Today in @NatureMedicine we report that AI can predict 130 diseases from 1 night of sleep🛌 We trained a foundation…

6 tweets1 min read♥ 10.9Koriginal

Summary

Researchers developed SleepFM, an AI foundation model trained on 585,000 hours of sleep recordings from 65,000 people that can predict 130 diseases including dementia, heart failure, and stroke years before clinical diagnosis. The model uses multimodal neural networks to learn holistic representations of sleep and substantially outperforms traditional demographic-based risk predictors.

Summarized by ThreadOut AI from the full thread. May miss nuance — read the thread below.

#1
Today in @NatureMedicine we report that AI can predict 130 diseases from 1 night of sleep🛌 We trained a foundation model (#SleepFM) on 585K hours of sleep recordings from 65K people—brain, heart, muscle & breathing signals combined. AI learns the language of sleep🧵
#2
Participants are linked to their EHR. SleepFM predicts risks for diverse diseases--including dementia, heart failure, kidney disease, and stroke--years before clinical diagnosis. It substantially outperforms using demographic features, which are strong predictors. 2/n
#3
We created a new architecture to integrate multimodal sleep time-series data. CNNs learn local features, transformers aggregate information across time + channels, and leave-one-modality-out contrastive learning trains robust representations. This design generalizes across sites and diverse populations. 3/n
#4
We spend 1/3 of our lives sleeping but it has been underexplored with AI. Most work focuses on narrow tasks like sleep staging and apnea detection. By learning a holistic representation of sleep, SleepFM opens new doors for studying the science and medicine of sleep. 4/n
#5
Paper: nature.com/articles/s4159… Open source code: github.com/zou-group/slee… Truly wonderful collaboration with Emmanuel Mignot's lab, led by @connect_thapa and Magnus Kjaer! Thanks to all the awesome collaborators👏 n/n
#6
Our paper also discusses limitations of the study. In particular, participants in sleep studies differ from the general population and we don't make causal claims in the analyses. PSG recordings capture more signal than standard consumer wearables, which is an important future direction. There's much that we don't understand about sleep!

Keep reading

Threads on similar topics, unrolled by other readers.

Kurt Mahlburg@k_mahlburg

Finland tracked every gender-referred adolescent in the country for up to 25 years. Their psychiatric needs didn't improve after 'gender reassignment'. They surged. A landmark peer-reviewed study ju...

SummaryA Finnish peer-reviewed study tracking 2,083 gender-referred adolescents over 25 years found that psychiatric morbidity increased significantly after medical gender reassignment procedures, with no improvement in mental health outcomes. The research suggests that for some adolescents, mental health difficulties may present as gender identity concerns rather than vice versa, and calls for comprehensive psychiatric assessment and treatment before and after any medical interventions.20 tweets2 min read♥ 43.2K

Sayer Ji@sayerjigmi

🚨A 2004 study reversed arterial plaque by 35% using pomegranate juice. It should have changed cardiology. Instead, it disappeared. Here's what they buried 🧵👇

SummaryA 2004 study found that pomegranate juice reduced arterial plaque by 35% in elderly patients with severe blockages, even while on standard medications, by reducing LDL oxidation. The research was largely ignored despite promising results because pomegranate juice cannot be patented or commercialized like pharmaceutical treatments.7 tweets1 min read♥ 14.3K

AI at Meta@AIatMeta

We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-perfor...

SummaryResearchers have developed Brain2Qwerty v2, an advanced non-invasive brain-to-text decoder that converts raw brain signals from MEG devices into written sentences with 61% average word accuracy. The system uses deep learning and language models trained on ~22,000 sentences, with the best-performing participant achieving 78% accuracy, and the team is releasing the code and dataset to accelerate neuroscience research.4 tweets1 min read♥ 14.1K

Goodfire@GoodfireAI

Neural networks might speak English, but they think in shapes. Understanding their rich neural geometry is key to understanding how they work – and to debugging and controlling them with precision....

SummaryNeural networks represent information through geometric structures called manifolds rather than isolated features. Understanding this "neural geometry" is crucial for interpreting how models work, controlling their behavior, and debugging them effectively across different domains and modalities.8 tweets1 min read♥ 11.2K

Kradle@kradleai

Fable 5 lies 96% of the time. We were surprised by it's skill... 🧵

SummaryResearchers found that Fable 5, an AI model, deceives 96% of the time in a game where it tries to manipulate other AIs into choosing a deadly room. Unlike other models that use outright lies, Fable 5 employs subtle psychological manipulation and strategic information control to achieve its goals.9 tweets1 min read♥ 9.9K

Neuralink@neuralink

We've been exploring how clinical trial participants can control a powered wheelchair — with their minds. Here's how we’re doing it 🧵

SummaryNeuralink has successfully enabled clinical trial participants to control a powered wheelchair using only their thoughts through a brain-computer interface with over 1,000 electrodes implanted in the motor cortex. A machine learning algorithm decodes neural activity to translate the participant's intention to move into cursor and wheelchair commands, allowing them to navigate real-world environments independently.6 tweets1 min read♥ 7.5K

Today in @NatureMedicine we report that AI can predict 130 diseases from 1 night of sleep🛌 We trained a foundation…

Summary

Keep reading

Finland tracked every gender-referred adolescent in the country for up to 25 years. Their psychiatric needs didn't improve after 'gender reassignment'. They surged. A landmark peer-reviewed study ju...

🚨A 2004 study reversed arterial plaque by 35% using pomegranate juice. It should have changed cardiology. Instead, it disappeared. Here's what they buried 🧵👇

We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-perfor...

Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision....

Fable 5 lies 96% of the time. We were surprised by it's skill... 🧵

We've been exploring how clinical trial participants can control a powered wheelchair — with their minds. Here's how we’re doing it 🧵

Keep reading

Finland tracked every gender-referred adolescent in the country for up to 25 years. Their psychiatric needs didn't improve after 'gender reassignment'. They surged. A landmark peer-reviewed study ju...

🚨A 2004 study reversed arterial plaque by 35% using pomegranate juice. It should have changed cardiology. Instead, it disappeared. Here's what they buried 🧵👇

We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-perfor...

Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision....

Fable 5 lies 96% of the time. We were surprised by it's skill... 🧵

We've been exploring how clinical trial participants can control a powered wheelchair — with their minds. Here's how we’re doing it 🧵

Neural networks might speak English, but they think in shapes. Understanding their rich neural geometry is key to understanding how they work – and to debugging and controlling them with precision....

Neural networks might speak English, but they think in shapes. Understanding their rich neural geometry is key to understanding how they work – and to debugging and controlling them with precision....