Imitate that one: dolphins spontaneously interpret human pointing gesture in a novel context.

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder caused by genetic and environmental factors affecting how people communicate and interact socially and behave. It is associated with a spectrum of symptoms that vary widely in severity. An early diagnosis, followed by programs of behavioral therapy, speech therapy, or structured support, can greatly enhance the quality of life of the individual. According to the data available, more than 2 million individuals are affected with ASD in India; with percentage prevalence of 0.11 for age group (1–18 year) and 0.09 for age group (0–15 year) in the rural and urban areas. The conventional diagnostic methods are mainly based on personalized evaluation and surveys, which take more time, making the symptoms more severe. This limitation stresses the importance of more advanced and precise methods that can enhance early diagnosis of ASD. There are research works that support early assessment and intervention but also challenges on the lack of empirically validated early interventions. However, the study concludes that early intervention is more cost effective and time saving, than a “wait and see” approach. Compared to other countries, many children in India are diagnosed much later than in developed nations like the US, UK, or Australia, where universal early screening is common. The US and Europe use AI-powered screening tools, digital apps, and teletherapy, whereas India is still catching up in technology adoption for ASD. In Indian scenario, Indian Scale for Assessment of Autism (ISAA) tool was developed to help in disability certification under the RPWD Act, 2016, using Indian-specific behaviors and expressions rather than Western standards. This throws an insight on the need for age-appropriate and multi-informant approaches to ensure accurate assessments and a more comprehensive approach to assessing ASD related disabilities in India. In an approach to automatically characterize brain Magnetic Resonance Imaging (MRI) images for detecting Autism Spectrum Disorders (ASD), by analyzing features from over 1,100 MRI scans of individuals with ASD, and identifies that there are structural differences in brain regions, such as the hippocampus, amygdala, and thalamus for individuals with ASD. These features serve as potential indicators of ASD and the framework shows promising accuracy, particularly in younger age groups, suggesting it could help with early ASD diagnosis. So, MRI (both structural MRI (sMRI) and functional MRI (fMRI)) helps identify ASD in early stages and provide elaborate information of the brain’s shape and connectivity, revealing ASD-related biomarkers such as cortical thickness and gray matter density. sMRI helps to detect anatomical differences in the brain of individuals with ASD and significant structural differences in their brain. sMRI can also detect abnormal brain growth patterns as early as 6–12 months of age. fMRI helps identify differences in brain activity and connectivity in ASD affected individuals, even before behavioral symptoms begin to appear. Recently, machine learning (ML) has emerged as an effective method for ASD detection, collecting data from an extensive range of sources in addition to behavioral patterns, and brain imaging data. The work in illustrates how ML is used for the identification of ASD using brain imaging and the methods for diagnosing ASD using functional and structural MRI data is available in. The preprocessing stage uses fMRI and sMRI for feature extraction with tools like FreeSurfer for structural analysis. Various ML algorithms were used for classification, including Support Vector Machines (SVM) that work efficiently for datasets with higher dimensions and medium-sized datasets, which achieved an accuracy between 60% and 95%. In the work proposed by, ASD is identified and predicted based on multi-site structural MRI with the help of machine learning. The two public datasets from the University of California, and Irvine (UCI) repository are used in combining features like age, gender, and diagnostic responses. The preprocessing includes handling missing values, one-hot encoding, and feature selection were included with the model. The Autism Brain Imaging Data Exchange (ABIDE) dataset was utilized, which contains numerous and diverse collections of MRI scans and phenotypic data from people with ASD and neurotypical controls. The results of this study concluded with the effectiveness of AI-driven techniques in ASD diagnosis, providing better accuracy and efficiency as compared to traditional methods.