With only 20 seconds of observation, artificial intelligence can detect children with autism

In the new study, the researchers used a wrist sensor and an AI algorithm to detect 'internal disorders' in children with an accuracy of up to 81% in just 20 seconds.

Anxiety and depression are alarmingly increasing in children. According to statistics, 1 in 5 children suffers from one of these conditions. Worryingly, many children have anxiety or depression before going to school.

Researcher Ellen and Ryan McGinnis, a representative of the University of Vermont team, successfully developed a wrist sensor to screen for symptoms of anxiety and depression in young children.

However, it is difficult to detect psychological disorders , also known as this internal disorder because symptoms often do not release to the outside. Therefore, even parents, teachers and doctors often misdiagnose their condition.

Psychological disorders in children are not simple. If left untreated, children with internal disorders often face the risk of drug abuse and the rate of subsequent suicide is many times higher.

'There is a need to find a way to screen early children with psychological disorders from which to directly take care of, treat and help them get better soon,' said Ryan McGinnis, a biomedical engineer. University of Vermont (USA) said.

So the McGinnis biomedical engineer collaborated with Ellen McGinnis, a clinical psychologist at the University of Vermont and colleagues at the Department of Psychiatry, University of Michigan, Maria Muzik, Katherine Rosenblum and Kate Fitzgerald, jointly develop a tool that can help screen children with internal disorders to help them get access to treatment soon. Their work was published in the journal PLOS ONE, issued on January 16.

The team used an 'emotional attack' test , a research method designed to find specific behaviors and sensations like anxiety. Researchers apply this test to 63 children. Some children in this group are diagnosed with an endocrine disorder.

In the experiment conducted, the children were led into a dimly lit room, while a researcher gave a choice board that included 'I have this thing to show you' and 'Keep quiet. Let it not wake up '.

The University of Vermont's new 20s method has an accuracy of up to 81%.

At the end of the room is a sealed box. Later, the researcher quickly peeled off the cloth covering the box, then opened the box, taking out a fake snake. They were then reassured by the researcher and allowed to play with the snake.

In the meantime, the researchers still observed and recorded the reaction of the group of children since the glass-covered cloth was removed and they saw the snake. Unlike previous experiments, this reaction is recorded not through video but with a sensor worn on the body.

This sensor will monitor every child's movements and then, the researchers use an AI algorithm to analyze these movements and find out the difference between children with anxiety and depression with children. rest. After processing the motion data, AI math detected the difference in the movements of these two groups of children.

'The children with internalized disorders move differently from normal children , ' said Ryan McGinnis.

The algorithm shows that the child's reaction just before the snake appears is the most obvious because people with psychological disorders tend to run away and turn their backs when they feel a potential threat.

Clinical psychologist Ellen McGinnis said the results are in line with psychological theory. Specifically, children with endocrine disorders often express anxiety, more precaution and running behavior is often classified by the researchers as a negative reaction.

This new tool helps distinguish children with psychological disorders and normal children with an accuracy of 81%, higher than the current popular method of questionnaires. With this new method, only 20 seconds of observation can identify a child with a psychological disorder, while traditional video methods can take up to several months. This opens up the possibility of conducting large-scale screening and early treatment

The number of children with anxiety and depression is increasing.

The world currently has about 20% of children with psychological disorders with symptoms such as anxiety and depression. However, symptoms of this disease are often difficult to identify.

The researchers said the next step of the study improved the algorithm and completed additional tests to analyze voice data and other information, thereby enabling discriminatory technology. depression and depression.

The group's ultimate goal is to develop a review that can be used right at school, to quickly screen children with anxiety and depression. From there, help children to be treated soon.