Studies explore impact of wearable devices in health tech

In this article we explore three research studies involving wearable devices, and their roles in supporting patients, research, population health initiatives and the potential clinical benefits.

The studies focus on the use of a wearable device for understanding brain function in a more naturalistic setting, population health learnings from diagnostics and screening, and remote digital biomarkers and heart rate variability.

Increasing flexibility

A study published in Nature Neuroscience describes how researchers developed a wearable bidirectional closed-loop neuromodulation system called Neuro-stack and used it to record single-neuron and local field potential activity behaviour in humans. The study notes that Neuro-stack “provides an opportunity to investigate the neurophysiological basis of disease, develop improved responsive neuromodulation therapies, explore brain function during naturalistic behaviours in humans and, consequently, bridge decades of neuroscientific findings across species.”

In humans who have macro-electrodes and micro-electrodes implanted for clinical purposes, the Neuro-stack can simultaneously record up to 128-channel local field potential activity during ambulatory behaviours in humans. In addition, it can deliver customisable closed-loop multi-channel stimulation, with configurable parameters such as pulse shape, frequency and amplitude.

In terms of functionality, the Neuro-stack can be controlled and powered externally via a USB cable, or remotely controlled through a secure local network using a battery-powered configuration. This flexibility, the study explains, allows researchers to perform recording and stimulation during stationary or ambulatory behavioural tasks.

The researchers say that the Neuro-stack differs from existing models which traditionally require the patient to be immobilised during the studies, as it is miniaturised and allows the patient to roam freely. They note that this is advantageous from a investigatory perspective, as it allows the technology to gather data from the patient whilst in a more naturalistic state – thus providing more accurate and true-t0-life results.

“Understanding brain function and its relation to cognition and behaviour requires the integration of multiple levels of inquiry,” the journal explains, and “although single-neuron studies in humans have yielded unique insights into memory, perception, decision-making as well as pathologies such as Parkinson’s disease and epilepsy, they have been solely done in immobile participants.”

Understandably, there are still large gaps in findings from human neuroscience studies, and so according to the study Neuro-stack presents “an unprecedented opportunity in the fields of neuroscience”, as it offers “a potential technological pathway toward more advanced implantable technologies with the development of a miniaturized bidirectional neuromodulation external device.”

It is believed that recordings of single-neuron activity in patients with brain disorders, particularly under naturalistic settings, would provide a “unique window into the neural mechanisms of brain pathology, symptoms and treatment response and would lead to more personalised and effective therapies.”

Citiation: Topalovic, U., Barclay, S., Ling, C. et al. A wearable platform for closed-loop stimulation and recording of single-neuron and local field potential activity in freely moving humans. Nat Neurosci 26, 517–527 (2023).

Wearables and machine learning

In an article published in The Lancet Digital Health, researchers state: “The increasing accessibility of wearable activity-tracking and health-tracking devices has prompted much research into passive diagnostics and screening that could contribute to infrastructure for population health testing and ultimately mitigate potential pandemics.”

In the past three years, they note, studies have shown the potential for wearable devices to contribute to population-level tracking of disease prevalence. However, the researchers urge caution around the design of these to ensure that estimated outcomes are accurate. The methods of evaluation proposed in COVID-19 detection studies using machine learning, according to the article, “do not replicate a realistic clinical use scenario.”

The article highlights how an investigation, carried out by Evidation Health, used previously published data to train a model using 48 features from wearable devices to identify COVID-19 on a daily basis. Data was trained on a random draw of 35 percent of the participants, and the model was retrained weekly. The researchers found that the randomly drawn test set “typically outperforms the prospective test set”, noting: “The performance of the model is clearly overestimated when using the retrospective test set, as the model implicitly learns the prevalence of the disease.”

The article also points out that it can be different for wearables designed to track disease to differentiate between different types of disease, as they are best equipped to identify symptoms rather than source; this does not negate their worth, but the researchers encourage clarification around purpose. “Previous work has shown that a model trained to detect the onset of influenza using wearables can transfer to COVID-19 detection without fine-tuning,” it notes. “This finding implies that the manifestations of influenza and COVID-19 overlap in terms of the distribution of data from wearable devices. Models that claim to predict COVID-19 symptoms might be better classed as models that can predict respiratory virus illness.”

Citation: Nestor, B. et al. (2023) ‘Machine learning covid-19 detection from wearables’, The Lancet Digital Health, 5(4). doi:10.1016/s2589-7500(23)00045-6.

Remote digital biomarkers and heart rate variability

A study published in Frontiers, in which the researchers highlight how heart rate variability (HRV) can offer insights into humoral, neural and nuerovisceral processes , but “has yet to be fully potentiated in the digital age.”

The researchers continue: “Remote measurement technologies (RMTs), such as smartphones, wearable sensors or home-based devices, can passively capture HRV as a nested parameter of neurovisceral integration and health during everyday life, providing insights across different contexts, such as activities of daily living, therapeutic interventions and behavioyral tasks, to compliment ongoing clinical care.” They add that many RMTs can be deployed as wearable sensors or digital cameras using photoplethysmography – an optical technique used to detect volumetric changes in blood in peripheral circulation.

When RMTs measure HRV, they provide the opportunity to identify digital biomarkers that can indicate changes in a person’s health, or disease stats in disorders where neurovisceral processes become compromised; for example, depression, epilepsy, substance abuse or dissociative disorders. The researchers note that other studies are using RMTs to “actively and passively measure neurophysiological, motor, functional, cognitive and affective digital biomarkers remotely in disorders” and as such RMT-based HRV could provide additional insight and context, by providing an easily deployable and scalable metric of health domains. .

Therefore, the study continues, these wearables have potential as “adjunct digital biomarkers” that provide “continuously updated, objective and relevant data to existing clinical methodologies, aiding the evolution of current ‘diagnose and treat’ care models to a more proactive and holistic approach that pairs established markers with advances in remote digital technology.”

Ultimately, the researchers say, the use of RMTs to capture this inform “can provide more detailed data across different contexts, such as the activities of daily living or interventions and behavioural tasks”.

Citation: Owens AP (2020) The Role of Heart Rate Variability in the Future of Remote Digital Biomarkers. Front. Neurosci. 14:582145. doi: 10.3389/fnins.2020.582145