Artificial Intelligence to Predict Autism Spectrum Disorders from Birth
Our groundbreaking project, Pelargos, represents a major advancement in predicting the risk of Autism Spectrum Disorders (ASD) in newborns. Using innovative computational methods, our research aims to develop a medical tool for prognosis and diagnostic support for these conditions.
Building on the promising potential of our published preliminary results, B&A Biomedical has launched a large-scale study divided into 3 phases, in collaboration with major French maternity hospitals and associated Autism Resource Centers.
Open to new collaborations, we invite you to connect with our team to discuss the different phases of our project in more detail and perhaps join us in the Pelargos adventure.
The Context
Autism Spectrum Disorders
Autism Spectrum Disorders (ASD) encompass a wide range of conditions that affect various aspects of behavior. They are characterized by impairments in social interactions, communication difficulties, limited adaptability to change, repetitive behaviors, and restricted interests. These disorders can be severe and pervasive, significantly impacting daily life and challenging both the autonomy and quality of life of individuals affected, as well as their families and caregivers.
The prevalence of ASD is significant, affecting approximately one in 50 children, and this incidence appears to be increasing, though some underlying causes remain unclear.
To date, no curative treatment exists for ASD. Psychoeducational approaches and behavioral therapies are the only alternatives approved by European authorities. These interventions aim to mitigate symptoms through compensatory strategies, improving long-term outcomes and patient autonomy.
However, the full potential of these therapies relies on early implementation, ideally between ages 2 and 3, when brain plasticity is at its peak, allowing for optimal learning and integration of compensatory techniques. Unfortunately, due to insufficient monitoring, limited awareness of ASD, delayed symptom recognition, and a lack of allocated resources, the average age of diagnosis remains around 5 years.
This situation highlights the critical need to raise awareness about autism and promote early access to appropriate interventions. Early diagnosis of ASD is essential to maximizing the chances of an autonomous and improved quality of life for individuals with autism and their families, making it a major public health priority.
Neuro-Archaeology
Yehezkel Ben-Ari, co-founder of B&A Biomedical, introduced the concept of Neuro-Archaeology in 2008 and has continued to develop it throughout his career. This theory posits that the brain establishes the foundations of its future development in utero, including predispositions to certain pathologies.
During the critical period of pregnancy, pathological events or external stressors, such as genetic mutations, pesticide exposure, stress, alcohol, or environmental factors, can disrupt brain maturation. These disturbances lead to mispositioned and improperly connected neuronal networks, directly contributing to neurological and psychiatric disorders by impairing essential brain functions.
These alterations, first identified in animal models of brain malformations, have also been observed in humans. In some cases, the resulting disorders may not manifest until much later in life.
However, numerous experimental and clinical data suggest that ASD “originate” in utero. This concept of Neuro-Archaeology is at the heart of our research. Building on these observations, we adopted an approach based on the unbiased exploitation of data available in maternity hospitals, aiming to determine whether it is possible, through cross-referencing, to identify typical patterns that could assess the significant risk of developing ASD from the first few days of life. Additionally, we seek to uncover new potential research avenues in understanding and identifying risk factors for the in utero pathogenesis of such neurodevelopmental disorders.
Our Goal
Our research project, Pelargos, aims to develop a medical device designed to:
- Predict the risk of developing ASD after birth, allowing for earlier diagnosis through continuous monitoring of development. This early detection would enable the prompt initiation of psychoeducational therapies crucial for the patient’s future autonomy.
- Identify new biomarkers related to ASD within clinical data, paving the way for new research avenues that expand our understanding of the origins and developmental alterations leading to ASD.
Our Method
Using cutting-edge data mining techniques, particularly machine learning, our unbiased analysis of health data collected throughout pregnancy, birth, and the early days of life opens up new perspectives in understanding ASD.
Medicine 2.0 and Machine Learning
Machine Learning is a branch of artificial intelligence (AI) that has revolutionized how we process and analyze data. Using sophisticated algorithms and statistical models, this technology enables computers to learn from data and improve with experience, instead of being explicitly programmed to perform tasks..
In the age of Big Data, machine learning plays a crucial role. It allows us to extract valuable insights from large amounts of data, even when they are complex or unstructured, in real time. This opens up new possibilities for informed decision-making and solving complex problems. Additionally, machine learning can detect hidden patterns and relationships within data, revealing unexpected insights and stimulating scientific discovery.
The healthcare field generates vast amounts of medical information (clinical, genomic, medical imaging, etc.). With its predictive power and ability to detect patterns, machine learning offers new opportunities to enhance healthcare, medical research, and our understanding of diseases. The analyses and interpretations that it provides in real-time on these large datasets can help healthcare professionals make more accurate and informed decisions, identify risk factors, predict disease progression, and personalize treatments based on individual patient characteristics.
By combining medical expertise with the capabilities of machine learning, we are on the brink of an exciting era where technological advancements can truly transform individuals’ lives and have a significant impact on global health.
Achievements
Our preliminary work, published, has demonstrated the full potential of this ambitious project.
B&A Biomedical was also awarded the E-Health Trophy in the Big Data / AI category in 2022 for this project.
Contact Us
If you are interested in our expertise and the scope of our services, or if you would like to share an idea for a collaborative project, please feel free to contact us :
To do so, kindly use the form below. We will respond as quickly as possible. (All fields are mandatory)
