Dyslexia detection using 3D convolutional neural networks and functional magnetic resonance imaging.

BACKGROUND AND OBJECTIVES: Dyslexia is a disorder of neurological origin which affects the learning of those who suffer from it, mainly children, and causes difficulty in reading and writing. When undiagnosed, dyslexia leads to intimidation and frustration of the affected children and also of their family circles. In case no early intervention is given, children may reach high school with serious achievement gaps. Hence, early detection and intervention services for dyslexic students are highly important and recommended in order to support children in developing a positive self-esteem and reaching their maximum academic capacities. This paper presents a new approach for automatic recognition of children with dyslexia using functional magnetic resonance Imaging. METHODS: Our proposed system is composed of a sequence of preprocessing steps to retrieve the brain activation areas during three different reading tasks. Conversion to Nifti volumes, adjustment of head motion, normalization and smoothing transformations were performed on the fMRI scans in order to bring all the subject brains into one single model which will enable voxels comparison between each subject. Subsequently, using Statistical Parametric Maps (SPMs), a total of 165 3D volumes containing brain activation of 55 children were created. The classification of these volumes was handled using three parallel 3D Convolutional Neural Network (3D CNN), each corresponding to a brain activation during one reading task, and concatenated in the last two dense layers, forming a single architecture devoted to performing optimized detection of dyslexic brain activation. Additionally, we used 4-fold cross validation method in order to assess the generalizability of our model and control overfitting. RESULTS: Our approach has achieved an overall average classification accuracy of 72.73%, sensitivity of 75%, specificity of 71.43%, precision of 60% and an F1-score of 67% in dyslexia detection. CONCLUSIONS: The proposed system has demonstrated that the recognition of dyslexic children is feasible using deep learning and functional magnetic resonance Imaging when performing phonological and orthographic reading tasks.

Reading networks in children with dyslexia compared to children with ocular motility disturbances revealed by fMRI.

Key PointsDyslexia is a neurological disorder with a genetic origin, but the underlying biological and cognitive causes are still being investigated.This study compares the brain activation pattern while reading in Spanish, a semitransparent language, in three groups of children: typically developing readers, dyslexic readers and readers with functional monocular vision.Based on our results Dyslexia would be a neurological disorder not related to vision impairments and would require a multidisciplinary treatment based on improving phonological awareness and language development. Developmental dyslexia is a neurological disorder the underlying biological and cognitive causes of which are still being investigated, a key point, because the findings will determine the best therapeutic approach to use. Using functional magnetic resonance imaging, we studied the brain activation pattern while reading in the language-related cortical areas from the two reading routes, phonological and orthographic, and the strength of their association with reading scores in 66 Spanish-speaking children aged 9-12 years divided into three groups: typically developing readers (controls), dyslexic readers and readers with monocular vision due to ocular motility disorders but with normal reading development, to assess whether (or not) the neuronal network for reading in children with dyslexia has similarities with that in children with impaired binocular vision due to ocular motility disorders. We found that Spanish-speaking children with dyslexia have a brain circuit for reading that differs from that in children with monocular vision. Individuals with dyslexia tend to hypoactivate some of the language-related areas in the left hemisphere engaged by the phonological route, especially the visual word form area and left Wernicke's area, and try to compensate this deficit by activating language-related areas related to the orthographic route, such as the anterior part of the visual word form area and the posterior part of both middle temporal gyri. That is, they seem to compensate for impairment in the phonological route through orthographic routes of both hemispheres. Our results suggest that ocular motility disturbances do not play a causal role in dyslexia. Dyslexia seems to be a neurological disorder that is unrelated to vision impairments and requires early recognition and multidisciplinary treatment, based on improving phonological awareness and language development, to achieve the best possible outcome.

Isolated and reversible lesions of the corpus callosum: a distinct entity.

The Reversible Splenial Lesion Syndrome represents a distinct clinicoradiological syndrome, associated with several disorders, including infection, high altitude cerebral edema, antiepileptic drug withdrawal, and severe metabolic disturbances (hypoglycemia and hypernatremia). Clinical presentation is nonspecific, most frequently as an encephalopathy or encephalitis. Outcome is favorable in most patients unless there is a severe underlying disorder. Magnetic resonance imaging findings are restricted to the splenium and consist of a nonenhancing oval lesion, hyperintense on T2-weighted images, including FLAIR. Findings on diffusion-weighted imaging are consistent with cytotoxic edema except for high-altitude cerebral edema, where vasogenic edema is present. Resolution after weeks or months is the rule.