Sometimes, a person’s biggest worry can simply be whether or not they’re getting a good night’s sleep. So it makes sense that sleep trackers have become such a popular and well-received technology.
The basic premise of a sleep tracker is straightforward. The device typically offers users two main pieces of information: how long they’ve slept and what the quality of their sleep was. These conclusions are reached by collecting various kinds of physical data during sleep, and through this information users can adjust their sleep schedules or sleep habits accordingly, leading to a healthier sleeping experience.
But tracking sleep is notoriously difficult; after all, most trackers are worn around your wrists or midsection, and not the brain–which is where all sleep activity occurs.
When it comes to tracking sleep, scientific sleep studies are considered the most accurate method. Participants in these studies are attached to wires that monitor brain waves, rapid eye movements, oxygen levels, breathing patterns, heart rate and more. One study can produce over 1,000 pages of data, and it’s no surprise that the direct analysis of brain activity by professionals is the definitive way to monitor sleep. But it is expensive and a hassle, while wearable or bedside sleep devices offer a convenient and affordable way to track sleep.
Since sleep trackers cannot monitor brain waves, improvements are focused on other data that is possible for the devices to record. The best sleep trackers are able to record much of the same data that sleep studies do, primarily tracking heart rate and breathing patterns to determine sleep behavior, and some sleep trackers can also provide environmental data including light levels, sound and temperature. Even further, trackers can prompt users to answer lifestyle questions that might affect sleep such as diet, stress, alcohol consumption and so on. Evidently, there are a multitude of ways that trackers can record sleep data, but the question then remains: If sleep studies are the most accurate form of sleep measurement, how close do sleep trackers come through their compensating methods?
The reality is that until sleep trackers are able to monitor brain waves, they will remain, as experts say, more of a guesstimate than an accurate measurement of sleep. However, this doesn’t mean that sleep trackers are useless. Sleep trackers still provide plenty of insight based on the other data it collects.
The main reason for performing a sleep study is usually to diagnosis sleep disorders, such as sleep apnea. If a person is worried that they might have a serious sleep disorder, this is likely the best route to go. But for the rest who are just looking for ways to improve their sleep, the information and guidance offered by sleep trackers is more than enough to do so. In addition, the act of tracking increases mindfulness, which encourages people to make better sleep choices.
As technology evolves, it becomes able to perform more complex tasks with less effort. It’s not far-fetched to say that with time, sleep trackers could realistically become as accurate as sleep studies, giving consumers full control over their sleep.
Fitness trackers are everywhere. From Fitbits to Mi Bands to Apple Watches, tech companies are jumping onto the bandwagon, for better or worse. The results (and consequences) have been wide-ranging: on the one hand, fitness trackers do seem to spur people to exercise more. On the other hand, fitness trackers have also revealed the locations of classified bases when military personnel unknowingly shared the details of their workouts with other users.
But do fitness trackers really work in the first place? Aside from their role as an incentive for exercise, can fitness trackers live up to the claims of their manufacturers?
The answers are mixed. While fitness trackers can measure certain indicators very accurately, they don’t always do so well with others. For instance, a Stanford study found that although fitness trackers could measure heart rate very well (the error rate was less than five percent), this was not the case for other factors, such as calorie burn. Researchers found that when it came to energy expenditure, the least accurate device was off by an average of 93 percent. Even the best device had an error rate of around 27 percent.
Granted, the Stanford study was not the end-all of fitness trackers, as the sample size was fairly small (60 volunteers used the seven most popular trackers). The research team also found that even small factors, like skin and BMI, affected accuracy, raising a troubling question: are manufacturers making claims that don’t measure up?
The answer might be less straightforward than you may think. An earlier study conducted by Japan’s National Institute of Health and Nutrition had 19 subjects testing 12 trackers–at the same time. Over the course of 24 hours, scientists found that while participants burned an average of 2,093 calories, fitness trackers deviated from the norm significantly, by as much as +/- 200 kilocalories. When volunteers tested the devices over a period of 15 days, researchers still found that the trackers reported consistently lower outputs (up to 800 kilocalories) than the actual calorie burn as measured by the team.
Even so, the Japanese group concluded that for the average person, fitness trackers were quite useful. Unlike in a research setting, where accuracy is paramount, the average user simply needs relative results: so long as someone knows whether they’re burning more or less calories than yesterday or last week, it’s not a huge issue. Besides, a good fitness tracker could be what they need to help themselves build consistent, lasting habits.
Sleep, on the other hand, is a whole different story. Though trackers can distinguish when someone is sleeping or awake thanks to their accelerometers, they don’t really have any good way of determining whether you are in deep REM sleep or simply dozing lightly.
In a 2011 study, researchers compared Fitbit data to a polysomnography test, the gold standard of evaluating sleep quality. While Fitbits tended to overestimate the sleep time of adults by 43 minutes, they often underestimated the sleep time of children by 109 minutes. More importantly, as one scientist pointed out, fitness trackers have no way to distinguish between the stages of sleep, given that such activity occurs solely in the brain. Such brain waves are measured by electroencephalography (EEG) machines–not heart-rate monitors.
As in fitness, trackers can help people pay attention to the quality and amount of their sleep; at the same time, given that some sleep disorders, like sleep apnea, are life-threatening, false claims on a sleep tracker could have some serious ramifications. For instance, a patient could forego seeing a sleep specialist simply based on positive (and inaccurate) data from a tracker, allowing their condition to worsen over time.
In the end, fitness trackers are something of a mixed bag. While they can track the more obvious indicators of health (such as heart rate) very clearly, they are less adept at measuring more subtle (but equally important) factors, primarily sleep quality and calorie burn. It’s certainly possible that manufacturers will be able to refine their devices as time goes on, but as it stands right now, it’s best to take the claims of fitness trackers with a grain of salt.
The quest to crack the machine learning code has been a dream of scientists since the invention of the computer. In 1950 the Turing Test captivated the public’s imagination with a question that likely seemed more based in science fiction than reality: could a computer ever match human intelligence?
We’ve come a long way since the technology of the 1950s, but the goal for machine learning is essentially the same. And we are living into the age when machine learning is bringing the kind of solutions to life that have previously been relegated to sci-fi.
Computers that can learn and adapt autonomously have fast become the new gold standard. And the startups leading the way into our AI future are answering questions even more interesting than whether androids dream of electric sheep. One of the most exciting industries ripe for an AI makeover is healthcare. Machine learning has the power to revolutionize how we detect and treat illness, and also how we approach patient care. Here are 5 machine learning startups in the healthcare sector to keep on your radar:
1. ID Avatars
ID Avatars recently raised $1M in funding to create emotional intelligence aimed at improving patient care for people with chronic diseases. By using avatars to interact with patients directly, this startup hopes to lead the charge in creating technology capable of providing empathy. This would have a huge ripple effect across the healthcare industry and would be as useful to hospitals and pharmaceutical companies as to patients.
The mission of Arterys is to make clinical care data driven with a machine-learning based imaging platform. They have developed the first 4D Flow Technology to measure blood flow more accurately and in a non-invasive way than current methods. The blood flow work can be done with any MRI. The tech is integrated with a SaaS platform for doctors to better interpret data on the go, and the potential applications for non-invasive blood analytics span across all kinds of medical needs.
London-based startup Babylon has raised more than $30M, and it’s easy to see why because their idea is so simple but brilliant: remote app-based medical care. In other words, Babylon is your virtual doctor, nurse and pharmacy in one. Imagine how much more streamlined emergency and on-demand care could be through app-based consultations. It’s already the highest rated service in UK healthcare and aims to expand to areas where healthcare is even more essential, like rural Africa, where doctors are scarce but cell phones are becoming commonplace.
Utilizing machine learning to create custom diet plans, Nuritas will be a game-changer in preventative health and overall wellness. Based on AI plus DNA analysis, Nuritas is designed to identify the healthiest ingredients specific to your dietary needs. The future will be full of bioactive peptides.
Developed by MIT scientists, Ginger.io uses predictive models to create a mental healthcare platform. This is a useful alternative model to traditional mental health care, which can be prohibitively expensive and frankly not that convenient for the modern world. App users can arrange a video call with a therapist, text with a health coach, learn coping strategies, and analyze their mood over time with the help of embedded sensors. This is mental health care for the 21st century.
2017 Agreement Between Players and NBA Says Practice Games Only
From Fitbit Trackers for the average consumer to the Zephyr Bioharness, which is allowed by Major League Baseball for players during actual games, wearable sports technology is hitting fields, tracks and gyms right and left. But not the basketball courts of the NBA, according to a recent decision between the organization and the Players Association.
The collective bargaining agreement released earlier this year states explicitly (and more than 250 pages into the document) that, “No Team may request a player to use any Wearable unless such device is one of the devices currently in use as set forth in Section 13(f)below or the device and the Team’s cybersecurity standards have been approved by the Committee.”
Section F makes it clear that players will only wear the device on a voluntary basis. The agreement further states that devices can be worn during practice— but not games.
Wearable devices for professional athletes measure everything from movement information (such as distance, velocity, acceleration, deceleration, jumps, and changes of direction), to biometric information (heart rate, heart rate variability, skin temperature, blood oxygen, hydration). Depending on the device, other health, fitness and performance information is gathered.
The technology is not, by today’s standard, new. The Adidas miCoach Elite Team System (one of approved devices for practice) hit the market in 2013. It’s touted, by Adidas, as the first of its kind “that uses physiological data in real time sending it straight to a coach’s tablet on the sideline. The system not only provides real-time insights during training, but tracks total training impact, collects and manages data and is highly portable.”
The goal of the system was to “offer insights into player performance and work rate, helping teams achieve and maintain peak physical performance.”
Apparently unconvinced about the need for wearable devices—but well aware that there’s no turning back from the wearable tech trend—the 2017 agreement sets up a committee to continue to explore the issue.
Wearable technology is far from limited to watches and glasses. Though smart clothing is admittedly taking longer to catch on, the possibilities are really endless when apparel is ascribed that “smart” quality that makes IoT products unique.
What if, for example, your pants were embedded with sensors that could tell you about your body, movements, and need throughout the day? Smart trousers are far from just a pipe dream. With various different visions in mind, several technology/apparel companies are developing pants that do more than just fit.
Sweetflexx, for example, sells leggings with “resistance band technology” to help wearers burn up to 255 extra calories a day, when worn. Their unique fabric technology is designed with comfort in mind, with crushed jade stone infused to lower body temperature by 10 degrees, and harness everyday movement to challenge muscles and tone your body.
Another type of smart athletic pants, developed by Athos, measures your muscular effort and maps it on a smartphone app. The app can tell you whether or not you are reaching your maximum muscle potential, if you are favoring one side of your body of another, or if some muscles are working harder than others.
There are also “smart tights” available for yoga enthusiasts. Sydney-based Nadi X comes with an app, and areas of the tights vibrate where posture and form need to be adjusted.
But as we’re discovering more and more, wearables have application outside of just sports. As one example, wearable tech trousers for tradesmen have been designed to keep workers safe. Developed by Snickers — the workwear company, not the candy — the pants house a device that collects data to alert wearers about knee protection and loud noise levels. The idea is to improve the health and safety of the employees wearing them, who may not realize when their health or safety is threatened. The data collection element of these wearables can help employers make adjustments that ensure safer, healthier conditions for workers.
Does wearing smart pants, necessarily, make you a smarty-pants? Maybe! The whole point of wearable technology is to add value and function, and it follows that the more value an item of clothing has, the smarter an investment it is. Though the examples listed probably aren’t for everybody, they do a great job of demonstrating that wearable technology has potential beyond wristwear.