The prevalence of Autism Spectrum Disorder (ASD) varies across different countries and regions, but it is generally considered a common developmental disorder. Early detection of ASD is crucial to ensure timely access to appropriate interventions, which can significantly improve outcomes for affected children and their families.

While there is no cure for ASD, early spotting and intervention can help children develop important social, communication, and cognitive skills, potentially reducing the severity of symptoms and improving long-term outcomes.

What is Autism Spectrum Disorder (ASD)?

ASD stands for Autism Spectrum Disorder. It is a developmental disorder that affects a person's social interaction, communication, and behavior.

Grown ups with ASD often have difficulties understanding social cues, engaging in conversation, and may exhibit repetitive behaviors or interests.

The term “spectrum” is used because the symptoms and severity of ASD can vary widely among individuals, with some people requiring significant support, while others may lead independent lives.

How common is ASD?

According to the Centers for Disease Control and Prevention (CDC) in the United States, the prevalence of ASD is approximately 1 in 54 children (~1.85%) in 2016. Other newer research centers like National Institute of Mental Health suggest slightly higher numbers (~2.3%). The prevalence is significantly higher in case of boys (~3%) than girls (0.9%). It is important to note that ASD prevalence has increased over the years, partly due to increased awareness and improvements in diagnostic methods.

Early intervention can significantly improve the quality of life for individuals with ASD, as it enables parents, caregivers, and professionals to address the unique challenges they face in social interaction, communication, and behavior.

By identifying ASD early on, tailored strategies can be implemented to support the child's development and enhance their potential to lead fulfilling lives. In this article, we will delve into various methods and studies aimed at recognizing the early signs of ASD, highlighting the importance of timely detection and intervention in promoting better outcomes for children on the autism spectrum.

Relevant New Studies, Books and Resources

Lung airway geometry as an early predictor of autism: A preliminary machine learning-based study, by Asef Islam, Anthony Ronco, Stephen M. Becker, Jeremiah Blackburn, Johannes C. Schittny, Kyoungmi Kim, Rebecca Stein-Wexler, Anthony S. Wexler (2023)

This research paper is about a study that looks into a possible way to predict autism in children at an early stage. They do this by examining the structure of the airways in the lungs using chest CT scans. The study included 54 children, 31 with autism and 23 without autism, who were of similar age and gender.

The researchers used two computer methods, called PCA and SVM, to analyze the CT scans and find differences between the lung airways of autistic children and those without autism. They focused on the angles where the airways branch out.

The study found that this method was able to correctly identify autism in 94% of cases. However, it also mistakenly identified some healthy children as having autism 22% of the time. This suggests that the angles of lung airways might be a useful clue for early autism detection, but more research is needed to improve the method's accuracy.

The study proposes an innovative AI (machine learning) based method for identifying ASD by analyzing lung airway geometry, which has not been widely explored as a potential biomarker for autism. This opens up a new area of research that could lead to more effective early detection techniques.

Chest CT scans are a non-invasive method for examining the child's biological structure, making it a more comfortable and accessible approach for detecting early signs of ASD in children compared to some other diagnostic procedures.

It is important to note that the study is still preliminary and has some limitations, such as the lower specificity rate of 78%, which means there were false positives (healthy children identified as having autism). Despite these limitations, the research is groundbreaking because it explores a previously under-investigated aspect of early autism detection.

Vision-assisted recognition of stereotype behaviors for early diagnosis of Autism Spectrum Disorders, by  Farhood Negin, Baris Ozyer, Saeid Agahian, Sibel Kacdioglu, Gulsah Tumuklu Ozyer (2021)

This research is about using computer technology to help doctors diagnose ASD more easily and accurately by recognizing certain behaviors in children. The researchers collected videos of children with ASD displaying common repetitive actions, known as stereotypic behaviors, which were selected with the help of professional clinicians.

The study tested different methods for analyzing these videos and identifying ASD-related behaviors. They used several computer techniques, including the Bag-of-Visual-Words approach, and machine learning tools like Multi-layer Perceptron (MLP), Gaussian Naive Bayes (GNB), and Support Vector Machines (SVM).

Additionally, the researchers developed a system that tracks the movements of children in the videos using a method called LSTM, which helps the computer understand how the children's movements change over time. They compared their results to two other advanced computer methods, called ConvLSTM and 3DCNN.

The study found that the best results were achieved when using a specific technique called Histogram of Optical Flow (HOF) with the MLP classifier. The promising findings suggest that a computer system based on action recognition could potentially help doctors diagnose ASD more reliably, accurately, and quickly.

There are one main limitation in this method. The accuracy of the system depends on the quality of the recorded videos. Poor lighting, low resolution, or obstructions in the video can affect the performance of the computer algorithms in recognizing stereotypic behaviors.

Eye Tracking-Based Diagnosis and Early Detection of Autism Spectrum Disorder Using Machine Learning and Deep Learning Techniques, by Ibrahim Abdulrab Ahmed, Ebrahim Mohammed Senan, Taha H. Rassem, Mohammed A. H. Ali, Hamzeh Salameh Ahmad Shatnawi, Salwa Mutahar Alwazer  and Mohammed Alshahrani (2022)

This research article is about using a technology called eye tracking to help detect Autism Spectrum Disorder (ASD) in children. Eye tracking follows the movement of a person's eyes while they look at images or videos, and it can provide information about how children with autism visually process information differently from others.

The study tested three different artificial intelligence (AI) techniques to analyze the eye-tracking data and diagnose autism. The first technique used two types of neural networks, which are computer algorithms inspired by the human brain. These networks were trained using a combination of two methods called Local Binary Pattern (LBP) and Grey Level Co-occurrence Matrix (GLCM). This technique achieved a very high accuracy rate of 99.8%.

The second technique used advanced AI models called convolutional neural networks (CNN), specifically GoogleNet and ResNet-18, which were already trained to recognize patterns in images. These models also showed high performance, with accuracy rates of 93.6% and 97.6%, respectively.

The third technique combined deep learning (GoogleNet and ResNet-18) with another AI method called Support Vector Machine (SVM). The deep learning models were used to extract important features from the eye-tracking data, while the SVM classified the features to determine if a child had autism. This hybrid approach achieved accuracies of 95.5% and 94.5% for the GoogleNet + SVM and ResNet-18.

The accuracy of the eye-tracking method depends on the quality and quantity of the eye-tracking data collected. Factors like participant fatigue, distraction, or calibration errors during data collection can affect the performance of the AI models.

Unmasking Autism, by Devon Price (2022)

In terms of books, Devon Price summarize some of the most important things you need to know about ASD including the term “masked Autism”. Dr. Price blends personal experiences with historical and social science research to shed light on the often-overlooked experiences of masked Autistic people, who frequently struggle for years before discovering their true identities and are more likely to be marginalized due to race, gender, sexual orientation, and other factors.

The book promotes unmasking by offering exercises that encourage self-expression and celebrating neurodiversity, ultimately calling for greater public acceptance and accommodation of difference so that both Autistic and neurotypical individuals can live authentically.

Conclusion

Several innovative approaches have been explored in recent years to improve early detection of Autism Spectrum Disorder (ASD). These methods often complement traditional diagnostic methods, such as clinical observations and interviews.

An ideal way to diagnose ASD could involve a combination of the innovative methods discussed earlier with traditional diagnostic methods, creating a comprehensive and holistic approach to early detection.

Begin with initial screening using parent questionnaires and developmental assessments by pediatricians or primary care providers. These tools can help identify children at risk for ASD and guide further evaluation.

The future of detection should be non-invasive like eye tracking and wearable devices to monitor visual attention, physiological markers, or behavioral patterns. These tools can provide additional information to support early detection efforts.

The use of machine learning and artificial intelligence techniques to analyze data from eye-tracking, CT images and other imaginary tests could radically improve the quality of detection. AI can help identify patterns that may indicate ASD and support clinicians in making a more accurate diagnosis.

If you are looking for more resources about ASD, check out further books in our ASD book category.

Featured on Joelbooks