Artificial intelligence (AI) in healthcare: an advancement that is here with no plans of slowing down. Even in the last several years, we have seen a massive surge in robotic surgeries, AI-predicted health assessments and treatment plans, and further detection of diseases through AI data analysis. But one area of healthcare that has seemed a bit more challenging for AI to interface with has been mental health. Psychology and mental health treatment have never been located in the black-or-white sphere of healthcare, but have always encompassed a gray space. A great piece of mental health care relies on human connection, emotions, and empathic understanding… things that robots and AI certainly aren’t known for. But perhaps bridging the gap is necessary, and learning ways of utilizing AI advancements can also be beneficial in a "gray space" field. Here are three ways that AI and augmented reality (AR)/virtual reality (VR) are already being used in the treatment of mental health.
The use of trackers is not novel. Fitness trackers have been around for some time, and have proven to be a great tool in aiding physical activity by tracking heart rate, steps, distance, sleep patterns, etc. Something similar is now being developed to track behaviors. Mental health struggles, such as anxiety or depression, have associated behaviors. Anxiety, for example, has been associated with behaviors such as picking, pacing, nail biting, hair pulling, or tapping. Furthermore, these tend to be automatic behaviors; things we likely engage in for some time before becoming aware of them. Research by Khan et al. (2021) demonstrated how motion sensors can be used to detect anxious behaviors, and Mastrothanasis et al. (2023) discussed how using wearables or sensors can help teachers and educators detect and manage students exhibiting anxious behaviors. Personal use of such trackers and motion sensors also has applicability, allowing individuals to become aware of these automatic behaviors sooner and applying appropriate coping skills.
These types of sensors are also beginning to be used in formal clinical settings. A major drawback of the increase of telehealth and tele-therapy has been the disadvantage that clinicians have with respect to body language and nonverbal cues. Through a screen, a clinician is limited in seeing nonverbal cues and anxious behaviors. AI is now being used to detect these nonverbal cues through video recording. Therapists can be given real-time or “after session” data to help assist their clients. Several students from the Samsung Electronics’ Samsung Innovation Campus in Valencia, Spain came together to develop CoteraplA, an AI program that can help therapists by “giving them a second set of eyes.” The students state the program gives mental health professionals a helping hand by recording patients’ verbal and non-verbal expressions via video, then providing actionable information in an easy-to-digest form.
In graduate studies, clinicians are taught numerous effective techniques to employ with clients. For example, theoretical frameworks ranging from client-centered humanistic styles, motivational interviewing techniques, cognitive-behavioral interventions, and dialectical-behavioral approaches are all widely used. Supervision and “human-rating” constructs have traditionally been used to assess the efficacy of clinician ability. Recent advances in linguistic AI programs have allowed more accurate scoring of psychotherapy tactics. Flemotomos et al. (2021) discussed creating a BERT (bidirectional encoder representations from transformers)-based model of automatic behavioral scoring for cognitive-behavioral therapy. This type of automatic behavioral coding and transcription of sessions can be extremely useful in not only training new clinicians, but also maintaining the skills of seasoned clinicians. The goal is not only to train more effective therapists, but continue real-time assessment of skills, areas of improvement, and efficacy across the scope of therapy. The following are several BERT-based models already being used that have strong applicability for mental health:
BERTweet: BERT model based on analyzing twitter data to screen mental health discussions on social media
BioBERT: BERT model trained on biomedical contexts including clinical notes
HealthBERT: BERT model based on health related text such as articles or mental health content on the internet
ClinicalBERT: BERT model that is trained on clinical text, electronic medical records, clinical notes, or mental health diagnosis or treatment
A subset of AI, virtual reality (VR) or augmented reality (AR), is perhaps the most widely discussed technology in recent times with respect to mental health treatment. The idea is to create an immersive world where the individual is able to work through immediate concerns. The bulk of current research regarding VR has been in the treatment of phobias or utilizing exposure therapy. Other areas of focus have been in the treatment of schizophrenia, social anxiety, eating disorders, and addiction. According to Bell et al. (2020), VR has shown to elicit similar physiological and psychological reactions to real-world environments. Further, they state that “superior capabilities for experimental manipulation and controlled exposure could significantly advance the field of mental health by improving methodological rigor as well as enabling more accurate and individualized assessment.”
The advances in AI and VR are certainly exciting, but come with their own questions and concerns. For instance, discussions about clinician access, training, and willingness to utilize AI technologies in sensitive personal matters such as mental health treatment are valid. Considerations of privacy, ethics, and transparency are warranted. In addition, the cost of utilizing such technologies comes with a steep price tag. Several VR software programs that are currently available to clinicians are selling for thousands of dollars, which does not include clinician training. Either way, it will be interesting to see how the field of mental health adapts to the inevitable advancements in AI and how clinicians and technology can work together to improve mental health treatment.
References
Bell, I.H., Nicholas, J., Alvarez-Jimenez, M., Thompson, A., Valmaggia, L. (2020) Virtual reality as a clinical tool in mental health research and practice. Dialogues Clin Neurosci 22(2): 169-177. https://doi.org/10.31887%2FDCNS.2020.22.2%2Flvalmaggia
Flemotomos N, Martinez VR, Chen Z, Creed TA, Atkins DC, Narayanan S (2021) Automated quality assessment of cognitive behavioral therapy sessions through highly contextualized language representations. PLoS ONE 16(10): e0258639. https://doi.org/10.1371/journal.pone.0258639
Khan N.S., Ghani M.S., and Anjum G., (2021). ADAM-sense: Anxiety-displaying activities recognition by motion sensors, Pervasive and Mobile Computing, Volume 78101485, ISSN 1574-1192
Mastrothanasis, K., Zervoudakis, K., Kladaki, M., and Tsafarakis, S. (2023). A bio-inspired computational classifier system for the evaluation of children’s theatrical anxiety at school. Educ Inf Technol 28, 11027–11050 (2023). https://doi.org/10.1007/s10639-023-11645-4
login
who are we?
contact us
qosheapp
Toi Staff
Gideon Levy
Belen Fernandez
Tarik Cyril Amar
Rami G Khouri
Dr Ramzy Baroud
Mort Laitner
Bradley Blankenship
visit website
