Research reportWorking-memory endophenotype and dyslexia-associated genetic variant predict dyslexia phenotype
Introduction
Developmental dyslexia is defined by severe difficulties in reading acquisition, often accompanied by spelling difficulties (American Psychiatric Association, 2000, World Health Organization, 2009), and affects 5–10% of children across literate societies (Katusic et al., 2001, Shaywitz et al., 1990). Dyslexia often persists into adulthood, and can affect individuals' everyday reading and writing activities as well as academic success (Raskind et al., 1999, Shaywitz et al., 1999). In cognitive science, classical dual-route (Coltheart et al., 2001, Ziegler et al., 2000) and connectionist models (Harm and Seidenberg, 1999, Seidenberg, 2005) have established the constituent components of the reading process from behavioral double dissociations. In contrast, contemporary approaches in cognitive neuroscience also focus on the neurological basis and potential genetic causes of dyslexia (see Carrion-Castillo et al., 2013, Richlan et al., 2009, Richlan et al., 2013), aiming to provide a more complimentary picture. Given the long-lasting and wide-ranging consequences of dyslexia, the long-term objective of these approaches is a diagnosis and intervention prior to reading onset. The current study follows this aim and considers behavioral, neurological, as well as genetic aspects in a multistage assessment of developmental dyslexia.
Evidence from cognitive science and neuroscience suggests that a phonological deficit in the processing and/or representation of speech sounds forms the basis for developmental dyslexia (see Goswami, 2015, Peterson and Pennington, 2015). Rules of grapheme-to-phoneme conversion and phoneme assembly need to be acquired in order to develop reading capabilities during early education. Once this is successfully completed, experienced readers can rely on their orthographic and lexical knowledge, mapping visual word forms directly to their associated lexicon and phonology (see Coltheart et al., 2001, Ziegler et al., 2000). In contrast, infants at risk for dyslexia have been found to display aberrant speech sound processing, suggesting a predictive relation of phonological skills to literacy acquisition prior to reading onset (Guttorm et al., 2010, van Zuijen et al., 2013). Individual differences in phonological processing have been consistently found to predict reading abilities across languages (Seymour et al., 2003, Ziegler et al., 2010, Ziegler and Goswami, 2005). Phonological processing deficits in dyslexia (Snowling, 2000, Vellutino et al., 2004) are often accompanied by deficits in auditory (Tallal, 1980) and visual processing (Stein & Walsh, 1997), attention (Facoetti et al., 2001, Valdois et al., 2003), and automaticity (Nicolson, Fawcett, & Dean, 2001). These cognitive manifestations have been suggested to dissociate different dyslexia subtypes (e.g., van Ermingen-Marbach et al., 2013, Heim et al., 2008). However, it has been argued that the non-phonological manifestations may primarily result from reduced reading experience in individuals with dyslexia (for discussion, see Goswami, 2015, Peterson and Pennington, 2015). On account of this evidence, our study focuses on phonological processing deficits of dyslexia. Thus, in a symptom-oriented approach we assessed clinically defining literacy impairments (i.e., reading and spelling) and accompanying phonological deficits that have been repeatedly reported for individuals with developmental dyslexia. Phonological deficits involve reduced verbal working memory (WM; Jorm, 1983, Stone and Brady, 1995), impaired phonemic awareness (Bradley & Bryant, 1983), and slowed rapid automatized naming (Catts, 1986). Note that impaired rapid naming has been controversially discussed as reflecting a phonological deficit (i.e., deficient access to phonological representations; see Ramus and Szenkovits, 2008, Wagner and Torgesen, 1987) or a non-phonological naming speed deficit (i.e., reduced automaticity in literacy-related processes; see Wolf and Bowers, 1999, Wolf et al., 2002).
As a neurobiological correlate of dyslexia, neuronal misplacements in left perisylvian regions have been proposed, potentially resulting from disturbed neuronal migration in language-related areas early during development (Galaburda et al., 1985, Humphreys et al., 1990). Indeed, morphometric and functional brain analyses in individuals with dyslexia revealed structural gray matter alterations (Brambati et al., 2004, Kronbichler et al., 2008, Pernet et al., 2009, Raschle et al., 2011, Richlan et al., 2013, Silani et al., 2005) and divergent activation patterns (Richlan et al., 2009) predominantly in left temporo-parietal and occipito-temporal regions. Converging on the gray matter findings, children and adults with dyslexia were also found to exhibit white matter alterations in temporo-parietal brain regions (Ben-Shachar et al., 2007, Carter et al., 2009, Eckert et al., 2005, Klingberg et al., 2000, Steinbrink et al., 2008), which were associated with reduced reading and phonological abilities (Beaulieu et al., 2005, Deutsch et al., 2005, Vandermosten et al., 2012).
On the genetic level, several dyslexia-associated genes have been identified (Burkhardt et al., 2012, Carrion-Castillo et al., 2013, Kere, 2014, Wilcke et al., 2012). Particularly well-replicated genes are located within so-called chromosomal susceptibility regions DYX1 to DYX9 (see Gibson and Gruen, 2008, Scerri and Schulte-Körne, 2010). A prominent example is the first gene found to be associated with dyslexia, that is, DYX1C1 on chromosomal region DYX1 (Paracchini, Scerri, & Monaco, 2007). Variant DYX1C1-rs3743204 has been, at least nominally, replicated as genetic risk factor in German- and English-speaking individuals (Dahdouh et al., 2009, Wigg et al., 2004). This genetic variant was also reported in a meta-analysis of dyslexic individuals from Austria, France, Germany, Switzerland, and the Netherlands (Becker et al., 2014), and was found in association with reading in non-affected individuals (Bates et al., 2010). On chromosomal region DYX2, the dyslexia-associated genes DCDC2 and KIAA0319 have been consistently replicated in different languages: A large deletion in DCDC2-rs71745442 was reported as genetic risk factor in German-, English-, and Italian-speaking participants with dyslexia (Cope et al., 2012, Marino et al., 2012, Meng et al., 2005, Wilcke et al., 2009), while variant KIAA0319-rs6935076 was confirmed as a risk factor in numerous studies on reading abilities in unaffected and dyslexic populations (Cope et al., 2005, Couto et al., 2010, Luciano et al., 2007, Newbury et al., 2011, Paracchini et al., 2008, Scerri et al., 2011). In addition to these well-known hypothesis-driven dyslexia-associated genes, a recent hypothesis-free microarray-based screening approach of validated dyslexia regions DYX1 to DYX9 identified further potential dyslexia candidate genes in a large German population, where variant TNFRSF1B-rs496888 on DYX8 showed the best nominal association with dyslexia in an independent replication (Kirsten, Wilcke, Ligges, Ahnert, & Boltze, 2011).
While these genetic-association studies provide evidence for a certain relationship between behavioral manifestations and underlying genetic predispositions, they neither allow for diagnostic inferences nor provide an immediate patho-mechanistical hypothesis. Instead, the moderate effect of individual genetic variations on behavior necessitates thorough assessment of such variations' impact on brain endophenotypes. The imaging genetics approach assumes that a particular brain structure provides a temporally stable endophenotype by which genetic variants and environmental factors affect behavior (Meyer-Lindenberg & Weinberger, 2006). This implicates an essential methodological advantage with respect to the dataset extent that delivers reliable results. In contrast to the high-input requirement in genetic association studies, imaging genetics frequently operates on participant numbers typically involved in imaging studies. The imaging genetics approach seems particularly promising in dyslexia: Imaging studies have revealed aforementioned dyslexia-associated gray matter alterations in language-related areas while, in turn, genetic variants can cause structural divergences being detectable by magnetic resonance imaging (MRI) (Blokland et al., 2012, Kremen et al., 2010, Peper et al., 2009), specifically in language-related areas (Thompson et al., 2001). In particular, variations in DCDC2 were found to associate with wide-spread gray matter changes and reading-related brain activation (Cope et al., 2012, Meda et al., 2008). Moreover, variations in DCDC2, DYX1C1, and KIAA0319 were shown to affect temporo-parietal white matter (Darki et al., 2012, Mahalanobis, 2014), even though these structural features may be only indirectly affected by genetically determined alterations in neuronal migration (Meng et al., 2005, Paracchini et al., 2006, Wang et al., 2006).
In sum, despite of strong evidence for dyslexia-associated phenotypes, candidate genotypes, and mediating neuronal endophenotypes, a tripartite link is still missing. We followed the imaging-genetics approach, conceptualizing brain structure as endophenotype through which genetic variants and environmental conditions are associated with behavior. Thus, we combined structural imaging data, dyslexia-associated behavioral data, and genetic data: First, we analyzed gray matter changes associated with dyslexia-relevant phonological disabilities (i.e., verbal WM, phonemic awareness, and rapid automatized naming), alongside measures clinically defining dyslexia (i.e., spelling and reading). Second, we tested whether the identified behaviorally-relevant gray matter changes are associated with candidate genetic risk variants reported for dyslexia. In a third step, we directly tested whether the gray matter changes and associated genetic variants are predictive of participants' diagnostic status (i.e., being diagnosed with dyslexia vs not being diagnosed).
Section snippets
Participants
Thirty-two male adults participated in the study. Based on participants' clinical history of having been diagnosed with dyslexia during literacy acquisition by a professional versus not having been diagnosed (i.e., participants' diagnostic status), there was a patient group (n = 16) and a control group (n = 16). To confirm affected individuals' diagnosis (and to ensure that control participants' were unimpaired) we obtained clinically relevant measures of reading and spelling. In acquiring the
Behavioral data
Participant groups differed in their performance in the assessed dyslexia-associated behavioral domains (Table 1). Participants with dyslexia generally showed significantly slower speed and decreased accuracy compared to control participants. Apart from significant group differences across all behavioral domains (although not present in all subtests), participants with dyslexia showed sub-standard performance in the diagnostic reading and spelling tests. As expected, reading was not as strongly
Discussion
Our study aimed at establishing a missing tripartite link between dyslexia phenotype, candidate genotypes, and potentially mediating neuro-endophenotypes (i.e., gray matter alterations related to a range of dyslexia-relevant behavioral domains). To this end, we assessed gray matter changes associated with dyslexia-specific impairments in phonological processing (i.e., verbal WM, phonemic awareness, and rapid automatized naming), reading, and spelling, and tested whether these brain-structural
Conclusion
Our study revealed verbal-WM-related gray matter alterations in HG/pSTG and the pSTS that interacted with genetic risk variants in TNFRSF1B in the prediction of dyslexia. Participants with a structural predominance of HG/pSTG and a low genetic risk tended to be classified as controls. Functionally, controls may thus employ early auditory-sensory stages of verbal WM as gateway for the conversion of sensory-based into phonological representations. Participants with a structural predominance of
Conflict of interest
A. W., H. K., and J. B. are co-inventors on a pending patent describing a method to identify risk for dyslexia by means of analyzing SNP rs496888. No further conflicts of interest, including financial ones, exist.
Acknowledgments
We thank Anke Kummer, Mandy Jochemko, and Domenica Wilfing for imaging data acquisition, and Elfi Quente for technical assistance. This work was supported by the Max Planck Society (C. M., L. M., and A. F.) and the German Federal Ministry of Education and Research (H. K. and A. W., grant number PtJ-Bio 0315883).
References (172)
A fast diffeomorphic image registration algorithm
NeuroImage
(2007)Computational anatomy with the SPM software
Magnetic Resonance Imaging
(2009)- et al.
Unified segmentation
NeuroImage
(2005) Working memory: theories, models, and controversies
Annual Review of Psychology
(2012)- et al.
The phonological loop as a language learning device
Psychological Review
(1998) - et al.
Sensitivity to lexical stress in dyslexia: a case of cognitive not perceptual stress
Dyslexia
(2012) - et al.
Dyslexia and DYX1C1: deficits in reading and spelling associated with a missense mutation
Molecular Psychiatry
(2010) - et al.
Imaging brain connectivity in children with diverse reading ability
NeuroImage
(2005) - et al.
Genetic analysis of dyslexia candidate genes in the European cross-linguistic NeuroDys cohort
European Journal of Human Genetics
(2014)
White matter pathways in reading
Current Opinion in Neurobiology
Controlling the false discovery rate: a practical and powerful approach to multiple testing
Journal of the Royal Statistical Society Series B
Genetic and environmental influences on neuroimaging phenotypes: a meta-analytical perspective on twin imaging studies
Twin Research and Human Genetics: the Official Journal of the International Society of Twin Studies
Developmental dyslexia: the visual attention span deficit hypothesis
Cognition
Annotation of functional variation in personal genomes using RegulomeDB
Genome Research
Categorizing sounds and learning to read – a causal connection
Nature
Regional reductions of gray matter volume in familial dyslexia
Neurology
The search for the phonological store: from loop to convolution
Journal of Cognitive Neuroscience
Human dorsal and ventral auditory streams subserve rehearsal-based and echoic processes during verbal working memory
Neuron
Association of rs2069459 in the CDK5 gene with dyslexia in a German cohort
Psychiatric Genetics
Molecular genetics of dyslexia: an overview
Dyslexia
A dual DTI approach to analyzing white matter in children with dyslexia
Psychiatry Research: Neuroimaging
Speech production phonological deficits in reading disordered children
Journal of Learning Disabilities
Context-dependent encoding in the human auditory brainstem relates to hearing speech in noise: implications for developmental dyslexia
Neuron
Neuroanatomical precursors of dyslexia identified from pre-reading through to age 11
Brain
Reading normal and degraded words: contribution of the dorsal and ventral visual pathways
NeuroImage
DRC: a dual route cascaded model of visual word recognition and reading aloud
Psychological Review
Variants in the DYX2 locus are associated with altered brain activation in reading-related brain regions in subjects with reading disability
NeuroImage
Strong evidence that KIAA0319 on chromosome 6p is a susceptibility gene for developmental dyslexia
American Journal of Human Genetics
Association of reading disabilities with regions marked by acetylated H3 histones in KIAA0319
American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics: the Official Publication of the International Society of Psychiatric Genetics
Further evidence for DYX1C1 as a susceptibility factor for dyslexia
Psychiatric Genetics
Three dyslexia susceptibility genes, DYX1C1, DCDC2, and KIAA0319, affect temporo-parietal white matter structure
Biological Psychiatry
Rapid “automatized” naming of pictured objects, colors, letters and numbers by normal children
Cortex
An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest
NeuroImage
Children's reading performance is correlated with white matter structure measured by diffusion tensor imaging
Cortex
Differential expression, cytokine modulation, and specific functions of type-1 and type-2 tumor necrosis factor receptors in rat glia
Journal of Neuroimmunology
Anatomical signatures of dyslexia in children: unique information from manual and voxel based morphometry brain measures
Cortex
Distinct neural signatures of cognitive subtypes of dyslexia: effects of lexicality during phonological processing
Acta Neurobiologiae Experimentalis
Orienting of visual attention in dyslexia: evidence for asymmetric hemispheric control of attention
Experimental Brain Research
Neuroimaging studies of word reading
Proceedings of the National Academy of Sciences of the United States of America
Neurodegenerative and neuroprotective effects of tumor necrosis factor (TNF) in retinal ischemia: opposite roles of TNF receptor 1 and TNF receptor 2
Journal of Neuroscience
Developmental dyslexia – 4 consecutive patients with cortical anomalies
Annals of Neurology
The human lexinome: genes of language and reading
Journal of Communication Disorders
A voxel-based morphometric study of ageing in 465 normal adult human brains
NeuroImage
Sensory theories of developmental dyslexia: three challenges for research
Nature Reviews. Neuroscience
Linkage studies suggest a possible locus for developmental dyslexia on chromosome 1p
American Journal of Medical Genetics
Newborn event-related potentials predict poorer pre-reading skills in children with dyslexia
Journal of Learning Disabilities
Linking sight and sound: fMRI evidence of primary auditory cortex activation during visual word recognition
Brain and Language
Phonology, reading acquisition, and dyslexia: insights from connectionist models
Psychological Review
Nonstationary cluster-size inference with random field and permutation methods
NeuroImage
Cited by (18)
Sensory auditory interval perception errors in developmental dyslexia
2020, NeuropsychologiaCitation Excerpt :The best performing group made fewer errors and had shorter response times and was comprised of more than twice as many NR as DD individuals; the slowest and most error prone group contained twice as many DD as NR cases, and the middle ground was made up from an equal number of each reader category. This suggests that the continuum of sensitivity to rapid auditory tones in both DD and NR is worthy of consideration alongside categorical/diagnostic attribution of reading ability (Männela et al., 2015; Protopapas, 2014; Rosen, 1999; van der Lely et al., 1998). The temporal frequency of our individual tone pairs were within the delta band (1.5 – 4Hz) which is consistent with rate of stressed syllables in speech across languages (Arvaniti, 2009) and at a frequency problematic in DD (Leong et al., 2011).
Phonemes, words, and phrases: Tracking phonological processing in pre-schoolers developing dyslexia
2019, Clinical NeurophysiologyChildren with dyslexia show a reduced processing benefit from bimodal speech information compared to their typically developing peers
2019, NeuropsychologiaCitation Excerpt :For reading, German orthography is transparent (i.e., consistent grapheme-phoneme correspondences), whereas for spelling (i.e., phoneme-grapheme correspondence) it is less consistent and thus more demanding (Landerl and Wimmer, 2008). Accordingly, spelling is more informative than reading for classifying literacy-impaired German school children (see also Cantiani et al., 2013; Männel et al., 2015; Neuhoff et al., 2012). We applied a passive visual-auditory oddball paradigm with video-recorded mouth movements while syllables were pronounced aloud.
Phonological abilities in literacy-impaired children: Brain potentials reveal deficient phoneme discrimination, but intact prosodic processing
2017, Developmental Cognitive NeuroscienceCitation Excerpt :Specifically for word-level prosodic processing, individuals with DD have been shown to perform normally in implicit tasks (e.g., priming tasks, Mundy and Carroll, 2012, 2013), but experience problems in explicit tasks (e.g., reiterative speech tasks, Goswami et al., 2010; Leong et al., 2011; Mundy and Carroll, 2012). Furthermore, deficient processing has been shown to arise in tasks with increasing memory load (Soroli et al., 2010), matching the well-documented finding of working memory deficits in children and adults with DD (Baddeley and Wilson, 1993; Jorm, 1983; Snowling, 1998; see also Männel et al., 2015). Thus, our finding of intact PB processing in children with DD might be rooted in both the processing advantage for larger prosodic units as well as the advantage of obtaining brain responses during passive listening, thus avoiding intervening task demands.
Present and past: Can writing abilities in school children be associated with their auditory discrimination capacities in infancy?
2015, Research in Developmental DisabilitiesCitation Excerpt :This can be explained by less consistent phoneme-grapheme correspondences (i.e., the relation between spoken and written language) compared to consistent grapheme-phoneme correspondences (i.e., reading direction) in the German language compared to, for example, English (Ziegler & Goswami, 2006). Thus, we expected writing to be more informative than reading to classify our German school children, which has also been shown for German adults (Cantiani, Lorusso, Guasti, Sabisch, & Mannel, 2013; Männel et al., 2015). Therefore, children at risk of developing reading and writing problems at preschool age (BISC; Jansen et al., 2002) and with a percentile rank (PR) < 25 in the writing test (i.e., according to the DERET's test norms; Stock & Schneider, 2008) represented the group with writing problems (N = 17).
- 1
Claudia Mannel and Lars Meyer contributed equally.
- 2
Holger Kirsten and Angela D. Friederici contributed equally.