The brain’s capacity to use and understand language expands rapidly in the first years of life, as babies start to make sense of the words they hear and eventually begin to piece together sentences of their own. The language-processing parts of the brain that make this possible continue to evolve in older children, as they expand their vocabularies and learn to use language more flexibly.
For the first time researchers have used an advanced AI model that understands both images and language allowing them to model dyslexia, paving the way for potential new treatments.
Dyslexia, the world’s most common learning disorder impacting reading, spelling and writing, is estimated to affect up to 20% of the global population. Until now, traditional approaches to studying dyslexia, such as behavioral and neuroimaging methods, have provided valuable insights but remain limited in their ability to test the underlying mechanisms of reading impairments.
Now, researchers from EPFL’s NeuroAI Lab , part of the Schools of Computer and Communication Sciences and Life Sciences have modelled dyslexia using next-generation Vision Language Models, that can fully model the whole pipeline from seeing words to processing and understanding the context.
A major study has revealed that where a child goes to school plays a role in whether they get diagnosed with a specific learning difficulty or not. Lead author, Dr Johny Daniel explains.
Two children sit in different schools. Both struggle to read. Both have similar low scores on national tests. But while one gets a diagnosis of specific learning difficulties and a package of support, the other is left to fall behind.
My colleagues and I have carried out new research analysing the records of around 540,000 primary school children across England. It reveals a troubling picture. Whether a child gets identified with specific learning difficulties – an umbrella term for conditions involving difficulties with reading and mathematics – depends not just on how they perform academically, but on the school they go to, their gender, their family’s income, their first language, and even the average ability of their classmates.
Sentence writing is a critical early writing skill (Kim et al., 2014) but is often overlooked in empirical literature (McMaster et al., 2018). The current pilot study evaluated the effect of a set of explicit sentence writing lessons on a first grader’s writing and identified the types of errors made before and during instruction.
Cognitive-linguistic deficits in kindergarten—especially in phonological awareness and letter knowledge—strongly predict early-emerging dyslexia by first grade.
Specific kindergarten deficits, particularly in letter knowledge, rapid automatized naming, and morphological awareness, remain significant risk factors for late-emerging dyslexia in fourth grade.
Difficulties in forming letter-speech sound associations may constitute a challenge for individuals with dyslexia. However, the learning trajectories of these associations remain poorly understood. This EEG study examined behavioral and neural changes while 31 typical and 31 dyslexic adult readers learned to map six novel symbols to Dutch spoken syllables with either high or low phonological similarity. Both groups demonstrated successful learning with learning-related ERP changes over frontotemporal, temporoparietal, and occipitoparietal regions. Phonologically similar vs. dissimilar pairs showed lower accuracy, slower reaction times, and reduced ERP responses, with earlier frontotemporal effects in dyslexic vs. typical readers (block 2 vs. blocks 3–4). As for learning outcomes, both groups showed temporoparietal (mis)matching responses in the last block. Dyslexic readers had lower post-training symbol reading scores, which correlated with their reading and phonological skills. Our findings indicate comparable learning during initial symbol-sound association in dyslexic readers, but difficulties applying novel associations during reading.