Educational content on VJNeurology is intended for healthcare professionals only. By visiting this website and accessing this information you confirm that you are a healthcare professional.

The Sleep Disorders Channel is supported through educational grants from Alkermes and Takeda.

VJNeurology is an independent medical education platform. Channel supporters have no influence over the production of content.

Share this video  

Sleep Europe 2026 | SLEEPYLAND: an open-source AI toolbox for building trust in automatic sleep staging

Luigi Fiorillo, PhD, Neurocenter of Southern Switzerland (EOC), Lugano, Switzerland, discusses the rationale behind SLEEPYLAND, developed to answer which automatic sleep staging model to use on a given dataset. Dr Fiorillo explains that it merges different models trained on the same data into an ensemble, much like relying on several scorers, to build trust in how models perform on specific datasets. He also highlights the aim of releasing these tools openly so people can test them on their own data, and of outputting hypnodensity graphs, which show sleep stage probabilities rather than discrete stages. This interview took place at Sleep Europe 2026 in Maastricht, the Netherlands.

These works are owned by Magdalen Medical Publishing (MMP) and are protected by copyright laws and treaties around the world. All rights are reserved.

Transcript

SLEEPYLAND, it’s about answering this question. So the main question will be, how should I use like an automatic sleep staging model on my data? Which one should I use? So to answer this question, that’s the reason why we decided to develop SLEEPYLAND. So to try to expose a solution that will merge different models altogether to give you, let’s say, an assembled answer like you would do with different scores...

SLEEPYLAND, it’s about answering this question. So the main question will be, how should I use like an automatic sleep staging model on my data? Which one should I use? So to answer this question, that’s the reason why we decided to develop SLEEPYLAND. So to try to expose a solution that will merge different models altogether to give you, let’s say, an assembled answer like you would do with different scores. And let the scorers try to rely on not a single model, but a merge of different models that actually are hosted under the same cap, let’s say. And mainly models that should have been trained on the same set of data. So this is about trust. It’s about, you know, understanding how this model will perform on a specific dataset. And the point is to also try to release all these tools that we are building open to everybody so that people can start to use them, can start to try to understand how they perform on their data. And mainly, I mean, I would say the most important thing of these models and those of SLEEPYLAND, it’s not about SLEEPYLAND, again, it’s about models built on the same knowledge and the same set of data is to also output what we call the hypnodensity graph. So a representation, so probabilities of sleep and not just the discrete sleep that we are used to. So to give you also an idea of not the uncertainty but the additional information that you would have out of an answer that a scorer or a group of scorers, a group of models will give to you.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.

Read more...

Choose your cookie settings

VJNeurology uses necessary cookies to make the site work. With your permission we'd also like to use analytics, preference and marketing cookies to understand how the site is used and improve it.

Choose an option to continue. You can change it at any time on our Cookie Policy page, linked at the bottom of every page.