The development of robust ML models in the mental health sector will enable medical experts to draw more accurate diagnosis results in a shorter period of time.
FREMONT, CA – The contributions of artificial intelligence (AI) to the healthcare sector are plenty, ranging from cancer to forecasting of premature death. The development of robust machine learning (ML) models have paved the way for innovative diagnosis approaches capable of identifying subtle and sophisticated conditions.
Among the conditions challenging to spot, anxiety and depression are one of the most prominent, with one in five children falling victim. Furthermore, the inability of toddlers to articulate their emotions makes it challenging for medical professionals to diagnose their ailments.
Hence, psychologists must be equipped with robust, objective tests to assist the kids when they are most vulnerable. In the case of mental disorders, healthcare providers need to diagnose the conditions early on. By treating children when their brains are in the developing stage, medical experts can avert adverse outcomes such as the risk of suicide and substance abuse.
A group of researchers recently tested the utilization of AI model to identify depression and anxiety. Seventy-one children between the ages of 3 and 8 participated in the Trier-Social Stress Task, a protocol designed to induce psychological stress in subjects. The children had to develop a three-minute story, which would supposedly be judged on its content.
The subjects were offered neutral and negative feedback during the test. The task was designed to induce stress in children and put them in a mindset of being judged by others. The researchers leveraged the ML model to analyze the audio recordings of each of the stories made up by the subjects. The results drawn by the ML model matched the structured clinical interview and parent questionnaire.
The AI algorithm identified the children with an internalizing disorder with an accuracy of 80 percent, which compared well with the accuracy of the results drawn from the parent checklist. Also, it was able to do it a matter of a few seconds. The researchers are planning on developing an enhanced algorithm and integrate it into a universal screening tool for clinical use. The algorithm can also be used in a smartphone app to record and analyze results immediately, alerting parents to any potential problems in their children.