Search

EP-4740860-A1 - INFORMATION PROCESSING DEVICE, SLEEP/WAKEFULNESS STAGE DETERMINATION DEVICE, COMPUTER PROGRAM, INFORMATION PROCESSING METHOD, AND SLEEP/WAKEFULNESS STAGE DETERMINATION METHOD

EP4740860A1EP 4740860 A1EP4740860 A1EP 4740860A1EP-4740860-A1

Abstract

Included are an electroencephalogram processing unit that obtains an amplitude and a phase from training electroencephalogram data measured from a living body, which is a mammal, for each of a predetermined plurality of frequency bands, and performs complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands, and a machine learning unit that performs machine learning processing of a machine learning model that determines sleep/wakefulness stages using a result of the complexity analysis processing and label data of the sleep/wakefulness stages corresponding to the training electroencephalogram data.

Inventors

  • SAKURAI, TAKESHI
  • FURUTANI, NAOKI
  • SAITO, YUKI

Assignees

  • University of Tsukuba

Dates

Publication Date
20260513
Application Date
20240708

Claims (18)

  1. An information processing device comprising: an electroencephalogram processing unit configured to obtain an amplitude and a phase from training electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and perform complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands; and a machine learning unit configured to perform machine learning processing of a machine learning model configured to determine sleep/wakefulness stages using a result of the complexity analysis processing and label data of the sleep/wakefulness stages corresponding to the training electroencephalogram data.
  2. The information processing device according to claim 1, wherein the electroencephalogram processing unit further extracts signals in the plurality of frequency bands, separately, from the training electroencephalogram data, and performs complexity analysis processing comprising of entropy analysis on each of the signals extracted in each of the plurality of frequency bands, and the machine learning unit performs machine learning processing of the machine learning model by further using a result of the complexity analysis processing.
  3. The information processing device according to claim 1, wherein the entropy analysis on the phase calculates an entropy of a cosine of the phase by performing entropy analysis on the cosine of the phase.
  4. The information processing device according to claim 1, wherein the electroencephalogram processing unit uses expanded sample entropy as entropy analysis.
  5. The information processing device according to claim 1, wherein when generating learning data used for machine learning of the machine learning model, the machine learning unit expands the learning data by standardization or normalization using a plurality of thresholds, or by both standardization and normalization using a plurality of thresholds.
  6. The information processing device according to any one of claims 1 to 5, wherein the living body is a mouse.
  7. A sleep/wakefulness stages determination device comprising: an electroencephalogram processing unit configured to obtain an amplitude and a phase from electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and perform complexity analysis processing comprising of entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands; and a determination unit configured to determine sleep/wakefulness stages from a result of the complexity analysis processing using a determination model.
  8. The sleep/wakefulness stages determination device according to claim 7, wherein the electroencephalogram processing unit further extracts signals in the plurality of frequency bands, separately, from the electroencephalogram data, and performs complexity analysis processing comprising of entropy analysis on each of the signals extracted in each of the plurality of frequency bands, and the determination unit determines the sleep/wakefulness stages by further using a result of the complexity analysis processing.
  9. The sleep/wakefulness stages determination device according to claim 7, wherein the entropy analysis on the phase calculates an entropy of a cosine of the phase by performing entropy analysis on the cosine of the phase.
  10. The sleep/wakefulness stages determination device according to claim 7, wherein the electroencephalogram processing unit uses expanded sample entropy as entropy analysis.
  11. The sleep/wakefulness stages determination device according to claim 7, further comprising a rescoring short epoch (RSE) method execution unit configured to apply an RSE method to determination results of the sleep/wakefulness stages determined using the determination model to eliminate epochs having short stage durations in the determination results of the sleep/wakefulness stages by lowering priority of epochs determined to have stage durations shorter than a certain time, starting from an epoch having the shortest stage duration.
  12. The sleep/wakefulness stages determination device according to claim 7, further comprising a skipping unit configured to suppress display of an epoch with a sufficiently high degree of confidence in the determination results of the sleep/wakefulness stages determined using the determination model.
  13. The sleep/wakefulness stages determination device according to claim 7, further comprising an accuracy display control unit configured to display accuracy of determination for the determination results of the sleep/wakefulness stages determined using the determination model.
  14. The sleep/wakefulness stages determination device according to any one of claims 7 to 13, wherein the living body is a mouse.
  15. A computer program for causing a computer to execute: obtaining an amplitude and a phase from training electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and performing complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands; and performing machine learning processing of a machine learning model for determination of sleep/wakefulness stages using a result of the complexity analysis processing and label data of the sleep/wakefulness stages corresponding to the training electroencephalogram data.
  16. A computer program for causing a computer to execute: obtaining an amplitude and a phase from electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and performing complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands; and determining sleep/wakefulness stages from a result of the complexity analysis processing using a determination model.
  17. An information processing method executed by an information processing device, the information processing method comprising: obtaining an amplitude and a phase from training electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and performing complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands; and performing machine learning processing of a machine learning model for determination of sleep/wakefulness stages using a result of the complexity analysis processing and label data of the sleep/wakefulness stages corresponding to the training electroencephalogram data.
  18. A sleep/wakefulness stages determination method executed by a sleep/wakefulness stages determination device, the sleep/wakefulness stages determination method comprising: obtaining an amplitude and a phase from electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and performing complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands; and determining sleep/wakefulness stages from a result of the complexity analysis processing using a determination model.

Description

Technical Field The present disclosure relates to an information processing device, a sleep/wakefulness stages determination device, a computer program, an information processing method, and a sleep/wakefulness stages determination method. The present application claims priority based on Japanese Patent Application No. 2023-111974 filed in Japan on July 7, 2023, the contents of which are incorporated herein by reference. Background Art In the past, in sleep research, experiments using mice have been conducted to determine sleep/wakefulness stages (wakefulness, non-REM sleep, REM sleep). As an automated technique for determining sleep/wakefulness stages, a technique that automatically determines sleep/wakefulness stages from mouse electroencephalogram data or electromyogram data is known (see, for example, NPTL 1). Known automated sleep/wakefulness stages determination techniques determine sleep/wakefulness stages based on frequency spectra of electroencephalogram data or amplitudes of electromyogram data. NPTL 2 describes that a combination of decomposition entropy, which is entropy decomposed into frequency band, phase, and amplitude using sample entropy (SampEn) of an electroencephalogram, and deep learning, is useful for diagnosis of Alzheimer's disease. NPTL 3 describes that a combination of decomposition entropy using expanded sample entropy (expSampEn) of an electroencephalogram and deep learning is useful for neural decoding (such as predicting what a person thinks). The expanded sample entropy is calculated using an autocorrelation of a single point to all points in a time series. NPTL 4 describes that expanded sample entropy of body movements (trunk acceleration) during night is useful for diagnosing autism spectrum disorder. In NPTL 4, decomposition entropy and machine learning/deep learning are not used. Citation List Non-Patent Literature NPTL 1: Masato Yamabe et al., "MC-SleepNet: Large-scale Sleep Stage Scoring in Mice by Deep Neural Networks", SCIENTIFIC REPORTS, 9:15793, 2019NPTL 2: Naoki Furutani et al., "Decomposed Temporal Complexity Analysis of Neural Oscillations and Machine Learning Applied to Alzheimer's Disease Diagnosis", Front. Psychiatry, 11:531801, 03 September 2020NPTL 3: Naoki Furutani et al., "Neural Decoding of Multi-Modal Imagery Behavior Focusing on Temporal Complexity", Front. Psychiatry, 11:746, 30 July 2020NPTL 4: Naoki Furutani et al., "Complexity of Body Movements during Sleep in Children with Autism Spectrum Disorder", Entropy, 23(4):418, 31 March 2021NPTL 5: Dorde Miladinovic et al., "SPINDLE: End-to-end learning from EEG/EMG to extrapolate animal sleep scoring across experimental settings, labs and species", PLOS Computational Biology, 15(4):e1006968, 18 April 2019 Summary of Invention Technical Problem However, with the known automated sleep/wakefulness stages determination techniques described above, it has been difficult to improve accuracy of determination. NPTLs 2, 3, and 4 do not disclose anything regarding the determination of the sleep/wakefulness stages. The present disclosure has been made in view of such circumstances, and an object thereof is to provide a technique that contributes to improvement of accuracy of automated determination of sleep/wakefulness stages. Solution to Problem According to an aspect of the present disclosure, an information processing device includes an electroencephalogram processing unit configured to obtain an amplitude and a phase from training electroencephalogram data measured from a living body being a mammal for a plurality of frequency bands according to individual frequency bands, the plurality of frequency bands being predetermined, and perform complexity analysis processing comprising entropy analysis on each of the amplitude and the phase obtained in each of the plurality of frequency bands, and a machine learning unit configured to perform machine learning processing of a machine learning model configured to determine sleep/wakefulness stages using a result of the complexity analysis processing and label data of the sleep/wakefulness stages corresponding to the training electroencephalogram data. According to an aspect of the present disclosure, in the information processing device, the electroencephalogram processing unit further extracts signals in the plurality of frequency bands, separately, from the training electroencephalogram data, and performs complexity analysis processing comprising entropy analysis on each of the signals extracted in each of the plurality of frequency bands, and the machine learning unit performs machine learning processing of the machine learning model by further using a result of the complexity analysis processing. According to an aspect of the present disclosure, in the information processing device, the entropy analysis on the phase calculates an entropy of a cosine of the phase by performing entropy analysis on the cosine of the phase. According to an aspect of the present disclosure,