Polish AI innovation detects depression in seconds through eye movement
Polish scientists have developed an innovative system that can detect depression and anxiety in just ten seconds by analysing eye movement. Thanks to artificial intelligence, the accuracy of this method reaches 70%, with researchers optimistic about the potential for further improvement.
Polish artificial intelligence may revolutionise the diagnosis of mental disorders through the analysis of eye movement. Researchers from three Polish universities have developed an AI-based system that facilitates the rapid detection of such disorders. This research was detailed in the International Journal of Marketing, Communication and New Media.
The study involved 101 participants, including patients with depression, individuals with social anxiety, and a control group of healthy participants. The subjects had to watch images of faces showing various emotions for ten seconds, while special sensors in eye trackers recorded their eye movements. The collected data was then used to create 'gaze paths', which were analysed using neural networks.
Eye movement patterns as an indicator of psychological condition
"Eye movement patterns can provide objective data about our psychological condition. We observe a tendency in depressed individuals to focus their attention on negative stimuli," said Dr Karol Chlasta from Kozminski University, co-author of the study and an expert in artificial intelligence, in a conversation with PAP.
He added, "Individuals with social anxiety demonstrate increased activity in scanning faces, a psychological phenomenon known as hyperscanning." The co-author also noted that this is shown by prolonged face-scanning paths, indicating these individuals' sensitivity to social stimuli.
The method is effective in up to 70% of cases
Psychologists and AI experts, including Dr hab. Krzysztof Krejtz and Dr hab. Izabela Krejtz from SWPS University, along with Dr Katarzyna Wisiecka from AEH in Warsaw, were involved in the project. The method achieves 60–70% accuracy in detecting cases of depression and social anxiety, comparable to traditional methods.
The new approach is quicker and less burdensome for the patient compared to traditional methods, making it easier to monitor changes in mental state. The system can be integrated with everyday devices, such as laptops, smartphones, or VR goggles. Dr Chlasta compares it to smartwatches that track sleep patterns; in this instance, it analyses vision.
Expanding research into voice analysis
Researchers are also exploring the application of AI to voice analysis for diagnosing depression and neurological disorders. Dr Chlasta mentions that changes in voice can be an early warning signal for conditions like depression, dementia, or Alzheimer's disease, allowing for quicker responses and consultations with a doctor.
"In many disorders, our voice changes subtly. It's akin to an overworked computer running somewhat slower, switching between different tasks. In humans, changes can be observed in speech organs, which can be difficult for individuals to detect, but a system based on artificial neural networks can identify these from short speech excerpts," explains Dr Chlasta.
The creators of the system emphasise that depression and social anxiety are among the most prevalent mental disorders, with numbers continuing to rise. WHO forecasts suggest that by 2030, depression will be the most frequently diagnosed disease globally. In Poland, around 4 million people already suffer from it, though many cases remain undiagnosed. Quick analysis of eye movements can offer valuable insights into mental states and serve as an essential prompt for seeking medical advice.
Need for further research and systemic changes
Further research is necessary to enable the widespread implementation of this new method. Dr Chlasta explains that additional data collection, currently not systematic, is needed, and social trust in AI remains low. Without additional data, moving beyond laboratory conditions to demonstrate prototypes in real-world settings will be challenging. Systemic changes are also required to facilitate the broader monitoring of mental health.
