GI-SleepNet: A Highly Versatile Image-Based Sleep Classification Using a Deep Learning Algorithm.

Tianxiang Gao,Masashi Yanagisawa,Yuji Watanabe,Masahiro Ohsawa,Akihiro Yamanaka,Chijung Hung,Kazuhiko Kume,Kazumasa Horie,Jiayi Li

doi:10.3390/clockssleep3040041

Abstract

Sleep-stage classification is essential for sleep research. Various automatic judgment programs, including deep learning algorithms using artificial intelligence (AI), have been developed, but have limitations with regard to data format compatibility, human interpretability, cost, and technical requirements. We developed a novel program called GI-SleepNet, generative adversarial network (GAN)-assisted image-based sleep staging for mice that is accurate, versatile, compact, and easy to use. In this program, electroencephalogram and electromyography data are first visualized as images, and then classified into three stages (wake, NREM, and REM) by a supervised image learning algorithm. To increase its accuracy, we adopted GAN and artificially generated fake REM sleep data to equalize the number of stages. This resulted in improved accuracy, and as little as one mouse’s data yielded significant accuracy. Due to its image-based nature, the program is easy to apply to data of different formats, different species of animals, and even outside sleep research. Image data can be easily understood; thus, confirmation by experts is easily obtained, even when there are prediction anomalies. As deep learning in image processing is one of the leading fields in AI, numerous algorithms are also available.

Highlights

Sleep is a stable systemic state that, in mammals, is controlled by homeostasis and endogenous circadian rhythms
It is widely known that sleep is composed of two parts, non-rapid eye movement (NREM) sleep and REM sleep
The upper and lower parts of a single data image were the EMG raw data graph and heatmap of the EEG power spectrum, respectively (Figure 1A). The latter was calculated by fast Fourier transform (FFT) and normalized by Python’s Scikit-learn library

Summary

Introduction

Sleep is a stable systemic state that, in mammals, is controlled by homeostasis and endogenous circadian rhythms. It is widely known that sleep is composed of two parts, non-rapid eye movement (NREM) sleep and REM sleep It took researchers nearly three decades, from 1924 to 1953, to separate these two types of sleep. Sleep is characterized using an electroencephalogram (EEG), which was first developed by Hans Berger in 1924 [1] He called the EEG a “brain mirror,” reflecting the “electrical psychic energy” within cortical tissue. Loomis used the six-channel EEG of 30 s because the record sheets were automatically cut by a scissor every 30 s, and this marked the earliest conceptual origin of the classification epoch These separate epochs were visually judged by researchers in a manner similar to the workflows conducted by modern polysomnography (PSG) technicians. Even most sleep classification algorithms use 30 s as one epoch length to determine the sleep stage

Methods

Results

Conclusion