) and gist (general, meaning-based). According to this theory, individuals often rely on gist memory for decision making and learning, especially when verbatim traces are weak or inaccessible. In the context of dyslexia, gist-based encoding may serve as a compensatory mechanism, allowing individuals to bypass phonological decoding deficits by relying more heavily on meaning-based representations.

Additionally, while phonological deficits affect the encoding of verbal material, evidence suggests that verbal learning within declarative memory remains largely unimpaired when encoding difficulties are considered [3]. Research indicates that individuals with dyslexia may initially struggle with learning verbal information due to deficits in phonological processing and working memory; however, once information is encoded, their ability to retain and recall it is comparable to that of typically developing individuals [2, 3, 9]. Furthermore, lexical knowledge and semantic learning appear to be preserved in dyslexia, further supporting the notion that declarative memory compensates for phonological processing deficits. Studies on receptive vocabulary have demonstrated that individuals with dyslexia can acquire and understand words at a level comparable to that of their typically developing peers, despite their difficulties with phonological decoding [10, 11]. This suggests that they may rely on meaning-based strategies to aid in word recognition and comprehension. Taken together, these findings highlight the potential role of declarative memory as a cognitive strength in dyslexia, allowing affected individuals to develop alternative strategies for learning and retaining information.

Declarative memory may play an adaptive role in dyslexia by facilitating reading through several mechanisms. One such mechanism is chunking and whole-word memorization, in which individuals with dyslexia rely on memorizing all the words rather than phonetically decoding them [2, 12, 13]. This approach, which depends on declarative memory, helps to compensate for phonological deficits and contributes to reading fluency. Semantic context utilization also plays a crucial role in supporting reading skills. Research indicates that children with dyslexia exhibit improved word recognition when words are presented within a strong semantic context, suggesting a reliance on meaning-based strategies rather than phonological decoding [2, 14]. This finding reinforces the idea that declarative memory allows individuals with dyslexia to bypass phonological difficulties by leveraging stored semantic associations. Another significant aspect is explicit rule learning, which is a common component of reading interventions for dyslexia [15]. These interventions often emphasize direct instruction in phonological rules, and studies suggest that explicit learning of these rules enhances reading skills, likely because of the engagement of declarative memory processes [15]. Together, these mechanisms highlight the role of declarative memory as a compensatory tool in dyslexia, allowing affected individuals to develop alternative pathways for reading and comprehension.

Neuroimaging studies further support this hypothesis. Functional and structural magnetic resonance imaging (MRI) studies have demonstrated increased activation and volume in the hippocampus and medial temporal lobe (MTL) following reading interventions, suggesting enhanced reliance on declarative memory [16‐18].

To the best of our knowledge, this study is the first to comprehensively investigate a wide range of declarative memory using Latent Profile Analysis, in both children and adolescents from the general population and those with dyslexia. Thus, the aim of this study was to investigate the characteristics of declarative memory in children and adolescents aged 10–19 years and to examine whether individuals with dyslexia exhibit specificity.

Additionally, we sought to understand whether their declarative memory functions in a unique or atypical manner, or if their memory profile is typical despite the presence of phonological difficulties. This could shed light on whether their memory abilities differ from those of neurotypical individuals or whether the mechanisms they rely on for learning and retaining information are consistent with general cognitive development. A deeper understanding of this specific memory profile may contribute to the design of more effective support programs and therapeutic interventions that leverage the strength of declarative memory in individuals with dyslexia.

The Test of Memory and Learning, Second Edition (TOMAL-2) [19], offers one of the most extensive memory assessments within a standardized battery. Its core section comprises eight subtests that are evenly split into four verbal and four nonverbal tasks. These were used to determine verbal, nonverbal, and composite memory indices. The core subtests evaluate various aspects of memory, including free and associative recall, abstract and meaningful retention, sequential recall, and learning. This comprehensive set effectively covers key areas of memory assessment. Beyond the core battery, TOMAL-2 features four supplementary verbal and two nonverbal subtests, allowing for a more in-depth analysis. These additional subtests are particularly valuable for neuropsychologists and researchers. Incorporating these enables the calculation of additional indices such as the Verbal Delayed Recall Index (VDRI), Attention/Concentration Index (ACI), Sequential Recall Index (SRI), Free Recall Index (FRI), Associative Recall Index (ARI), and Learning Index (LI). The VDRI was obtained by applying a delayed recall procedure to two core verbal subtests.

To address the research questions and test the hypotheses, a three-step analytical strategy was employed, integrating person-centered and variable-centered approaches, to provide a comprehensive understanding of declarative memory functioning in individuals with dyslexia.

First, a Latent Profile Analysis [20, 21] was conducted to identify subgroups of participants based on their declarative memory performance across six indices derived from TOMAL-2: the Verbal Delayed Recall Index (VDRI), Attention/Concentration Index (ACI), Sequential Recall Index (SRI), Free Recall Index (FRI), Associative Recall Index (ARI), and Learning Index (LI). Latent Profile Analysis is widely used in cognitive and clinical psychology research to uncover underlying heterogeneity in cognitive performance profiles [22, 23]. Given that dyslexia is associated with variability in cognitive functioning, this approach was chosen to identify potential subgroups with distinct declarative memory patterns. A series of models with varying numbers of profiles (ranging from two to five) was compared using statistical fit indices to determine the optimal number of memory profiles [24]. These indices included the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), where smaller values indicated a better fit to the data. Additionally, entropy (ranging from 0 to 1) was considered, with higher values indicating better separation between profiles. The minimum average latent class probabilities for the most likely latent class membership (range: 0–1) were also examined to reflect the quality of the profile assignment. Models in which any profile represented less than 5% of the sample were considered insufficient for meaningful subgroup identification. The evaluation of the models also included the Bootstrapped Likelihood Ratio Test, which assesses whether a model with k profiles provides a significantly better fit than a model with k-1 profiles. A significant result (p < 0.01) indicates that the additional profile meaningfully improves the model fit, whereas a non-significant result suggests that adding more profiles does not significantly enhance the model’s explanatory power.

Second, to determine whether individuals with dyslexia were significantly more likely to belong to at least one of the identified latent memory profiles, a chi-square test was used.

Third, to examine specific differences in declarative memory performance, independent samples t-tests were conducted to compare declarative memory indices between individuals with dyslexia and the neurotypical controls. This method was selected because of its ability to assess mean differences between groups while controlling for variability within each group. Cohen’s d effect sizes were computed to determine the magnitude of the observed differences, providing additional insights into the practical significance of group-level differences in memory performance [25]. This analysis was critical for identifying the specific aspects of declarative memory that differentiated individuals with dyslexia from their neurotypical peers, particularly in relation to associative and learning-based recall.

Finally, a logistic regression model was constructed to assess the predictive power of declarative memory indices in distinguishing individuals with dyslexia from the neurotypical controls. This method was chosen because it allows multiple predictor variables to be examined while estimating their relative contributions to the dyslexia classification [26]. Backward selection was applied to identify the most significant predictors among the eight memory indices to ensure that only the most relevant variables were included in the final model. Nagelkerke R² was computed to evaluate the explanatory power of the model, indicating the proportion of variance in dyslexia diagnosis explained by the selected memory indices [27]. This analysis was crucial for determining whether specific aspects of declarative memory serve as reliable predictors of dyslexia, supporting the hypothesis that declarative memory functions as a compensatory mechanism in affected individuals.

To prevent age and sex from confounding the results of the independent samples t-tests and logistic regression model, a matched subsample was drawn from the neurotypical group. For each participant with dyslexia, a neurotypical individual of the same age and sex was randomly selected from the control group. This matching procedure minimized the influence of demographic differences on memory performance comparisons, thereby enhancing the validity of the findings.

All calculations were performed using Mplus 8.11 [28] and the R environment [29].

A logistic regression analysis was conducted to assess the predictive utility of TOMAL-2 indices in identifying individuals with dyslexia. The model included eight indices (two general [verbal and nonverbal] and six supplementary) as predictors, with the presence of a dyslexia diagnosis as the outcome variable. To refine the model, a backward selection procedure was applied to systematically remove non-significant predictors and improve model efficiency. Table 4 presents the results for both the baseline model (which included all predictors) and the final model (resulting from backward selection). Both models were statistically significant and accounted for 26% and 25% of the variance in dyslexia diagnoses, respectively, as indicated by Nagelkerke R² (see Table 4). The final model demonstrated a classification accuracy of 72%, suggesting a moderate ability to distinguish between individuals with and without dyslexia based on their memory performance.

The results indicated that lower scores on the Sequential Recall Index and Free Recall Index, combined with higher scores on the Nonverbal Memory Index, were significant predictors of a dyslexia diagnosis. This suggests that individuals with dyslexia tend to struggle with sequential and free recall tasks, which are closely linked to verbal memory deficits, while exhibiting relative strengths in nonverbal memory tasks.

This study examined the declarative memory profiles of individuals with dyslexia and their neurotypical peers using Latent Profile Analysis. The results revealed four distinct memory profiles, demonstrating similarities and differences between individuals with and without dyslexia.

Most participants, including those with dyslexia, were classified under Profile 3 (Typical Memory Performance), accounting for 78% of the total sample and 68% of the group with dyslexia. This finding is consistent with previous research suggesting that declarative memory remains largely intact in individuals with dyslexia [2]. However, a significant proportion of individuals with dyslexia displayed distinct memory characteristics, particularly within Profile 4 (Divergent Memory Abilities) and Profile 1 (Globally Impaired Memory).

Profile 4 (Divergent Memory Abilities) was observed in 19% of individuals with dyslexia, compared to only 7% of neurotypical participants. This group exhibited strengths in Associative Recall, Learning, and Verbal Delayed Recall but weaknesses in Attention/Concentration, Sequential Recall, and Free Recall. These results suggest that some individuals with dyslexia may rely on alternative memory strategies, such as semantic association and chunking, to compensate for phonological deficits. This interpretation is consistent with prior studies indicating that declarative memory supports reading and learning through meaning-based strategies rather than phonological decoding [2, 14]. These findings can also be interpreted through the lens of fuzzy-trace theory, which distinguishes between verbatim and gist memory representations [7]. Individuals with dyslexia may compensate for phonological processing deficits by increasingly relying on gist-based strategies that emphasize meaning and conceptual associations. Supporting this, Obidziński and Nieznański [30] found that adolescents with dyslexia were more susceptible to false memories for semantically related words, suggesting a cognitive shift toward gist-based processing. Their study offers empirical support for the idea that dyslexic memory functioning may involve both compensatory and maladaptive mechanisms—a duality that mirrors the divergent profiles observed in our study. Additionally, it is worth noting that the effectiveness of such learning strategies may depend on individual differences in metacognitive abilities. While the literature on metacognition in dyslexia remains limited, existing research suggests that children with dyslexia may show reduced metacognitive monitoring and regulation skills, potentially impacting their ability to plan, implement, and evaluate compensatory strategies effectively [31]. These differences may partly explain the observed variability in how declarative memory is leveraged across dyslexic learners.

Conversely, Profile 1 (Globally Impaired Memory) was more prevalent among individuals with dyslexia (12%) than in the neurotypical group (5%). The participants in this group exhibited below-average performance across all memory indices, with the most pronounced deficits in Associative Recall. This suggests that a subset of individuals with dyslexia may experience broader cognitive impairments beyond phonological deficits, potentially affecting their overall learning efficiency.

The findings from independent samples t-tests further corroborated these profile distinctions. Individuals with dyslexia scored significantly lower than their neurotypical peers across all declarative memory indices, with the largest effect sizes observed in Free Recall Index (d = 0.80) and Sequential Recall Index (d = 0.75). These results highlight the difficulties in recalling non-cued information and ordering information sequentially, which are critical for reading fluency and comprehension. Notably, a lack of sensitivity to sequential information has recently been reported in adults with dyslexia [32, 33]. The Serial-order Learning Impairment in Dyslexia (SOLID) hypothesis proposed by Szmalec et al. [32], is based on evidence that sequence learning is important for language learning [34‐36]. According to the SOLID hypothesis, individuals with dyslexia are less sensitive to such repetitions and are consequently less likely to learn (and automatize) sequences. Sensitivity to sequential patterns in environmental input has been argued to be crucial for language acquisition and has been identified as a potential core deficit in individuals with dyslexia. However, the mean scores in the group with dyslexia remained within the normative range, suggesting that, while declarative memory functions were relatively preserved, specific deficits may impact learning strategies.

The logistic regression analysis provided additional insights into the predictive utility of memory indices for identifying dyslexia. The final model, which retained Sequential Recall Index, Free Recall Index, and Nonverbal Memory Index as significant predictors, demonstrated a classification accuracy of 72%. Lower Sequential and Free Recall scores were associated with dyslexia, whereas higher Nonverbal Memory scores suggested relative strengths in nonverbal cognitive abilities. This finding aligns with the hypothesis that individuals with dyslexia may rely on nonverbal memory processes to compensate for verbal memory deficits [3].

These results have important implications for educational and therapeutic interventions. The presence of Divergent Memory Abilities in nearly 20% of individuals with dyslexia suggests that leveraging strengths in associative and meaning-based recall may be beneficial in instructional strategies. Tailoring interventions to support declarative memory functions, such as explicit rule learning and semantic reinforcement, may enhance learning outcomes in individuals with dyslexia. Although declarative memory can play a compensatory role, a subset of individuals with dyslexia still experience significant difficulties in this domain. This variability highlights the need for individualized assessment, as assuming that all individuals with dyslexia have intact declarative memory may overlook those who require additional support. Therefore, investigating declarative memory in dyslexia is crucial for designing targeted interventions that address both its strengths and weaknesses. The variability observed in declarative memory profiles may also be better understood considering theoretical models of dyslexia. While the phonological deficit theory remains one of the most widely supported explanations [37, 38], our findings align more closely with multifactorial or multiple-deficit models [39, 40], which propose that dyslexia arise from the interplay of multiple cognitive risk and protective factors. The presence of both compensatory memory strengths and broader memory impairments in our sample suggests that declarative memory could function not only as a buffer against phonological difficulties but also as an area of vulnerability in some individuals. Our findings support the view that dyslexia is a heterogeneous disorder, and models that accommodate variability in cognitive functioning—such as the multiple-deficit framework—may offer more comprehensive explanatory power.

Although this study provides valuable insights into the declarative memory profiles of persons with dyslexia, several limitations warrant consideration. The cross-sectional design precludes causal inferences, limiting our ability to determine how declarative memory compensation evolves over time. Although the sample size was substantial, it may not fully represent the diversity of individuals with dyslexia across different socio-cultural and linguistic contexts, potentially restricting the generalizability of the findings. Standardized tests such as TOMAL-2 may not capture all the nuances of memory performance in real-world learning environments. Furthermore, while efforts were made to control for confounding variables such as age and sex, other factors, such as comorbid learning disorders or socioeconomic background, could have influenced the observed memory profiles. To address these limitations, future studies should employ a longitudinal design to track the developmental changes in declarative memory and its compensatory role over time. Expanding studies to include more diverse populations and linguistic contexts will be essential for a more comprehensive understanding of the functions of declarative memory in dyslexia. Such research could also help design targeted interventions that leverage declarative memory strength to support individuals with dyslexia more effectively.

While most individuals with dyslexia demonstrate typical memory performance, a significant subset exhibits either pronounced strength of associative recall or generalized memory impairment. The identification of distinct memory profiles highlights the heterogeneity of cognitive functioning in dyslexia and underscores the need for personalized learning approaches that accommodate individual memory strengths and weaknesses.