Computer-Assisted Mental Health Solutions in Depression Treatment
Article information
Recent Fesearch on AI Applications in Depression Treatment
It is becoming an essential part of the health system (Abd-Alrazaq et al., 2023; Wright et al., 2019) With the advent of artificial intelligence (AI), the potential applications of computer-assisted mental health (CAMH) solutions are more expanding (Xu et al., 2021).
Depression analysis (DEPRA) chabot is combined with a social media platform to provide a practically structured and authoritative early detection solution for depression, which is built on Dialogflow as a conversation interface. This guide in cludes 27 questions using 2 scientific scoring systems; the Hamilton Depression Scale and the inventory of depressive symptomatology (Health Information Science, 2021). The DEPRA chatbot as a messenger in facebook is a web and mobile application-based platform, which is the proxy between the user and it. It has suggested challenges as a feasible and easily accessible option for early detection of metal health problems (Kaywan et al., 2023). DEPRA is used for the early detection of depression using 2 scientific scoring systems called QIDS-SR (Quick Inventory of Depressive Symptoma-tology-Self Report) and IDS-SR (Inventory of Depressive Symptomatology- Self Report), but it does not provide a
solution for the treatment of depression.
In the United States, Woebot is a widely used chatbot designed to help individuals manage their mental health using cognitive behavioral therapy (CBT) techniques. Although it is frequently used for routine therapeutic sessions and is particularly targeted at addressing symptoms of depression and anxiety, including postpartum depression (Fitzpatrick et al., 2017; Woebot Health, 2024). Woebot Health started a pivotal clinical trial to be approved by the U.S. Food and Drug Admi nistration (FDA) as an investigational digital therapeutic for postpartum depression since 2023 that was granted breakthrough device designation by the FDA in 2021.
In contrast, Rejoyn, an FDA-approved medical device used as an adjunctive therapy for major depressive disorder, was developed by Otsuka Pharmaceutical, a subsidiary of Otsuka Holdings in Japan, in collaboration with the U.S.-based digital therapeutic company Click Therapeutics. This digital therapeutic app, launched in 2024, combines CBT with cognitive training exercises (Business Wire, 2024). It involves training users to infer emotions from facial expressions shown on the screen and modifying their approach to grasping objects, thereby improving symptoms. The app implements each therapeutic approach three times a week over a total of six weeks to alleviate depressive symptoms (Pokhrel et al., 2024).
In South Korea, research is underway on using AI chatbots, such as the ‘SimSimi’ robot, for treating depression (Choi, 2022). Domestic companies, including Y-Brain, Meditrix, and Rowan, are also actively engaged in research and development in this area.
Strengths and Limitaions of AI Applications in Deression Treatment
Research has shown that chatbots in the field of mental health counseling can improve accessibility by providing unbiased and common-sense information. It is known that chatbots can be useful for individuals experiencing mild mood disorders or those seeking to verify simple psychiatric infor mation. However, when direct psychiatric intervention is required, such as in cases of suicide or domestic conflict, AI cannot replace expert counseling with skilled intervention. Since each individual's situation and environment are unique, the experiences shared between a professional and a patient cannot be replicated by AI. The therapeutic role based on the challenges arising from human relationships is an area where AI cannot and should not intervene.
On the other hand, CAMH solutions are continuously accessible, effective for nonsevere cases, and can play a com-plementary role in clinical treatment. Patients can enhance their awareness of their own psychiatric issues, maintain regular attention to their mental health, and thus prevent progression to more severe conditions. This self-awareness can also help recognizing the need for treatment. Therefore, if combined with digital therapeutics, the effectiveness on treatment in clinical settings could be significantly improved.
Several researches suggest that chatbot-based counseling tools can be particularly effective as supplementary aids for postpartum depression, which is significantly influenced by hormonal changes (Miura et al., 2023; Suharwardy et al, 2023). Since over 80% of postpartum depression cases are nonsevere and often improve within 2-3 weeks after de livery, these tools are well-suited for addressing nonseverer forms of postpartum depression. Mothers and their partners experiencing prenatal or postpartum depression can facilitate a shared understanding on both pregnancy and childbirth process, as well as offer opportunities for overcoming the depression together. Additionally, mothers living in underserved areas can benefit from AI tools, which can provide access to support and treatment for postpartum depression. As an alternative to the lack of healthcare facilities and healthcare workforce
infrastructure in medically underserved areas, AI-based tools are expected to be particularly helpful in treating perinatal depression.
Research and development are steadily progressing in com bining bio-data measured from wearable devices, such as electroencephalogram (EEG) (brainwaves) and heart rate variability (Pawlowski et al., 2017), and AI voice, conversation recognition (Gaffney et al., 2019), and facial expressions technologies (Fu et al., 2023). By integrating these data, we aim to predict and diagnose users' mental health conditions. With such dramatic advancements, recently the European Union passed the world's first AI regulation law to restrict the indiscriminate development and use of AI (World Laws Information Center, 2024). The law prohibits human rating systems, emotion recognition systems, and personal behavior prediction systems, suggesting that the development of technologies capable of identifying individuals through voice, facial expressions, and conversations will be banned. The development of AI using specific data that enabled personal identification is expected to be somewhat limited.
The recent development and application of chatbots are also expected to have great expectations in the diagnosis and treatment of depression during pregnancy or after childbirth. Through the diagnosis of depression, severe cases will be recommended an immediate consultation with a specialist, and nonsevere cases will be provided various treatment interventions such as meditation and music therapy. This approach is expected to be especially useful for populations in medically underserved areas with limited access to healthcare. Although the diagnosis of depression related to pregnancy and infertility allows for treatment intervention according to severity, and also its utility is expected to increase, but the limitations of mobile-based depression treatment in practice are still debatable.
Conclusion
The continuity and effectiveness of mental care can depend on an individual's voluntary commitment. In addition, treatment interventions that are based on interpersonal relationship problems according to the individual's situation and environ-
ment cannot be treated through mechanical intervention. Because of these limitations, although CAMH solutions are primarily used as supportive tools in mental health treatment, AI has increasingly gained momentum in behavioral health interventions. Considering the acceleration of AI development as a supplementary tool in mental health treatment and its po tential impact, researchers need to continue efforts to enhance the performance and effectiveness of chatbots. Given the potential of chatbots in prescreening assessment and early detection and treatment intervention for depression and anxiety, increased investment and research are warranted. By advancing chatbot development, researchers and developers, as well as clinical psychiatrists, can enhance their capabilities and improve mental health treatment outcomes.
Conflict of Interest
The authors have nothing to disclose.
Acknowledgments
This research was supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute, funded by the Ministry of Health and Welfare, Republic of Korea (Grant No. RS-2024- 00427361).