Fuzzy Assessment for Autism Spectrum Disorders

Autism spectrum disorder is characterized by being difficult to diagnose due to limited knowledge of its etiology and pathophysiology. There are, however, biomarkers that correlate to a limited extent the presence of the disorder, as well as its severity. A fuzzy inference system that simultaneously aggregates multiple biomarkers for assisting in the assessment of the autism spectrum disorder is proposed in this work. When combining the different biomarkers, the fuzzy decision support system is suitable to be applied in identifying the presence and the severity of the autism spectrum disorder. The Mamdani fuzzy model is employed to relate the input variables oxytocin, serotonin, and vitamin D levels to obtain an autism scale as the outcome. The assessment of autism spectrum disorder is given in the form of stratification (classes) and their respective score (score). The proposed system is consistent with the scientific literature, detaching a greater influence of serotonin on the onset of ASD and its severity. Moreover, the fuzzy assessment detaches that when the level of the vitamin D is reduced, there is a worsen on the onset and the severity of the autism spectrum disorder.