A major new release now makes available genetic and behavioural information about over 1,500 youth with autism who were hospitalised in six U.S. psychiatric units. Most previously studied children lived at home or in outpatient settings. This dataset highlights those with the most severe autism challenges—often called "profound autism"—featuring very limited speech and high dependency on help. Such individuals have rarely been included in past research.

The project is called the Autism Inpatient Collection (AIC). It is led by Dr Matthew Siegel of Boston Children’s Hospital and funded by the Simons Foundation Autism Research Initiative (SFARI) with support from the Nancy Lurie Marks Family Foundation. According to the AIC website, enrolment started in March 2014 and concluded in April 2024, gathering data from 1,555 young people aged 4 to nearly 21.

The Autism Inpatient Collection (AIC) has released a uniquely comprehensive dataset that combines several interconnected layers of information, offering an in-depth look at children and adolescents with profound autism. At the core of this release is phenotypic data, which includes detailed assessments of each participant’s behaviour, communication skills, emotional regulation, cognitive ability, sleep patterns, daily functioning, and even the stress and coping levels of their carers. These real-world observations offer helpful information about how autism affects everyday life.

Complementing this behavioural information is whole-exome sequencing, a form of genetic analysis that focuses on the protein-coding regions of DNA—areas known as exons. These regions are crucial because they hold the instructions for making proteins, which carry out most of the body’s functions. By sequencing both the child and their biological parents, researchers can identify rare or inherited gene variants that may help explain certain behavioural patterns or neurological traits seen in severe autism.

To make this data even more robust, biosamples such as blood or saliva were collected from participating families. These samples are securely stored and managed by a biobanking service called Sampled, ensuring they remain available for future scientific use. Together, these three components—observable traits, genetic code, and physical samples—form a powerful, interconnected dataset that gives researchers the tools to explore the biological and environmental roots of autism with high support needs.

Children with profound autism—who often exhibit intellectual disabilities or speak very little—make up roughly 27% of community samples, according to data from the U.S. Centres for Disease Control. However, such individuals are vastly under-represented in autism research, with studies including them shrinking from 95% in 1990 to only 35 % in 2013. Without robust data on this group, scientists cannot fully understand autism's most severe forms or design effective care strategies.

Dr Siegel highlights that this dataset “is the largest single collection of information on hospitalised children with autism, many of whom meet the criteria for profound autism”. He and his team aim to use this rich combination of genetic and behavioural data to develop more accurate, personalised treatments. These would help reduce serious issues such as aggression, self-harm, and emotional breakdowns.

To understand the significance of this new autism research, it helps to look closely at three key concepts: phenotypic data, whole-exome sequencing, and profound autism—and how they connect.

Profound autism refers to individuals who face the most intense challenges on the autism spectrum. These are children and adolescents who often speak very little or not at all, may have intellectual disabilities, and usually require round-the-clock support in daily life. Because their needs are so complex, they have historically been left out of many studies, especially those that focus on higher-functioning individuals who can attend school or therapy in outpatient settings.

To understand this group better, researchers collect phenotypic data. This means gathering detailed information on how a person behaves, thinks, feels, communicates, and functions day-to-day. For example, it includes things like sleep patterns, emotional control, social responses, and even how stressed their parents feel while caring for them. Such data helps paint a clear picture of their lived experience—something numbers alone can’t explain.

At the same time, scientists also use whole-exome sequencing, a method that scans the protein-making parts of a person’s DNA. Proteins are like the body’s workhorses, and any errors in the genes that create them can lead to disorders in brain function or behaviour. By studying the DNA of children with profound autism—and their parents—researchers can look for patterns or mutations that might explain why certain behaviours or difficulties arise.

When you connect these three elements, the picture becomes powerful: scientists now have both what is happening in a child’s daily life (phenotypic data) and why it might be happening at the biological level (whole-exome sequencing). And because the study focuses specifically on those with profound autism, this information has the potential to shape better support systems, treatments, and interventions tailored to those who need it most.

The AIC collected various assessments, including behaviour and cognitive tests (such as IQ or attention checks), emotional control surveys, sleep tracking data, and feedback from parents regarding their stress levels and confidence in managing their child's needs. Genetic information is matched to the child and, when possible, their biological parents. All records avoid personal details and are available to researchers through a secure portal called SFARI Base.

Kelsey Martin, SFARI’s executive vice president for autism and brain science, says this data pool is a “valuable resource” that could deepen our understanding of autism in those who require the most help.

Researchers worldwide can now access genetic and behavioural information after receiving approval. The dataset is expected to support new studies, including those testing “wearable biosensors” to predict aggressive behaviour before it starts.

This data resource marks a shift toward addressing a significant gap in autism science: the most vulnerable individuals have finally been included in rigorous study. With deeper insight into how genes, behaviour, and emotions link together, future care plans—from medication to training programmes—may become far better suited to those with profound needs.