A newly published Google patent describes a system that monitors what happens after an initial cardiac arrhythmia event, looks for patterns in subsequent episodes and uses those patterns to predict when another event is likely to occur. The US patent application, numbered US 2026/0294263 A1, was published at the start of this month.

The watch would learn your own pattern

The basic idea is actually pretty simple. A wearable detects an initial cardiac arrhythmia event using its biometric sensors. The patent repeatedly uses atrial fibrillation as an example, although its claims cover cardiac arrhythmias more broadly.

That first event then triggers an observation period. During this window, the wearable continues collecting heart data and records additional arrhythmia events.

The system looks at their timing and frequency to see whether there is a pattern specific to that person. If there is, Google proposes using it to generate a prediction of when a future event may occur.

So this is not simply a general risk score telling someone they are more likely to develop AFib at some point. The patent describes learning from actual arrhythmia events and looking for temporal patterns that could help forecast another one. For example, if someone’s episodes repeatedly occur at certain times or tend to appear at fairly regular intervals, that information could potentially become part of the prediction.

This goes beyond today’s AFib alerts

Current wearables can already look for signs of AFib. Google’s own devices are part of that. I recently looked at how AFib detection works on Fitbit Air, where heart rhythm and movement data collected while the user is still can trigger an irregular rhythm notification.

Pixel Watch goes further by offering ECG hardware on compatible models. But these features are still fundamentally looking for something that is happening or has already happened.

Prediction flips that around. Rather than asking whether the wearer’s rhythm appears irregular now, the system starts asking what might happen next.

It also fits well with where Google has been taking its health platform lately. Health Guardian already uses existing wearable data to identify longer-term patterns linked to blood pressure and insulin resistance, rather than relying on dedicated sensors that directly measure either one.

AFib prediction would push that idea further. The sensors gather the raw signals, but more of the value comes from interpreting patterns across time.

But patents are not product roadmaps. Companies routinely patent technologies that never appear in a commercial device, or that arrive years later in a substantially different form.

There would also be a much higher bar for something that claims to predict a future cardiac event. Detecting signs of AFib is already a regulated health feature. Telling someone that another episode may be approaching would need convincing clinical validation before Google could put that sort of warning on millions of wrists.

Still, the direction is worth watching.

The post Google is working on predicting AFib before it happens appeared first on Gadgets & Wearables.

The study, published in BMC Medicine, analysed 1,290 people from the Northern Finland Birth Cohort. Participants were aged 33 to 35 and were asked to wear an Oura Ring for two weeks while also completing established questionnaires covering anxiety and depression symptoms.

The pattern shows up in several metrics

People reporting moderate-to-severe anxiety tended to spend less of the night in REM and deep sleep and more time in light sleep. Similar differences appeared among participants whose questionnaire results indicated depression and anxiety symptoms.

And it was not just sleep stages. Those groups also tended to have a higher heart rate during sleep, lower nighttime HRV and lower levels of physical activity during the day.

None of those findings on their own are particularly surprising. Poor sleep, reduced activity and changes in autonomic nervous system activity have all been associated with mental health before. What makes this study more interesting is that all of these signals were picked up passively by an ordinary consumer smart ring while people went about their normal lives.

HRV is particularly worth watching here. We have written before about how wearables use HRV to track recovery and stress, and lower readings can show that the body is dealing with some kind of additional load. The problem is that HRV cannot tell you what that load actually is.

Alcohol, illness, poor sleep, training, stress and plenty of other things can all move the number. So seeing low HRV on an Oura Ring obviously does not mean someone has depression or anxiety.

The study becomes more useful when those signals are looked at together. Lower HRV combined with higher sleeping heart rate, less deep and REM sleep and reduced daytime movement starts to create a broader physiological and behavioural pattern rather than relying on any single measurement.

This was not Oura’s latest hardware

The researchers actually used the second-generation Oura Ring. Which makes the results a little more interesting. The study was not relying on some new sensor or experimental mental health feature. It was using fairly basic metrics that smart rings and watches have already been collecting for years.

Participants also did not receive feedback from the ring during the study, so their behaviour was not being nudged by readiness scores or recommendations in the Oura app. The researchers were essentially looking at the raw pattern left behind by sleep, nighttime heart rate, HRV and movement.

There are still some big limitations. This was a cross-sectional study, so it shows an association at one point in time rather than demonstrating that these wearable changes predict worsening mental health. The participants were also all around the same age and came from the same Finnish birth cohort.

The post Oura Ring study finds sleep and HRV changes linked to depression and anxiety appeared first on Gadgets & Wearables.

Huawei’s Watch GT 7 product page says the feature requires between three and 14 days of continuous wear. It arrives through an OTA update and users need to join the study through the Huawei Research app.

How the diabetes risk study works

So how does this actually work?

The Watch GT 7 uses pulse-wave signals collected by its PPG sensor, so the same tech that monitors hear rate. It then combines this with information such as age, sex, height and weight. After enough data has been collected, the system produces a diabetes risk assessment.

Now Huawei is not claiming the watch measures glucose directly. There is no glucose value in mmol/L or mg/dL, and the company says the feature cannot be used to monitor diabetes treatment. But it could flag up an issue that you could investigate further.

The rollout appears to be research-led rather than a standard health feature available everywhere. In Malaysia, Huawei says the study is initiated by Universiti Malaya, while availability depends on region.

We already flagged the high blood sugar risk assessment when looking at the Watch GT 7 generation. The latest regional rollout gives a clearer picture of how it actually works.

Huawei has explored similar ideas before. Earlier this year, the company brought diabetes risk alerts to the Watch GT 6 Pro, again using smartwatch sensor data to identify patterns linked with elevated risk.

The Watch GT 7 takes that approach into Huawei’s latest mainstream watch range. Both the standard GT 7 and GT 7 Pro are included.

Huawei is also working on a blood sugar device

The more interesting context is what Huawei has said about its next step. Just a few days ago Huawei Consumer Business Group CEO confirmed that the company plans to release a blood sugar testing device. Apparently, it’s been work in progress for the past four years.

Huawei has not said what that device will look like or how it will measure blood sugar. It could be a watch, another wearable or a separate health product.

Huawei is clearly pushing further into blood sugar-related health tracking, both through existing smartwatch sensors and new dedicated hardware. Hopefully, we’ll see exciting developments here in the coming months.

The post Huawei Watch GT 7 can now assess diabetes risk in selected markets appeared first on Gadgets & Wearables.

The comment came from Huawei Consumer Business Group CEO He Gang during a recruitment event at Peking University. He said Huawei plans to release the device and that the technology behind it took four years to get to a point the company considered acceptable and ready for industrialisation.

That is about as close as Huawei has come to saying something new is actually on the way. The problem is that He Gang stopped there, so we still don’t know what the device looks like, how it works or whether it is even something you wear on your wrist.

Non-invasive glucose monitoring?

The obvious guess is that Huawei is working on some kind of non-invasive glucose monitoring. We already have finger-prick meters and invasive continuous glucose monitors, so it would be a bit strange to spend four years on research just to come up with another version of something that already exists.

Huawei still hasn’t actually said that, though. Gang didn’t explain how the technology works, whether it tracks glucose continuously or even whether the device sits on the wrist. But reading between the lines, non-invasive sensing looks like the most likely direction.

A wearable also feels like the obvious fit. Huawei has been pushing harder into health features for years, and the latest Watch D3 already uses a tiny inflatable airbag in the strap to take proper oscillometric blood pressure readings.

So Huawei clearly isn’t afraid of building unusual hardware around health tracking. Whether this blood sugar device ends up as a watch, another type of wearable or something separate is still unknown, but a conventional glucose meter would be the least interesting explanation.

There have been hints before. A report in September 2025 claimed Huawei was developing an advanced blood sugar monitoring wearable for 2026, although that came from a Weibo tipster rather than Huawei itself. At the time there was very little to back it up.

Our takeaway

This week’s comment carries considerably more weight than previous rumours because it comes directly from the head of Huawei’s consumer business. It also changes the wording from something being researched to something Huawei plans to release.

Separately, a few Huawei watches can already assess a user’s risk of elevated blood sugar, but that system doesn’t measure an actual glucose value. The recently launched Watch GT 7 series, for example, supports high blood sugar risk assessment in compatible markets. But again, it isn’t a glucose monitor.

So this is one to keep an eye on. There is still plenty of room between ‘ready for industrialisation’ and a product sitting on a shop shelf. Regulatory approval would also become a major part of the story if Huawei intends to market the device as something that genuinely measures blood glucose.

Sources: IT Home, Huawei Support, Weibo, Huawei Central

The post Huawei says a blood sugar testing device is coming after four years of research appeared first on Gadgets & Wearables.

The patent was published on September 17 and describes a miniaturized impedance plethysmography sensor (IPG sensor), built into a computing device. The drawings make the wearable angle pretty obvious too. One example is a smartwatch, while others show small finger-worn devices and even other types of electronics.

Your fingertip becomes part of the sensor

The basic idea is quite neat. Google places a set of miniature electrodes underneath the fingertip, with two used to send a tiny electrical excitation signal through the tissue and another two used to measure the resulting electrical response.

Blood flowing through the finger changes its electrical impedance. By tracking those changes, the system can produce a pulsatile waveform that reflects the blood pulse, and Google says that signal can then be used to determine blood pressure or a proxy for blood pressure. The patent’s process diagram boils the whole thing down to generating IPG data, deriving the pulsatile signal and then estimating blood pressure from it. 

And Google clearly has smartwatch use in mind. Figure 6A shows the system incorporated into a watch worn on the wrist, with the sensing area positioned on the watch body where a finger could be placed against it. Earbuds are another option. As well as smartphones.

That’s not the only idea

And this is not even the only recent Google patent dealing with the problem. A separate application published last month describes a small finger cuff containing an inflatable bladder, multiple pressure sensors and optical sensors. 

That system would inflate around the finger while an array of barometric sensors maps pressure around the digital artery. But as you can see from the image below, it’s not exactly a discreet smart ring.

Optical sensors simultaneously measure the pulse, and the device combines those signals to calculate blood pressure. The patent’s flow diagram specifically shows inflation, pressure mapping, optical measurements and the final blood pressure estimate as parts of the process. 

Where this fits with Google’s plans

It is worth noting that neither patent names a Pixel Watch or Fitbit product. Google has been busy around blood pressure lately, though. It ran a Fitbit Hypertension Lab study with Pixel Watch 3 owners last November, and Blood Pressure Trends is coming to the Pixel watches this fall as part of Health Guardian.

We also covered an earlier Fitbit patent that hinted at blood pressure checks from the wrist in April. Put the two finger filings next to that and Google is clearly working on blood pressure from several directions. Patents do not confirm a release though, so treat these as a look at what Google is exploring.

Source USPTO: September patent, August patent

The post Google patent puts a tiny blood pressure sensor in a smartwatch appeared first on Gadgets & Wearables.

Its new Dream Recall feature uses real-time EEG data to work out when you are in REM sleep, then tries to wake you at the right moment while the dream is still fresh. The feature is now in beta as part of Smart Wakeup for Muse Premium subscribers using a Muse S or Muse S Athena.

The basic idea is pretty simple. Muse thinks remembering dreams has a lot to do with when you wake up, so instead of just sounding an alarm at a fixed time, the headband watches your brain activity overnight and looks for a better moment to pull you out of sleep.

It isn’t actually reading your dreams

It is worth being clear about what the headband is actually doing. Muse isn’t reading your dream, working out what you are dreaming about or recording anything happening inside your head.

Instead, its EEG sensors monitor brain activity and identify sleep stages in real time. When the system detects REM, Dream Recall can time the wake-up around that period rather than waiting until the dream has already started fading from memory.

Of course, dreaming and REM are not the same thing. Dreams can happen outside REM as well, but vivid dreaming is strongly associated with it, giving Muse a practical window to work with.

This is where EEG comes in

Muse has an advantage here over most smartwatches and smart rings. Those devices usually estimate sleep stages from signals such as movement and heart rate, while Muse measures electrical brain activity directly from the forehead.

That was one of the stronger parts of the Muse S Athena when I tested it. Its sleep tracking was particularly good at separating awake time from light sleep, an area where wrist-based trackers can sometimes struggle.

Muse has increasingly been using that EEG data while you are still asleep rather than simply handing you another chart the following morning. Deep Sleep Boost targets slow-wave sleep with timed audio, while Smart Wakeup looks for a lighter stage of sleep before sounding the alarm.

Dream Recall takes the same idea and points it at REM. There is also an Auto-Snooze function that reduces the sound as you begin to wake, with the idea of keeping you closer to that half-awake state where returning to the dream may be easier.

Muse is experimenting with sound cues too. Users can rehearse a cue while awake and then have it played during sleep, potentially making the sound recognisable inside a dream or helping with recall afterwards.

Wearables are starting to look at dreams

The idea of wearables moving into dreams isn’t unique to Muse. Earlier this year we uncovered an Oura patent describing dream detection and dream recall, including prompts and tools for recording what users remember after waking.

The big difference is that Muse has now moved beyond the design stage and put its version into a beta. Its EEG hardware also makes this a fairly natural extension of what the device is already measuring overnight.

Whether it actually leaves you remembering noticeably more dreams is the part that will need testing. Dream recall is subjective by nature, and waking people deliberately during REM also raises the obvious question of whether remembering more dreams is worth interrupting sleep for.

Those ready to dive in can pick up the Muse S Athena for $474.99 from the choosemuse.com website. To save 15%, use code “gadgetsandwearables” at check-out.

The post Muse can now wake you during REM to help you remember dreams appeared first on Gadgets & Wearables.

Until now, BeamO could record those sounds, but that was basically where its job ended. You could listen to them yourself or send the recordings to a doctor, but the device and app did not actually tell you what they were hearing.

The stethoscope finally does more than record

That makes StethO Sense a pretty substantial software upgrade. BeamO already combines a thermometer, ECG and digital stethoscope in one handheld device, but the stethoscope was always the part that required someone else to make sense of the data.

We have been following BeamO for a while. It started rolling out in Europe in 2025, adding another unusual health device to a Withings line-up that already includes watches, scales and blood pressure monitors. The hardware was interesting from day one, but StethO Sense starts closing the gap between simply collecting medical-style data at home and getting something immediately useful back from it.

The new feature sits inside the Withings app. You choose whether you want to analyse a heart or lung recording, BeamO guides you through the correct positions on the body, and the analysis comes back within seconds. According to Withings, heart recordings use four standard positions, while lung recordings use six positions for children under 12 and up to eight positions from age 12 upwards.

Two AI systems are doing the listening

Withings is not building all of this analysis itself. StethO Sense combines technology from eMurmur for heart sounds and StethoMe for lung sounds.

eMurmur Heart AI looks for murmurs and sorts recordings into three broad outcomes. It can identify no murmur, an innocent murmur or an abnormal murmur, while also returning heart rate and a confidence score. Withings says the system has shown roughly 90% sensitivity and specificity for detecting pathological murmurs across clinical studies, and the underlying technology has FDA clearance for use with adults and children in the US.

The lung side works a little differently. StethoMe generates a Breathing Index from 0 to 100 using breathing sounds and the overall quality and pattern of breathing. That gives you something much easier to follow than an audio recording sitting in the app with no obvious indication of what you are supposed to be listening for.

There is an important regional distinction here. In the US, StethoMe’s Breathing Index is being offered as a general wellness feature and does not have FDA clearance, while the eMurmur heart analysis does. So although both analyses live inside the same StethO Sense feature, they do not carry the same regulatory status.

There is a subscription catch

StethO Sense will not simply appear as a free BeamO software update. It is part of Withings+, so using the analysis features will require a subscription once any free trial period ends.

Existing BeamO owners who bought the device before October 1 will get a free month of Withings+ when StethO Sense launches. New BeamO buyers will also get a one-month trial. BeamO itself remains priced at around $250 in the US, so anyone buying it specifically for the new AI analysis will need to factor the subscription into the longer-term cost.

The post IFA 2026: Withings BeamO is getting AI heart and lung analysis in October appeared first on Gadgets & Wearables.

It gives you a value, not just a score

According to Actxa’s launch announcement⁠, the one-minute AI Glucose Snapshot produces an estimated glucose range and an indicative value within that range. You can take a quick reading whenever you want or use Meal Snapshots before and after eating to see how the estimate changes.

This is not continuous monitoring, so the band is not quietly tracking glucose throughout the day like a CGM. Actxa also makes it clear that the feature is intended for general wellness and should not be used to diagnose, monitor or manage diabetes.

That is important to note. As we covered when the FDA warned about wearables claiming to measure glucose⁠, no smartwatch or smart ring had been authorised by the agency to measure blood glucose without piercing the skin.

Spark Band uses Actxa’s BGEM technology, which analyses optical PPG signals with an AI model. The company already offers a similar feature through its Core Smart Ring, but it has not published accuracy figures specifically for the new wrist-worn band.

If the Actxa name means nothing to you, that is understandable. The Singapore company is not widely known internationally, although it says it has supplied more than 1.9 million trackers for Singapore’s government-backed National Steps Challenge since 2015.

The rest looks more familiar

Away from glucose, this is a fairly conventional screenless health tracker. It records sleep stages, sleep score, overnight signals, activity, Intensity Minutes, VO2 max and Fitness Age, while its daily recovery reading compares sleep HRV against the wearer’s own baseline.

There is no display, but LEDs and vibrations provide target heart-rate-zone guidance during exercise. The same cues can also handle movement reminders and wake-up alarms without turning the band into another source of wrist notifications.

The tracker weighs around 22 grams and combines a zinc-alloy casing with a braided nylon strap. It carries an IP68 rating, stores up to seven days of data and takes up to 90 minutes to charge.

Actxa claims up to 14 days of battery life and does not require a subscription. That combination places Spark Band in an interesting position alongside the established options in our screenless fitness tracker comparison⁠, especially since none of those devices currently provides a glucose estimate.

Pricing and availability

Spark Band will cost $149 or SGD 199, with pre-orders opening on September 24. Actxa expects to begin fulfilment on November 18 across selected markets in North America, Europe, Asia-Pacific and the Middle East, although it has not named the individual countries.

There is enough here to make Spark Band interesting, but the glucose feature needs independent testing before anyone should put much faith in the number it produces. If Actxa can show that its estimates reliably follow real changes around meals, the band could offer something genuinely different from Fitbit Air, Garmin CIRQA and Whoop.

The post IFA 2026: This $149 screenless band gives you a glucose estimate appeared first on Gadgets & Wearables.

The patent (US 12,721,568 B2) is called “Sleep age determination from wearable-based physiological data”. The original application goes back to August 2023 and was published in early 2025, but Oura has now secured the patent itself.

Oura’s Sleep Age would look at the bigger picture

The basic idea is pretty easy to understand. Instead of simply giving you another score for last night, Oura would work out what your sleep looks like over a longer period and translate that into an age.

So if you are 40 but your sleep patterns look more like those expected from someone who is 32, your Sleep Age could come in below your chronological age. Go the other way, and you probably have something to work on.

Oura describes Sleep Age as a longer-term metric calculated using weeks or months of data. So don’t expect the metric to bounce around wildly because you stayed up late one Saturday night. At least if Oura implements it roughly as described in the patent.

Underneath it all is quite a lot of processing. The ring collects PPG data and can also use information such as heart rate, HRV, temperature, respiratory data and sleep measurements. One machine-learning model classifies the night’s data into wake, light, deep and REM sleep, while another model uses the resulting sleep features to calculate the Sleep Age.

Those features can include how much time you spend in each sleep stage, sleep fragmentation, the midpoint of your sleep and other patterns. Oura also describes comparing the data with baseline sleep characteristics associated with someone’s chronological age rather than looking at each measurement in isolation.

The app mock-up is the interesting bit

The patent drawings give us a much better idea of what Oura has in mind. One app screen shows a Sleep Age Metric of 29 alongside a chronological age of 26, so there is no ambiguity about how Oura imagines presenting this.

Below that is an explanation. In Oura’s example, the user’s deep sleep is lower than normal and this is contributing to the higher Sleep Age. The app then recommends trying cold showers before bed to increase deep sleep and improve the metric.

That explanatory part could make Sleep Age more useful than simply sticking another number on the Today screen. The patent describes providing insights and recommendations based on what is affecting the result, including recognising temporary changes when someone is unwell rather than treating every rise in Sleep Age as a long-term deterioration.

The Oura Ring 5 already puts plenty of emphasis on longer-term interpretation rather than just raw sensor readings, while Oura has also been expanding features such as Health Radar and other health insights to older rings. So if this metric sees the light of day, it will fit in quite nicely.

Oura also already has Cardiovascular Age, so translating another part of your health into an age isn’t exactly unfamiliar territory. Sleep Age would apply a similar idea to something Oura has focused on since the earliest versions of its ring.

Source: Google Patents US20250064387A1

The post Oura patents a Sleep Age metric for its smart rings appeared first on Gadgets & Wearables.

When we first covered Garmin’s HbA1c patent, the idea was pretty easy to understand. The watch would shine several different wavelengths of light into the wrist, read the PPG signals and use that to estimate glycated haemoglobin. So this would not be live glucose tracking like a CGM, but more of a longer-term view of where your glucose levels have been.

This goes beyond one glucose patent

There has been some movement since February. Garmin’s HbA1c patent application remains pending and received a non-final USPTO Office Action in March, meaning the examination process is now underway.

That is not an endorsement of the technology, and an Office Action can include objections or rejected claims. What is more interesting is what shows up when Garmin’s other optical-sensing patents are viewed alongside it.

A separate Garmin filing called Pressure compensation for wrist-based pulse spectrometry explicitly discusses HbA1c measurement. It describes compensating for the pressure a watch applies to the wrist, because compressing blood vessels and tissue can alter the optical signals used to calculate physiological measurements.

The filing goes into considerable detail, including pressure compensation across multiple optical wavelengths between roughly 600 and 1,000nm. It also describes calculating glycated haemoglobin from three PPG signals at different wavelengths, and two of its inventors, Paul MacDonald and Christopher Kulach, also appear on Garmin’s dedicated HbA1c patent.

That is a more interesting clue than another generic mention of glucose in a patent. Garmin appears to have been working not only on how to calculate HbA1c, but on one of the practical problems that could make those calculations unreliable on an actual wrist.

There is another piece of the puzzle. Garmin’s wrist-worn optical cardiac monitor patent describes sensor arrangements using three optical transmitters with different wavelengths, as well as configurations operating across several red and infrared wavelength bands.

The HbA1c patent also requires three separate wavelengths travelling along substantially the same optical path. That does not prove these filings describe parts of one future Garmin sensor, but the overlap is difficult to ignore.

And then there is Fenix 9

The timing makes Fenix 9 particularly interesting. Garmin’s HbA1c work claims priority back to June 2023, the pressure-compensation filing dates to October 2023, and the optical-monitor work goes back to 2022, so these projects have been underway long enough to influence a 2026 hardware generation.

We already know Garmin is preparing new high-end watches. Our Garmin Connect 5.27 teardown uncovered six coordinated outdoor-watch records that could belong to a future Fenix 9 family, although Garmin has not confirmed their identity.

There is also Garmin model A05170, an unannounced premium wrist-worn device that recently appeared at the FCC with LTE and satellite connectivity. Fenix 9 or Pro version of the same is one possible identity for that watch. But the filing keeps its external and internal photographs confidential, so it tells us nothing about the optical sensor on the back.

And that sensor is now the bit worth watching. If Fenix 9 arrives with a substantially redesigned optical module, particularly one with additional red and infrared emitters, the connection with Garmin’s spectroscopy work becomes much more interesting.

Why Fenix 9 is still a long shot

There is still a big gap between Garmin researching wrist-based HbA1c estimation and putting the feature into a shipping watch. So far, we have found no public evidence of clinical validation, regulatory work or anything else that ties the technology directly to Fenix 9.

Garmin’s wider interest in metabolic health is clearly growing. In May, Garmin Health published a Diabetes Insight Report looking at Garmin wearable data alongside CGMs, AI platforms and third-party diabetes programmes, but that is separate from Garmin measuring glucose or HbA1c itself.

Even if Fenix 9 introduces a new optical sensor, that alone would tell us very little. Garmin could redesign the sensor for better heart-rate accuracy, Pulse Ox, ECG performance, power efficiency or entirely different health metrics.

What would make the HbA1c connection more interesting is evidence that a new sensor uses the sort of multi-wavelength optical setup described in Garmin’s patents. Until we see something like that, linking Fenix 9 directly to glucose tracking would be mostly guesswork.

The post Garmin’s glucose tracking plans look more serious ahead of Fenix 9 appeared first on Gadgets & Wearables.