Menu
Circadian Sleep Sufficiency – a lifeline for night shift workers?
You may be waking up to your usual morning routine, but for millions of night shift workers (around 17% of adults worldwide – 2024 Statistics on America’s Off-Hour Workforce) – their body clocks may have no idea what time it is.
Night-shift workers are up to three times greater risk of occupational accidents – Shift Work Hazards, and crashes on their way home, with one study finding 37% near-crashes after 45 minutes at the wheel following a typical night shift – High risk of near-crash driving events following night-shift work.
The obvious answer is that night-shift workers just need longer in bed.
But this team working with nurses in Korea became curious when they spotted some dramatic differences between participants – even when they had spent the same time in bed, some suffered from crippling daytime sleepiness – and the other felt just fine. Personalized sleep-wake patterns aligned with circadian rhythm relieve daytime sleepiness.
So they decided to take a closer look and came up with a radical new model for optimising sleep efficiency and a new parameter, Circadian Sleep Sufficiency.
This model calculates the timing of sleep episodes to find the intersection between the circadian cycle, which responds to patterns of light and dark, and homeostatic sleep pressure – the natural feeling of fatigue that builds up over the course of the day.
As you sleep, that homeostatic drive falls away and you naturally move into the ‘wake zone’, where you can wake up without needing an alarm.
This team used actigraphy and mathematical models to create personalised sleep and wake times in real time, effectively optimising the effectiveness of time in bed while reducing sleepiness on the job – or on the way home.
Complementing these tailored profiles with dynamic lighting cycles that improve alertness during work periods and help the body clock to tell the time could help us all to get a refreshing nights’ rest – whenever that may be – Dynamic lighting schedules to facilitate circadian adaptation to shifted timing of sleep and wake.
🌟 Doing Well by Doing Good: Innovators in Healthcare
This week’s spotlight: Kim Crowe, Chief Executive of Parkhaven Trust
Kim and her team at The Beeches, a 45-bed dementia care facility in Liverpool, have implemented circadian lighting as part of an integrated approach to create bright, active days and calm nights with remarkable results:
Despite a 50% increase in lighting costs, the benefits far outweigh the investment. Kim was able to fund the cost of the initial installation with an innovation grant. But as Parkhaven Trust breaks ground on an ambitious new build programme, circadian lighting is firmly in the budget.
Kim’s advice:
Kim is one of a growing number of healthcare professionals who are harnessing smart, circadian technologies to create healthier, happier environments for patients and staff alike. Listen to Kim talking about her experience here – Innovators in Healthcare October 2024, Kim Crowe.
Machine learning and new insights into autism on National Change Your Password Day!
You, like me, might one of the 78% of people who use the same password on multiple platforms – leaving us wide open to hacking, theft and fraud – 78% of people use the same password across multiple accounts.
So I’m very relieved that mosst of the apps on my phone work with face recognition now. But I’ve often wondered how on earth they work.
After all, it’s often dark, the phone can be a log way from my face and not always at a flattering angle!
The simplest models build a map of critical points on your face – eyes, nose, mouth, jaw, and the relationships between them, create a basic equation and compare that equation with the one generated when you peer at your phone. The more sophisticated ones use infra-red, 3-d and basic motion capture to check it’s a living face, not a photo or a mask.
But the really cool stuff is happening in deep neural network research, where models like VGG-Face and FaceNet are as good as you and me at identifying faces.
It turns out that there are some striking parallels between the development of face recognition in these machine learning models and in children with autism, with huge potential for early diagnosis and support.
In people and in machines, as exposure to faces increases, the neural distances between them grow – the system gets more better at telling faces apart. But similar to the wiring and firing of the developing visual cortex in children if the information is poor or distorted in critical duty cycles called ‘critical periods’, the model will never ‘learn’ to differentiate those features correctly, however much additional information you feed in. This team uses face recognition models as a ‘digital twin’ to learn how to support the developing brain.
I wish they’d create a model to make it easier to remember passwords too! – A critical period for developing face recognition.