Reported by Money Today / Aug 28, 2023
On the 28th, it was announced that a sleep biosignal AI interpretation system developed by South Korean medical researchers has obtained approval from the U.S. Food and Drug Administration (FDA). As a sleep disorder diagnostic solution, it is the first of its kind in Asia.
The newly developed AI sleep biosignal interpretation system is named 'SOMNUM', and it was developed in collaboration with the team of Professor Jiho Choi from the Department of Otorhinolaryngology at Soonchunhyang University Bucheon Hospital (Professor Seokhoon Jung from the Department of Psychiatry at Seoul Asan Hospital, Professor Hyunjun Kim from the Department of Otorhinolaryngology at Ajou University Hospital, Professor Yongmin Kim from the Department of Otorhinolaryngology at Chungnam National University Hospital, and Professor Jaehoon Cho from the Department of Otorhinolaryngology at Konkuk University Hospital) along with the AI sleep tech company, HoneyNaps.
Sleep biosignals refer to various physiological signals monitored during sleep to assess sleep quality or diagnose sleep disorders, including brainwaves (EEG), eye movements, chin and leg electromyography (EMG), electrocardiography (ECG), respiratory airflow, oxygen saturation, posture, snoring, and more. Previously, it would take 24 hours for skilled personnel to manually interpret sleep biosignals because the 68-hour-long sleep biosignals are interpreted in 30-second intervals of biosignals.
To overcome these limitations, extensive research on AI-based interpretation systems has been actively conducted worldwide. However, due to the complex and heterogeneous nature of biological signals, it has been extremely challenging for AI interpretation systems to reach the level of human interpretation.
However, the newly developed "SOMNUM" is capable of analyzing sleep patterns with the same level of accuracy and speed as human interpretation. The analysis can be completed in approximately 5 minutes. Unlike previous AI sleep diagnostic systems that focused on image recognition, SOMNUM utilizes a multi-channel, time-series biosignal-based diagnostic algorithm. It leverages deep learning to analyze multi-channel and large-scale data in real-time.
Professor Jiho Choi mentioned that for the development of key technologies to enhance the performance of SOMNUM, HoneyNaps collaborated with renowned research teams from leading universities in the United States, as well as Professor Shinil Kang's team from Yonsei University and the nCOMS (National Center for Optically-Assisted Mechanical Systems) under the Ministry of Science and ICT.
Research utilizing SOMNUM has been prominently featured in recent international conferences and SCIE-level journals. At the World Sleep Congress held in 2019, a study titled "Validation of an Automated Sleep Stage Scoring Internet Algorithm: Neural Network Algorithm" was presented. In June of this year, the Sleep 2023 conference organized by the American Academy of Sleep Medicine featured a study titled "Robust Hybrid Algorithm for Scoring Adult Respiratory Events." Furthermore, "Validation Study on Automatic Sleep Stage Scoring using Deep Learning Algorithms" was published in the SCIE-level international journal "Medicina" in 2022.
Professor Jiho Choi anticipated that future advancements in AI interpretation technology for biological signals will significantly enhance the quality of sleep for people worldwide. He said, "I hope that the continuous improvement of AI interpretation technology for biological signals will not only contribute to diagnosing sleep disorders but also enable the detection and prediction of certain cardiovascular, neurological, and musculoskeletal conditions."
On another note, Professor Choi is an internationally recognized sleep medicine expert who became the first in Asia to obtain the certification of "Sleep Medicine Specialist" from the three major sleep societies in the world: the American Academy of Sleep Medicine, the European Sleep Research Society, and the World Sleep Society, in 2018. Additionally, he holds the credential of Registered Polysomnographic Technologist (RPSGT) in the United States. Recently, he has been actively dedicated to promoting sleep health for humanity through the development of technologies such as sleep biological signal AI interpretation systems, digital therapeutics for insomnia, and non-contact biosignal monitoring devices.
Comments