Download In this thesis, a research is illustrated that involved 40 secondary school children who were diagnosed with DSA. The data from the tests administered were correlated with each other and analyzed with modern machine learning techniques. We identified a direct correlation between the IRP Perceptual Reasoning Index and the Trog-2.
Regarding reading, in the dyslexic subjects of our sample there is a more evident (and lasting) deficit in reading speed, when compared with correctness.
With regard to the writing tests, carried out in normal conditions and in conditions of articulatory suppression, it emerges that writing performance worsens considerably if carried out in conditions of articulatory suppression, both for the correctness aspect (word dictation) and for the speed aspect. (writing numbers in letters).
Significant correlations emerged between working memory (IML), analyzed using WISC IV and WAIS IV e all The indices of BDE-2, in particolare with the area of the calculation. The index of the memory di Work IML results be BEN related also with lo chip oral numerico in way direct.
The most interesting result obtained from our analyzes is certainly the creation of a “black” model memory box”. of work according to which, having only the age of the subject, the verbal memory span and the Number Quotient of the BDE-2 available, we would also be able to estimate the value of the Working Memory Index of WISC IV and WAIS IV, without administer the relevant tests, with an accuracy of 83%.
Models di this type they throw le of semi-permanent base gels for un new scenario future di review neuropsychological of the DSA: if we had, in fact, models train su a great number of subjects of age different, it would be possible, during un path di assessment, administer required i test o even required The items and are be from model more predictive with the specific I disturb. Tale scenario è anything but and utopian, in as such mode is already applied each day in an infinity di areas (to example la dealer targeted and is done in basis preferences of identification Other immense è and un model and have used personal data previously acquired).