Correlation between language and cognitive skills in Egyptian children with reading problems

Abstract

Background

Reading problems are the most prevalent type of specific learning disability. Reading problems usually result from phonological deficits; however, it is unclear how much other linguistic abilities are impacted as well.

Objective

To study different language parameters and cognitive skills in a group of Egyptian children with reading problems to better understand the difficulties that need to be considered in rehabilitation programs for those children.

Methods

A case–control study was conducted on 30 children with reading problems having IQs above 85 and 30 normal control children matched for age and gender (their ages ranged from 7 to 9 years and 11 months). They underwent testing for dyslexia by the Arabic dyslexia assessment test, language evaluation using the Receptive Expressive Arabic Language Scale (REAL scale), and the Stanford-Binet intelligence scale, fifth edition.

Results

There were highly significant correlations between the severity of dyslexia and total language scores, most of REAL scale subtests. Also, highly significant correlations were detected between the severity of dyslexia and the total IQ, verbal IQ, and verbal working memory subtests. The linear regression model revealed a highly significant association between dyslexia and expressive language scaled scores, total language scaled scores, and a significant association with expressive vocabulary and understanding oral instructions subtests of REAL scale. Additionally, a highly significant association has been detected between dyslexia and the verbal working memory subtest of the Stanford-Binet test. Correlations between total language scores and intellectual abilities (total IQ, verbal IQ, and non-verbal IQ) revealed highly significant positive results.

Conclusion

Deficits in both receptive and expressive language were observed in children with reading problems. Expressive vocabulary, listening comprehension, and verbal working memory subtests are the most affected language and cognitive skills.

Background

Reading is a psycholinguistic process in which the reader deduces the writer's intended meaning by applying a variety of skills [1].

Coordination of a number of component skills is necessary before reading ability can emerge. Phonological (sound-letter) decoding, letter knowledge, rapid naming, verbal memory, pseudo-word repetition, and vocabulary are among these skills [2].

For about 5–12% of individuals worldwide, learning to read is very challenging. These individuals are affected by a complex neurodevelopmental disorder called developmental dyslexia, which represents the most prevalent learning disability among school-aged children and across all languages [3].

Developmental dyslexia is a neurobiologically based learning disability. It is characterized by challenges with precise and/or fluent word recognition, as well as weak decoding and spelling skills. These difficulties usually stem from a phonological impairment, which is unanticipated regarding other cognitive abilities and classroom-appropriate instruction. Comprehending difficulties and a decrease in reading experience are examples of secondary effects that could impede the development of background knowledge and vocabulary [4].

Various criteria have been put up to diagnose dyslexia, one of which takes into account a child’s IQ. A disparity between unexpectedly poor academic achievement and normal or high intellectual functioning is part of the diagnosis [5]. Even though they have strong general intellectual abilities, children with dyslexia may have particular deficiencies in certain cognitive abilities, such as working memory [6].

For many years, dyslexia has been classified as a language-based disorder; this classification has focused primarily on phonological deficits as a core feature of dyslexia. Thus, if phonological deficits are present in their profile, children with developmental language disorders are significantly more likely to have dyslexia. However, it is less clear how other areas of language development, like vocabulary, morphology, syntax, and discourse, are affected as well in dyslexic children [7].

Two types of reading disorders can be distinguished: decoding and comprehension. Children with decoding disorders struggle to learn how words’ spelling patterns correspond to their pronunciations. Children with comprehension disorders can read words accurately and fluently, but they have difficulty comprehending what they have read [8].

Reading difficulties are more common among children who have previously had difficulty with oral language. The kind of reading problem varies depending on the child’s language profile; reading comprehension problems are more closely associated with grammar, vocabulary, and receptive language, while poor decoding is directly tied to phonological inadequacies [9].

Language deficiencies beyond the phonological domain are common in dyslexic children, both with and without co-occurring developmental language disorders. Regardless of language abilities at diagnosis, those with dyslexia may be at risk for slower language acquisition across their lifetime [10]. Poor word-reading abilities have a negative impact on vocabulary development and reading comprehension. This becomes more clear around the third to fourth grades when students make the switch from learning to read to reading to learn. Reading deficits impede a child’s ability to improve his/her language skills through reading texts [11].

Assessment of multiple language skills is necessary to comprehend the complex language and reading development of dyslexic children, both at the time of diagnosis and years later.

Objectives

This study aimed to assess different language parameters and cognitive skills in a sample of Egyptian children with reading problems in order to better understand the challenges that must be taken into consideration when setting therapeutic objectives and creating rehabilitation programs for those children.

Methods

Thirty children with reading problems whose IQs were over 85 and 30 normal control children who were matched for age and gender participated in this case–control study. They were picked up from the Phoniatrics Unit outpatient clinic at Menoufia University Hospital between June 2022 and March 2023. Their ages ranged from 7 to 9 years and 11 months. The sample size had been estimated at a power of 80% and a confidence interval of 95% on the basis of the past literature review [12]. This study was approved by the Menoufia Faculty of Medicine’s ethical committee under reference number ENT20, registered on October 18, 2020. Before every child was enrolled in the study, their parents provided their written informed consent.

Exclusion criteria involved children with mental retardation, hearing impairment, autism spectrum disorder, or attention deficit hyperactive disorder.

Pathway of assessment

• I- Elementary diagnostic procedures

  • A- Parents interview: to fulfill

    • 1- Personal history.

    • 2- Complaints of both parents and teachers.

    • 3- Family history of similar conditions.

    • 4- Information about perinatal and neonatal periods.

    • 5- Milestones of development: motor development and language development.

    • 6- History of early childhood illness.

  • B- General examination.

  • C- Ear, nose, and throat examination.

  • D- Vocal tract examination.

  • E- Neurological examination.

• II- Clinical diagnostic aids:

  • A- Language evaluation:

Language skills were evaluated through the Receptive Expressive Arabic Language Scale (REAL-scale) [13], including 13 in-depth subtests. REAL scale subtests can be categorized as receptive, expressive, and receptive-expressive subtests.

Receptive subtests include

Receptive Vocabulary (RV), Sentence Comprehension (SC), Understanding Oral Instructions (UOI), and Comprehending Orally Presented Paragraphs (COPP).

Expressive subtests include

Expressive Vocabulary (EV), Morpho-Syntax (MS), Sentence Repetition (SR),Forming Sentences 1 (FS1), and Forming Sentences 2 (FS2).

Receptive-Expressive subtests include

Verbal Categorization Receptive 1 (VCR1), Verbal Categorization Receptive 2 (VCR2), Verbal Categorization Expressive 1 (VCE1), and Verbal Categorization Expressive 2 (VCE 2).

The raw score for each subtest is calculated and then converted into its equivalent scaled score (a mean of 10 and a standard deviation of 3) based on the child's chronological age. To obtain total scores, the constituent subtest raw scores for each language domain are grouped to give two principal raw total scores: the Receptive Language Score (RLS) and the Expressive Language Score (ELS). After that, the Total Language Score (TLS) is obtained by adding RLS and ELS. The raw scores are converted into their corresponding total scaled scores (having a mean of 100 and a standard deviation of 15) based on the child's chronological age.

• B- Arabic Dyslexia Test [14]

There are eleven items in the test. These include rapid naming, bead threading, one-minute reading, postural stability, phonemic segmentation, two-minute spelling, backward digit span, non-sense passage reading, one-minute writing, verbal fluency, and semantic fluency.

Each item was assigned one of the following scores: (+), (0), (-), (--), (---), after the scoring key suited for the child's age was located. Lastly, the test produced an at-risk quotient (ARQ). An ARQ of ≥ 1 indicated a child's strong suspicion of dyslexia.

• C- Psychometric evaluation using the Stanford Binet intelligence scale, fifth edition [15] for determination of total IQ, verbal IQ, and non-verbal IQ. The test includes five subtests distributed across verbal and non-verbal domains. These subtests are knowledge, fluid reasoning, quantitative reasoning, visuospatial processing, and working memory. The study included only children whose IQ was higher than 85.

Statistical analysis

Data collection, tabulation, and statistical analysis were performed using the Statistical Package for Social Science (SPSS) version 26 [16]. Number (No) and percent (%) were utilized to represent qualitative data, whereas mean and standard deviation (SD) were utilized to represent quantitative data.

The correlation between two quantitative variables was examined using the correlation coefficient test (r-test), and the findings may be positive or negative.

Probability of error: P-values greater than 0.05 were regarded as non-significant, P-values less than 0.05 as significant, and P-values less than 0.01 as highly significant.

Results

Two groups participated in this study (cases and control): 21 males and 9 females for each group. Among the studied cases, 83.3% have a history of delayed language development, and 36.7% have a family history of dyslexia.

Based on their age, cases were separated into three subgroups:

• Subgroup (A) consisted of 13 cases aged 7 to 7 years and 11 months.

• Subgroup (B) consisted of 11 cases aged 8 to 8 years and 11 months.

• Subgroup (C) consisted of 6 cases aged 9 to 9 years and 11 months.

In the same way, the control group was divided into three subgroups (AI, BI, CI).

Comparison between cases and control groups as regards REAL scale raw scores revealed significant differences in RLS, ELS, and TLS (Table 1).

Table 1 Comparison between cases and control groups as regards total raw scores of REAL scale

By studying the correlations between ARQ of the Arabic dyslexia test and total raw scores of the REAL scale in the studied cases, there were highly significant negative correlations between ARQ and RLS, ELS, and TLS (Table 2).

Table 2 Pearson correlation between ARQ and total language raw scores, IQ scaled scores in studied cases

Regarding the correlation between ARQ and total scaled scores of the Stanford-Binet intelligence scale in the studied cases, there were highly significant negative correlations between ARQ and total IQ and verbal IQ in three subgroups (Table 2).

By studying the correlation between ARQ and raw scores of REAL scale subtests in the studied cases, results revealed significant negative correlations with most of the REAL scale subtests in the three groups (Table 3).

Table 3 Pearson correlation between ARQ and raw scores of REAL scale subtests in studied cases

Regarding the correlation between ARQ and scaled scores of verbal IQ subtests among studied cases, there was a highly significant negative correlation between ARQ and verbal working memory subtest in groups A and B and a significant negative correlation in group C (Table 4).

Table 4 Pearson correlation between ARQ and scaled score verbal IQ subtests

There were highly significant positive correlations between total language scores and total IQ, verbal IQ, and non-verbal IQ (Table 8).

Discussion

Reading acquisition is a complex process, and oral language components are crucial to this process. Most research has given less attention to oral language abilities in dyslexia. To understand how efficient reading is acquired, it is important to explain the associations between several components of reading development, including spoken language, word reading, and reading comprehension [17].

This study showed that over two-thirds of the cases with reading difficulties were male, which is in agreement with the findings of Al Dakhil et al. [18], who found that male students were twice as likely as female students to have dyslexia. Hormonal factors, such as fetal testosterone levels during late pregnancy, may play a crucial role in male predominance in most of these conditions [19].

Nearly one-third of the cases in this study had a family history of learning disabilities. This finding supports the theory that dyslexia may have a genetic basis, which has been the subject of numerous studies such as Catts [20]. Catts’ study indicates that children who have a first-degree relative who has reading difficulties are 40–66% more likely to struggle with reading themselves, compared to 6–14% of children who do not have a reading-challenged family member.

Results of our study showed that delayed language development had previously occurred in 83.3% of cases (most of them started language therapy before the age of 4 years). This finding agrees with the results of Snowling et al. [21], who studied language and literacy outcomes in three groups of children (preschool-aged children with language impairment, children at family risk of dyslexia, and typically developing controls) at three time points: t1 (age 3 years), t2 (age 5 years), and t3 (age 8 years). They found that persistently low oral language proficiency at the time of reading acquisition is associated with a high risk of future reading difficulties.

First, we compared cases and control groups as regards receptive and expressive language skills. Results revealed that there was a significant difference between cases and control groups in RLS, ELS, and TLS in the three age groups. These results agree with Price et al. [22], who detected that children with low reading competence have poor receptive and expressive language skills. Their study looked at the association between reading skills and early language delay in three groups of children: those with developmental dyslexia, intermediate readers, and skilled readers. They concluded a significant association between developmental dyslexia and early language delay, with parents of children in the developmental dyslexia and intermediate readers groups reporting their children put words together later than those in the skilled readers group.

Another study by Osman et al. [23] found that students with poor academic achievement have below-average performance levels in receptive, expressive, and total language scores.

By comparing cases and control groups as regards total IQ, verbal IQ, and non-verbal IQ, there was a significant difference between cases and control groups as regards the raw score of total IQ in all age groups, a highly significant difference as regards the raw scores of verbal IQ and non-verbal IQ in subgroups A and B, and a significant difference as regards the raw scores of verbal and non-verbal IQ in subgroup C, as shown in Table 5. This goes with the results of D’Angiulli and Siegel [24], who compared the intelligence performance of normal students with that of students who had specific disabilities in arithmetic and students who had specific disabilities in reading. They found that the mean IQ scores for the students with arithmetic or reading disorders were significantly lower than the mean score of the typically developing group.

Table 5 Comparison between cases and control groups as regards raw scores of total IQ, verbal IQ, and non-verbal IQ

By studying the correlation between the degree of severity of dyslexia and performance in REAL scale total raw scores, there were significant negative correlations with RLS, ELS, and TLS in all studied cases except RLS in subgroup C. The negative correlations are due to the method of scoring. The more severe the dyslexia, the lower the scores on the REAL scale. These results supported the findings of Gathercole et al. [25], who indicated that language, phonological awareness skills, and complex memory were significantly correlated with the degree of reading difficulties.

Regarding the correlations between ARQ and IQ scores, there were highly significant negative correlations with total IQ and verbal IQ in all age groups, a highly significant negative correlation with non-verbal IQ in subgroup A, and a significant negative correlation with non-verbal IQ in subgroup B (Table 2).

Another correlational study between the severity of dyslexia and performance in REAL scale subtests revealed that:

• For subgroup A (aged 7–7 years and 11 months), there were highly significant negative correlations between ARQ and the raw score of subtests UOI, VCR2, VCE2, and SR and significant negative correlations with the raw score of subtests RV, SC, VCR1, COPP, EV, MS, VCE1, and FS2.

• For subgroup B (aged 8–8 years and 11 months), there were highly significant negative correlations between ARQ and the raw scores of subtests SC, UOI, VCR2, EV, and SR and significant negative correlations with the raw scores of subtests COPP, VCE1, and VCE2.

• For subgroup C (aged 9–9 years and 11 months), there were significant negative correlations between ARQ and the raw scores of subtests RV, SC, UOI, VCR2, COPP, EV, FS2, and SR.

Receptive and expressive vocabularies contribute significantly to pre-reading skills (e.g., letter naming, rapid naming, morphology, and phonological awareness).

Receptive language has been proposed as a central precursor to reading due to its role in speech perception and speech processing in relation to word reading development [26]. This supports the lexical restructuring model by Metsala and Walley [27], which postulated that children’s phonological representations gradually become segmented as a result of a restructuring process driven by expanding vocabularies in their lexicon. Impaired receptive skills are a major barrier to reading acquisition. Learning phoneme-letter connections during the early stages of reading acquisition is crucial, and one of its main influences may be the inability to distinguish between separate phonemes due to perceptual problems.

The acquisition of reading fluency and comprehension later on is linked to receptive and expressive language abilities. Among late talkers, reading comprehension and fluency both declined more rapidly, particularly if receptive and expressive language were delayed [28].

The sentence repetition subtest in the REAL scale depends mainly on the auditory verbal memory of the child by testing the child’s ability to reproduce sentences of varying length and syntactic complexity. The results of the sentence repetition subtest of the REAL scale showed a significant negative correlation with ARQ in all studied cases. Plaza et al. [29] also observed notable deficiencies in digit span, unfamiliar word repetition, and sentence repetition among a sample of dyslexic children when they studied auditory memory skills.

Also, our study showed that REAL scale subtests targeting listening comprehension (COPP, UOI, and SC) significantly negatively correlated with the ARQ of dyslexia tests. The same findings were reached by Cain et al. [30]. Recently, Stevens et al. [31] discovered that dyslexic students’ listening comprehension performance was noticeably worse than that of age-matched skilled readers.

In terms of IQ subtests, subgroups A and B showed highly significant negative correlations with the verbal working memory subtest, while subgroup C showed a significant negative correlation with the same subtest. These results are consistent with those of Roid and Barram [32], who found that the full-scale IQ of the Stanford-Binet fifth edition and its subtests, mainly working memory, knowledge, and quantitative reasoning subtests, are directly related to reading and arithmetic skills.

The results of the study by Knoop-van Campen et al. [33] supported that working memory indirectly affects word reading efficiency; this effect is mediated through phonological awareness.

To predict the association between dyslexia and Stanford-Binet total and subtest scaled scores, a linear regression model was used. The results showed that the verbal working memory subtest is the only one that has a highly significant negative association with dyslexia (Table 6 and Fig. 1).

Table 6 Linear regression model to predict association between ARQ and scaled scores of verbal IQ subtests

Fig. 1

Linear regression model to predict association between ARQ and scaled scores of verbal working memory subtest

Alloway’s study [34] demonstrated a strong correlation between working memory capacity and academic progress. Also, Pham and Hasson [35] investigated the association between working memory and reading fluency in students at varying reading levels. They found that assessments of verbal working memory significantly predicted reading fluency.

In contrast, De Carvalho et al. [36] studied students with and without dyslexia in the third to eighth grades. They found that reading fluency and phonological working memory were positively correlated in non-dyslexic students but not in the dyslexic group.

To predict the association between dyslexia and REAL scale total and subtests scaled scores, linear regression revealed that for the total scores, there were highly significant negative associations with RLS and ELS (Table 7). For subtest scores, there were significant negative associations with EV and UOI subtests (Table 7 and Fig. 2).

Table 7 Linear regression model to predict association between ARQ and scaled scores of REAL scale

Fig. 2

Linear regression model to predict association between ARQ and scaled scores of REAL scale subtests (UOI, EV)

Expressive vocabulary knowledge better predicts word decoding skills because it involves using both phonological representations and semantic understanding. Word identification is supported by expressive vocabulary in two ways: First, students with large vocabularies have well-established phonological representations of words. The second route is related to the depth of vocabulary knowledge. Students with deep vocabulary knowledge can efficiently recall the phonological representations of words. These two pathways of vocabulary effect imply that receptive and expressive vocabulary may relate to measures of reading achievement differentially. Word decoding performance may be more closely correlated with expressive vocabulary than receptive vocabulary because expressive vocabulary involves accessing semantic knowledge and phonological representations [37].

The UOI subtest assesses listening comprehension and places an extra load on working memory. Our findings agree with those of Cain et al. [30], who discovered a relationship between reading comprehension abilities in children aged 7 to 10 and assessment of working memory, vocabulary, comprehension monitoring, and inference-making. Listening comprehension difficulties might arise as secondary consequences or are often co-occurring with dyslexia.

Finally, we studied the correlation between total language scores and total IQ, verbal IQ, and non-verbal IQ. The results indicated highly significant positive correlations (Table 8). These results are in line with those of Liao et al. [38], who found that performance IQ was correlated with language scores and that children with language impairment had lower IQ scores.

Table 8 Pearson correlation between total language scores and (total IQ, verbal IQ, non-verbal IQ)

Conclusion

Deficits in both receptive and expressive language were evident in children with reading difficulties. Expressive vocabulary and listening comprehension subtests are more associated with reading difficulties. Although the studied children had average intelligence, they demonstrated particular deficiencies in verbal working memory skills.

Limitations of the study

This work was a cross-sectional case–control study. We studied language and cognitive skills in children with reading problems at a certain point in time, so further longitudinal studies are needed to follow dyslexic children after designing rehabilitation programs for targeting the affected language parameters and cognitive skills. Furthermore, our sample size was small, and further studies with larger sample sizes are recommended.

Recommendations

A thorough evaluation of both linguistic and cognitive abilities in children with reading problems is a must to identify the more subtle difficulties and to act early to facilitate timely intervention and better outcomes. Further studies are recommended to apply intervention programs targeting the affected language parameters (expressive vocabulary, listening comprehension) and cognitive skills (verbal working memory) in dyslexic children.