The research was undertaken in partnership with The Black Dog Institute, which recommended four interventions, including mindfulness, sleep, diet and exercise programs.

The AI data is gathered in a smartphone app. Participants complete a questionnaire, including their general information such as age and cultural background, along with levels of depression, anxiety and stress. “AI will look at this information and, from its statistical knowledge, will do the number crunching to suggest an intervention,” says Prof Gupta.

The student then completes the same questionnaire two weeks later. “From these comparison surveys we get to know whether levels of stress have gone down or up,” he explains.

“This data is again ingested by AI and then AI algorithms update their understanding, so when new people complete the survey they will receive interventions more and more accurately.”

In the future, Prof Gupta hopes that AI can be refined to be instrumental in identifying more complex mental-health issues and even be able to recommend therapies such as CBT (cognitive behavioural therapy) by capturing more data, including behavioural information, such as movement and smartphone usage.

Meanwhile, Professor Truyen Tran, head of AI, Health and Science at Deakin University’s Applied Artificial Intelligence Institute, is spearheading a project to develop an AI tool to improve universal access to early diagnosis for babies with cerebral palsy.

“Movements of infants that are nine to 16 weeks, to the untrained eye, are quite random and that’s exactly where machine learning comes into play, because we can detect certain patterns that are indicative of future disorder,” explains Prof Tran.

“So far, in collaboration with the Cerebral Palsy Alliance and hospitals throughout Australia, we have collected and analysed several thousand video clips of babies,” he explains. “The algorithm learns from the experts who have provided the annotation for those videos and then the algorithms learn how to detect the kind of movement that is highly indicative of developing cerebral palsy in the future.”

Currently, cerebral palsy is mostly diagnosed when the parents notice some abnormality in the baby’s movements when they start crawling and walking, between one and one-and-a-half years old. “Intervention is less effective as the baby gets older,” says Prof Tran.

Earlier diagnosis at nine to 16 weeks could significantly improve the outlook for the child, he asserts. “There are a lot of interventions at an early age, such as 10 weeks or 20 weeks, that are effective and reduce the effect of disability later.”

Smartphone-accessed technology will make screening for cerebral palsy accessible to most people, no matter your location or economic status, he adds.

“Right now, this early screening, especially in low-income countries is really missing,” says Prof Tran. In the future, video clips could be recorded on mobile phones and the AI algorithm would analyse the results in real time, which would inform the parents of the results and refer them to specialist clinicians if necessary.

“The global efforts of this research as a whole is to develop the AI solutions, so that we can provide access to every single family in the world, because most people will now have a smartphone, so when the smart AI solutions become available in the next few years, every single infant in the world will have access to this early screening.”

Such exciting developments to medical care already — and this is just the beginning.

To find out more about how Deakin research is leading the way, click here.