Early care aims at the early identification of the population at risk for autism during their early stages of neurodevelopment. Maternal and paternal factors, such as the use of drugs or a history of autoimmune disease, have been linked to autism. Even fetal ultrasound parameters have been documented as related to autism. Preclinical and clinical research suggest that autism spectrum disorder begins to develop during pregnancy.
The increased incidence of autism has been linked to viral infections, activation of the immune system, maternal exposure to drugs such as valproic acid, vitamin D deficiency and in-utero hormonal influence. It has also been observed that an autism diagnosis is more frequent in children who have been the product of obstetric complications, such as gestational diabetes, preeclampsia and acute fetal distress resulting in cesarean birth or preterm delivery and those who have had APGAR < 5, as well as feeding difficulties and meconium aspiration syndrome.
One of the main objectives in this line of research is to test autism prediction models during early stages. To achieve this goal we use machine learning methods with special emphasis on the mother’s and father’s medication intake and prenatal data. Currently, we are identifying clinical factors that should be included in autism prediction models, through retrospective and prospective clinical cohorts such as the SIGNATURE project of emotional and inflammatory stress in pregnant women with COVID-19.