It is a well-known fact that autism condition cannot be detected easily during the first few months of infant’s life. However, a new study has showcased EEG’s which are used to measure the electrical activities of the brain can accurately predict whether or not an infant will be diagnosed with autism, later in his life.
“EEG’s are non-invasive, low-cost devices that can be easily incorporated during the course of routine baby checkups,” says Charles Nelson based out of Boston Children’s Hospital and has also co-authored the study.
Charles continues, “The reliability of these devices in predicting whether an infant will go on to develop autism condition helps in developing possible interventions much earlier than autism symptoms could emerge. This indirectly leads to fruitful outcomes and could also end up preventing behavioral complexities that could be associated with autism disorder.”
The study authors analyzed existing data from the earlier Infant Sibling Project, a collaboration between Boston University and Boston Children’s hospital. The project aims to identify and map the early developmental risks in infants with a probability of developing autism condition or other language and communication difficulties.
William Bosl’s research data suggests, even if an EEG displays as one being normal, it still contains an enormous amount of data that reflects the functions of the brain as well as the underlying structure and connectivity patterns that can be found using sophisticated computerized algorithms.
Bosl mined the EEG data from the Infant Screening project from 99 infants who were considered at a higher risk of autism condition. The EEG scans were regularly taken with frequency intervals of 3, 6, 9, 12, 18, 24 and 36 months. The EEG was modified to have 128 sensors and a net was fit on the baby’s scalps. An experimenter was engaged who had to blow bubbles in order to keep the young ones distracted.
Every child undergoing EEG analysis was also subjected to behavioral evaluations with Autism Diagnostic Observation Schedule.
The computational algorithms developed by Bosl analyzed six different components using various signal measurement complexities. These measurements also reflected how the brain components are wired internally and how the brain actually integrates and processes information.
The algorithm could predict the clinical diagnosis of autism symptoms with higher sensitivity, specificity and exceeded with 95 percent accuracy at some ages.
Bosl exclaims, “The results were absolutely stunning.” The algorithmic predictive accuracy by as early as 9 months was seen to be nearly 100 percent. We could also understand the severity of symptoms using the ADOS scale with decent pace.”
Bosl suggests that early differences in signal complexities, with underlying aspects of underlying brain activities, showcases that autism condition begins during the early development of brain but can also take different trajectories. In simpler terms, early predispositions to autism condition can also be influenced by a majority of other factors along the way.
Nelson explains, “Nevertheless, we believe there lies a greater possibility of autistic older siblings carrying genetic liabilities for being diagnosed with autism condition.”