Source: Karolina Grabowska / Pexels
New research has found that wearable AI is promising to help detect and predict anxiety but is currently best combined with a professional clinical evaluation. Wearable devices can be useful to track, monitor, and share data with your doctor or therapist. Wearable AI that detects and predicts anxiety in combination with a professional clinician can be used to help personalize treatment of anxiety disorders.
Wearable AI is the combination of data obtained from wearables and machine learning algorithms that can analyze the data to help detect and predict anxiety.
There are four types of wearable devices:
- Noninvasive on-body wearables that are fixed directly on the body or skin (like smart rings, smart wristbands, or smartwatches).
- Near-body devices that are fixed close to the body, but no direct contact with the body or skin.
- In-body devices (implantable electronics).
- Electronic textiles.
Newer versions of wearable devices incorporate AI technology within the device or, more commonly, the data is sent to a computer, smartphone, or cloud for computing power and processing. Many wearable AI devices currently focus on anxiety detection and prediction, though some wearable devices seek to implement treatment of anxiety through vibration or neurofeedback (but these treatment devices were not the focus of this study).
The review study examined 21 studies and included 17 in a meta-analysis, which focused on noninvasive wearable on-body devices like smartwatches, smart glasses, smart wristbands, smart clothes, and smart rings. Wrist devices were the most common in the research studies. Across the studies, 20 different algorithms were used. All the commercial wearable devices studied had AI embedded in a separate device like a computer but used data collected from the wearable device. None of the devices use neuroimaging to predict anxiety.
Researchers concluded that, given the state of the current research, wearable AI for anxiety detection and prediction is a promising tool but is not yet ready to be used clinically for diagnosis without additional clinical assessment. One major problem is that physiologically anxiety can look like other medical issues. It is also difficult to distinguish from physiological data alone anxiety from other conditions like depression or stress.
Furthermore, none of the devices in the study could distinguish between different types of anxiety disorders: panic disorders, social anxiety, phobias, obsessive-compulsive disorder (OCD), and posttraumatic stress disorder (PTSD). Knowing the specific type of anxiety disorder is important, especially since different types of medication and psychotherapy modalities target specific anxiety disorders. Therefore, while wearable AI can provide helpful screening and data, working with a clinician with the information from these devices is most helpful in diagnosing and treating anxiety.
This technology is rapidly evolving so there will likely be innovative ways to improve models, including being able to detect specific types of anxiety disorder.
Here are promising applications for wearable AI devices for anxiety:
1. Screening for anxiety to increase awareness and education around anxiety treatment. People may use a wearable device for other reasons—like tracking sleep and fitness—but then be notified that they are dealing with signs that may be anxiety. This early awareness is a good prevention strategy to alert people before anxiety worsens. Many people may not realize that they are experiencing anxiety until it becomes so overwhelming that it affects their work and relationships, and often people do not realize that anxiety is treatable. The sooner one seeks help for anxiety, the better.
2. Early prediction and recommendation of interventions for anxiety and panic attacks. The majority of the reviewed research studies (86%) used AI to detect anxiety and only 14% looked at AI to try to predict anxiety. Predicting anxiety has interesting clinical applications because it creates a window of opportunity to prevent the downward spiral of anxiety.
When wearable AI devices notify users that anxiety may be imminent, the user or the AI could recommend individualized strategies like deep breathing, visualization, or deep relaxation exercises. There may be some challenges because some forms of anxiety like panic attacks are known to come “out of the blue” but AI models could potentially find ways to predict panic attacks early enough to nip them in the bud with tailored tools. The ability to predict anxiety could also be useful for certain types of therapies like dialectical behavioral therapy, which often seeks to identify and implement strategies early on, before situations become overwhelming.
3. Sharing data from wearable AI with clinicians for personalized treatment of anxiety. Anxiety can have varying patterns, ranging from generalized anxiety which can be constant, to nighttime anxiety which interferes with sleep. Tracking the timing of anxiety throughout the day can identify recurring patterns and triggers. Sharing these patterns with clinicians can help people create targeted and individualized treatment along with their doctors, therapists, or treaters. Different interventions can be prescribed based on such patterns.
4. Development of wearable AI devices that integrate treatment. The recent research review focused only on studies that looked at wearable AI devices for anxiety detection and prediction, not treatment. However, wearable AI devices that can predict, detect, and then guide and deliver treatment is the future. There are AI devices that could detect and then also offer real-time personalized anxiety treatment, tailored to the type of anxiety detected or predicted.
Wearable AI devices are an exciting frontier with the potential to augment the diagnosis and treatment of anxiety and make the process more efficient, personalized, and interactive, along with a professional clinician’s guidance.
To find a therapist, visit the Psychology Today Therapy Directory.
Marlynn Wei, MD, PLLC Copyright © 2023 All Rights Reserved.