AI achieves 80% in Autism detection

Dear Reader,

Flipped.ai’s weekly newsletter read by more than 75,000 professionals, entrepreneurs, decision makers and investors around the world.

In this newsletter, we feature a new study from Karolinska Institutet, published in JAMA Network Open, which reveals that AI can predict autism spectrum disorder (ASD) in toddlers with about 80% accuracy. Led by Dr. Kristiina Tammimies, the research used data from the Simons Foundation SPARK database, analyzing information from 30,660 participants to enhance early detection of ASD using basic medical and background details.

Before, we dive into our newsletter, checkout our sponsor for this newsletter.

Streamline Your Entire Business with a Free CRM

HubSpot provides a comprehensive customer relationship management platform to help you grow. With powerful features to manage leads and improve customer relationships, HubSpot’s CRM is completely free, with no restrictions on users or data, making it ideal for businesses at any stage.

AI model shows promise in early Autism detection, achieving 80% accuracy

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by difficulties in social interaction, communication, and repetitive behaviors. Early diagnosis and intervention are critical for improving developmental outcomes. Traditional diagnostic methods are time-consuming and costly, leading researchers to explore alternative approaches. A recent breakthrough by Karolinska Institutet introduces a machine learning model, 'AutMedAI,' which has shown an impressive 80% accuracy in predicting autism in children under two years old. This article delves into the study’s details, its implications, and future directions for AI in autism diagnosis.

The study

Overview of the research

The study, led by Associate Professor Kristiina Tammimies at Karolinska Institutet and published in JAMA Network Open, investigates a machine learning model that evaluates 28 parameters to detect autism. Utilizing data from the SPARK (Simons Foundation Powering Autism Research for Knowledge) database, which includes comprehensive information on autism diagnoses, the research aimed to develop a model capable of early autism detection in young children.

Development of AutMedAI

The researchers developed and evaluated four machine learning models to identify autism-related patterns. The models were trained on a dataset of approximately 30,000 individuals, split evenly between those diagnosed with autism and those without. The AutMedAI model emerged as the most effective, achieving nearly 80% accuracy in predicting autism in children under two years old. This model’s performance highlights its potential in facilitating early diagnosis and intervention.

Machine learning techniques in Autism detection

How machine learning models work

Machine learning involves training algorithms to recognize and predict patterns from large datasets. For autism detection, these algorithms analyze features associated with developmental milestones to identify patterns indicative of autism. The AutMedAI model used 28 parameters, including age at first smile and age at first words, chosen for their relevance to early autism indicators.

The training process involved supervised learning, where the model was trained on labeled data and validated on separate datasets. This approach allows for assessing the model’s accuracy and reliability in making predictions on new data.

Key features analyzed

The features analyzed by AutMedAI include:

• Age at first smile: Reflects early social engagement and developmental milestones.

• Age at first words: Indicates language development and cognitive progress.

• Eating difficulties: Can be associated with sensory sensitivities and behavioral challenges in autism.

• Developmental milestones: Various milestones were assessed to gauge overall developmental progress.

These parameters were selected based on their potential to serve as early indicators of autism, leveraging basic medical and background information to provide practical diagnostic support.

Methodology

Framework of the diagnosis

Above figure illustrates the framework for diagnosing autism in children using machine learning approaches. The framework outlines the process of data collection, model training, and evaluation, showcasing how machine learning algorithms analyze various features to predict autism.

Dataset

The dataset used for training and evaluating the machine learning models was gathered through an app designed for autism screening called ASDTests. Created by Thabtah et al. (2018) from the Nelson Marlborough Institute of Technology, the app collected data from children and was incorporated into the ML repository used for this study.

The dataset contains 20 features and two classes (normal or ASD). These features were derived from AQ-10 questionnaires and demographic information about the children, which were evaluated to identify variables associated with an increased likelihood of having ASD. Table 1 summarizes the specific characteristics of the dataset considered in the study, including the features collected and their relevance to autism detection.

Experiments

Comparison of classification methods

To evaluate the effectiveness of AutMedAI, several classification algorithms were compared, including K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), and Random Forest (RF). The comparison was conducted in both binary classification scenarios (normal vs. ASD) and multiclass environments.

• Accuracy: Measures the proportion of correct predictions.

• Precision: The ratio of true positive predictions to the sum of true and false positive predictions.

• Recall: The ratio of true positive predictions to the sum of true positives and false negatives.

• F1-Score: The harmonic mean of precision and recall, providing a balance between the two.

• Mean Squared Error (MSE): The average of the squares of the errors or deviations.

• Root Mean Squared Error (RMSE): The square root of MSE, providing error magnitude.

• R2 Score: Measures the proportion of variance in the dependent variable that is predictable from the independent variables.

In the experiments, a dataset of 20 features was used to create binary classifications. The correlation method was employed to determine the strength of associations between features. Results were compared with other existing methods to evaluate the performance of AutMedAI.

Splitting dataset

To achieve high accuracy, the dataset was split into training and testing sets, with 80% allocated for training and 20% for testing. This method ensured a robust evaluation of the model's performance. The complexity of ASD data necessitated the use of specialized hardware and software to build the ASD detection system effectively.

Results

Performance of classification methods

The results of the classification experiments are detailed in above table. The performance of KNN was notably less satisfactory, with an accuracy of only 52%. High misclassification rates were observed, indicating that KNN was less effective for detecting ASD in the studied dataset. The accuracy performance of KNN compared to the target and predicted values is illustrated in Figure 1. Given the high misclassification rates, KNN was not considered an appropriate classification approach for ASD detection.

Implications for early intervention

Importance of early detection

Early detection of autism is critical for effective intervention. Research has shown that early intervention can significantly improve developmental outcomes, including social skills, communication abilities, and overall quality of life. The high accuracy of the AutMedAI model suggests its potential to enable earlier diagnosis and intervention, leading to more timely support and improved outcomes for children with autism.

Potential benefits for families

Early detection through models like AutMedAI offers several advantages for families:

• Access to Resources: Early diagnosis allows families to access support systems sooner, aiding in navigating autism-related challenges.

• Informed Decision-Making: Early diagnosis enables families to make informed decisions about interventions, therapies, and educational support.

• Reduced Stress: Early identification reduces uncertainty and stress, providing clarity and direction for families.

The AutMedAI model’s ability to use basic data for early detection could also alleviate some of the burdens on healthcare professionals, allowing them to focus on complex cases and tailor their care based on the AI model’s results.

Challenges and limitations

Need for clinical validation

Despite its promising results, AutMedAI requires further clinical validation to ensure its reliability and effectiveness. Rigorous testing across diverse populations and settings is essential to confirm the model’s performance and generalizability. Ongoing research is needed to address these validation needs.

Ethical considerations

The use of AI in healthcare raises important ethical considerations, including data privacy and security. Ensuring that personal data is handled with strict privacy standards is crucial. Transparency in how AI models use data and make decisions is necessary to maintain public trust and ensure ethical use.

Limitations of the model

AutMedAI, while promising, has limitations. Its performance varied across different datasets, indicating the need for further refinement. The model may not account for all variables influencing autism diagnosis, such as genetic factors or nuanced developmental behaviors. Continued research is required to enhance the model’s accuracy and applicability.

Future directions

Enhancing model accuracy

Future research aims to improve AutMedAI by incorporating additional data sources and exploring new techniques. Integrating genetic information and more complex data, such as neuroimaging and detailed behavioral assessments, could provide a more comprehensive assessment and enhance the model’s accuracy.

Broader applications of AI in Autism research

The success of AutMedAI highlights the potential for AI in various aspects of autism research and diagnosis. Future applications could include:

• Analyzing Speech Patterns: AI could assess speech and language patterns to identify early communication difficulties.

• Assessing Social Interactions: Machine learning models could evaluate social behaviors to provide insights into social development.

• Personalized Treatment Plans: AI could assist in developing personalized treatment plans based on individual characteristics and needs.

Collaboration with healthcare professionals

Successful integration of AI into clinical practice requires collaboration between researchers, healthcare professionals, and families. Effective communication and cooperation are essential to ensure appropriate and ethical use of AI models. Collaboration can maximize the benefits of AI while addressing challenges and concerns.

Conclusion

The introduction of the AutMedAI model represents a significant advancement in AI-assisted autism detection. With nearly 80% accuracy in early diagnosis, this model has the potential to revolutionize the diagnostic process and facilitate earlier intervention. While further validation and refinement are needed, AutMedAI’s success demonstrates the promise of AI in enhancing autism diagnosis and improving outcomes for affected children. The future of autism diagnosis may indeed be shaped by these technological advancements, leading to more effective and timely interventions and a better quality of life for those affected.

Hire top-quality Indian tech talent with SourceTalent.ai by Flipped.ai

Need to build a world-class tech team in India without breaking the bank? SourceTalent.ai offers an AI-powered, cost-effective hiring solution just for you!

Key Benefits:

• Instant access: Tap into a vast pool of 24M+ Indian candidates with personalized recommendations.

• AI-powered matching: Our advanced algorithms connect you with candidates that perfectly fit your job requirements.

• Automated hiring: Simplify the process with AI-driven job descriptions, candidate screening, and tailored recommendations.

• Seamless video interviews: Conduct unlimited interviews effortlessly and gain valuable insights.

Why SourceTalent.ai?

• Affordable excellence: Prices start at just Rs400 / $5 per job posting.

• Top talent pool: Access a diverse selection of India’s best tech professionals.

• Efficient hiring process: Enjoy a streamlined recruitment process with video assessments.

• Global reach: US companies can also leverage India’s premier tech talent!

Get started today at SourceTalent.ai and take advantage of our exclusive launch offer: [Link]

For more information, reach out to us at [email protected].

Experience smarter, faster, and more affordable hiring with SourceTalent.ai!

Want to get your product in front of 75,000+ professionals, entrepreneurs decision makers and investors around the world ? 🚀

If you are interesting in sponsoring, contact us on [email protected].

Thank you for being part of our community, and we look forward to continuing this journey of growth and innovation together!

Best regards,

Flipped.ai Editorial Team