RU2754779C1 - Method for identifying depression based on eeg data - Google Patents

Abstract

FIELD: computing technology.

SUBSTANCE: invention relates to computing technology, namely, to identification of depression based on EEG data. A method is proposed, containing: a preparatory stage wherein at least one EEG rest signal is preprocessed; informative features are extracted from at least one EEG rest signal, namely, channel synchronisation indicators and spectral power indicators; vectors are built based on the informative features extracted from said at least one EEG rest signal; vectors are built based on the informative feature vector and demographic data; a neural network is trained, wherein at least one vector built at the previous stage is supplied to the input of the neural network, a trained neural network is obtained at the output; a working stage wherein informative features are extracted from at least one EEG rest signal; vectors are built based on the informative features extracted from said at least one EEG rest signal; vectors are built based on the informative feature vector and demographic data; informative feature vector of the EEG rest signal are supplied to the input of the trained neural network, the result of the predicted diagnosis is obtained at the output.

EFFECT: ensures identification of depression in a patient based on the data of the EEG rest signal.

2 cl, 2 dwg

Description

FIELD OF TECHNOLOGY

The present technical solution relates to the field of computing, in particular, to a method for detecting depression based on electroencephalogram data (hereinafter - EEG), using machine learning.

LEVEL OF TECHNOLOGY

From the prior art known source of information RU 2 689 886 C1, publ. 05/29/2019, revealing a method for the diagnosis of affective disorders, consisting in conducting electroencephalographic studies using leads "10-20" and ear ipsilateral electrodes as reference electrodes with spectral and coherent electroencephalogram (EEG) analyzes. Determine the indices of the spectral power in the frequency range of 2-4 Hz on the leads T6-A2 and T5-A1, the indices of the square of the coherence modulus C ² between the leads O1-P3 in the range of 20-21 Hz; between leads T5-P3 in the range of 7-8 Hz; between leads F7-F3 in the range of 14-15 Hz; between leads T3-C3 in the range of 9.5-11 Hz; between leads P4-F7 in the range of 24.5-26 Hz; calculate the indices of normalized coherence (NK) by the formula NK = Ln (C ² / (1-C ² )), then calculate the index Y by the formula Y = -A + 1.6B + 0.69C-0.63D-0.32E + 0.097F + 1.5;

where

A - NC between leads O1-P3 in the range of 20-21 Hz;

B - Ln (P (R) / P (L)),

where P (R) is the spectral power in the range of 2-4 Hz at the T6 lead; P (L) - on lead T5;

C - NC between leads T5-P3 in the range of 7-8 Hz;

D - NC between leads F7-F3 in the range of 14-15 Hz;

E - NC between leads T3-C3 in the range of 9.5-11 Hz;

F - NC between leads P4-F7 in the range of 24.5-26 Hz;

and when the values of the Y indicator are equal to or more than zero, depressive affective disorder is diagnosed, when Y is less than zero, anxiety affective disorder is diagnosed.

Also from the prior art from the information source KR20080082665A, publ. 09/11/2008, a solution is known that discloses a system and method for analyzing and assessing depression and other mood disorders using electroencephalographic (EEG). A wealth of brain wave data associated with an individual is obtained. Determine at least one static component of the EEG data. Determine the static asymmetry in the static component of the set of EEG data and

determine indications as to whether an individual is at risk for a mood disorder based at least in part on static asymmetry in the static component of portions of the plurality of EEG data.

The proposed technical solution differs from the solutions known from the prior art in that the analysis is carried out on the EEG data of the state of rest. In addition, the neural network is trained on a vector of informative features, which include channel synchronization data and spectral power indicators, and patient demographic data.

SUMMARY OF THE INVENTION

The technical problem to be solved by the claimed technical solution is the creation of a method for detecting depression based on EEG data, which is described in an independent claim. Additional embodiments of the present invention are presented in the dependent claims.

The technical result consists in determining the patient's depression on the basis of EEG data.

The claimed result is achieved by implementing a method for detecting depression based on EEG data, containing:

The preparatory stage, at which:

carry out preprocessing of at least one EEG rest signal;

extract informative signs from at least one EEG rest signal, namely, signs of channel synchronization and spectral power indicators;

constructing vectors based on informative features extracted from at least one EEG rest signal;

build vectors based on the vector of informative features and demographic data;

training of the neural network is carried out, while at least one vector constructed at the previous step is fed to the input of the neural network, and the trained neural network is obtained at the output.

The working stage at which:

extract informative signs from at least one EEG rest signal;

constructing vectors based on informative features extracted from at least one EEG rest signal;

build vectors based on the vector of informative features and demographic data;

vectors of informative signs of the EEG signal are fed to the input of the trained neural network; the result of the predicted diagnosis is obtained at the output.

In another particular embodiment of the proposed method, the logistic regression machine learning algorithm is used to classify EEG rest signal data.

DESCRIPTION OF DRAWINGS

The implementation of the invention will be described in the following in accordance with the accompanying drawings, which are presented to clarify the essence of the invention and in no way limit the scope of the invention. The following drawings are attached to the application:

1 illustrates an exemplary embodiment of the method.

FIG. 2 illustrates an example of a general arrangement of a computing device.

DETAILED DESCRIPTION OF THE INVENTION

In the following detailed description of an implementation of the invention, numerous implementation details are set forth in order to provide a thorough understanding of the present invention. However, it will be obvious to those skilled in the art how the present invention can be used, with or without these implementation details. In other instances, well-known techniques, procedures, and components have not been described in detail so as not to obscure the details of the present invention.

In addition, it will be clear from the above description that the invention is not limited to the above implementation. Numerous possible modifications, alterations, variations and substitutions, while retaining the spirit and form of the present invention, will be apparent to those skilled in the art.

The proposed method for detecting depression based on the data of the EEG rest signal is performed on a computing device and is shown in Fig. 1.

At the preparatory stage, the computing device receives multimodal biomedical data of the patient, which includes:

- patient demographic data (gender, age);

- raw neuroimaging data of the subject: 1 minute of resting EEG recording with eyes closed and 1 minute of resting EEG recording with eyes open.

Preprocessing of one EEG signal of rest is carried out, while informative signs are extracted from one EEG rest signal, by fixing a set of procedure parameters selected by the user, such as the length of the signal time window and the metric used to calculate the synchronization (spectral coherence, correlation between signal envelopes). Informative features can be, but are not limited to, channel synchronization indicators and spectral power indicators. Vectors are constructed based on informative features extracted from one EEG signal of one patient.

The EEG resting signal preprocessing procedure is used for neuroimaging data of each patient, while using a fixed set of parameters, namely the length of the signal time window and the metric used to calculate the synchronization (spectral coherence, correlation between signal envelopes). Vectors are constructed based on the informative features of each patient.

The choice of a set of parameters, as well as the fixation of this set and the extraction of features is carried out using scientific research (conducting computational experiments, analyzing them and applying optimization procedures) carried out in the development of a predictive model. Using the MNE software package, which is designed to study, visualize and analyze human neurophysiological data (MEG, EEG, sEEG, ECoG and others), informative signs are extracted and preprocessing of the EEG rest signal is performed.

EEG resting signal preprocessing includes at least detecting artifacts of their resting EEG signal, interpolating bad channels, filtering and counting data, etc. Artifacts are parts of the recorded signal that arise from sources other than the source of interest.

The demographic data of each patient is added to the resulting vectors.

The constructed set of vectors of informative signs of all patients, including the patient's demographic data along with their diagnoses (depression or depression is absent) is considered as a training sample.

A neural network is trained on a training sample by applying the developed machine learning procedure based on the "Logistic regression" method.

The result of the trained neural network will be the decision rule required to make a predictable diagnosis and its reliability characteristics, estimated from the training set using the "5 fold cross-validation, 5 repeats" cross validation technology.

The trained neural network has the following reliability characteristics:

• specificity of the model: 95%,

• sensitivity of the model: 96%.

The result of applying this model to the vector of informative signs (built from neuroimaging and demographic data) of a survey subject with an unknown diagnosis is a predictable diagnosis ("Depression" or "Healthy"). The choice of a machine learning procedure is carried out using scientific research (conducting computational experiments, analyzing them and applying optimization procedures, for example, enumerating and comparing features, selecting the best features and checking the stability of features at different time intervals of EEG recording) carried out in the development of a predictive model.

After training the neural network, the working stage of the proposed method is carried out, in which multimodal biomedical data of the patient are obtained, preprocessing of at least one EEG rest signal is carried out. Vectors are constructed based on informative features extracted from at least one EEG rest signal, as previously indicated.

The vectors of informative signs of the EEG rest signal are fed to the input of the trained neural network; at the output, the result of the predicted diagnosis "Depression" or "Healthy" is obtained.

FIG. 2, a general diagram of a computing device (200) will be presented below, which provides data processing necessary for the implementation of the claimed solution.

In general, the device (200) contains such components as: one or more processors (201), at least one memory (202), data storage means (203), input / output interfaces (204), I / O means ( 205), networking tools (206).

The processor (201) of the device performs the basic computational operations necessary for the operation of the device (200) or the functionality of one or more of its components. The processor (201) executes the necessary computer readable instructions contained in the main memory (202).

Memory (202), as a rule, is made in the form of RAM and contains the necessary program logic that provides the required functionality.

The data storage medium (203) can be performed in the form of HDD, SSD disks, raid array, network storage, flash memory, optical information storage devices (CD, DVD, MD, Blue-Ray disks), etc. The means (203) allows performing long-term storage of various types of information, for example, the aforementioned files with user data sets, a database containing records of time intervals measured for each user, user identifiers, etc.

Interfaces (204) are standard means for connecting and working with the server side, for example, USB, RS232, RJ45, LPT, COM, HDMI, PS / 2, Lightning, FireWire, etc.

The choice of interfaces (204) depends on the specific implementation of the device (200), which can be a personal computer, mainframe, server cluster, thin client, smartphone, laptop, etc.

As means of I / O data (205) in any embodiment of a system that implements the described method, a keyboard should be used. The hardware design of the keyboard can be any known: it can be either a built-in keyboard used on a laptop or netbook, or a stand-alone device connected to a desktop computer, server or other computer device. In this case, the connection can be either wired, in which the connecting cable of the keyboard is connected to the PS / 2 or USB port located on the system unit of the desktop computer, or wireless, in which the keyboard exchanges data via a wireless communication channel, for example, a radio channel, with base station, which, in turn, is directly connected to the system unit, for example, to one of the USB ports. In addition to the keyboard, I / O data can also include: joystick, display (touch screen), projector, touchpad, mouse, trackball, light pen, speakers, microphone, etc.

Networking means (206) are selected from a device that provides network reception and transmission of data, for example, Ethernet card, WLAN / Wi-Fi module, Bluetooth module, BLE module, NFC module, IrDa, RFID module, GSM modem, etc. The means (205) provide the organization of data exchange via a wired or wireless data transmission channel, for example, WAN, PAN, LAN, Intranet, Internet, WLAN, WMAN or GSM.

The components of the device (200) are interconnected via a common data bus (210).

In the present application materials, the preferred disclosure of the implementation of the claimed technical solution has been presented, which should not be used as limiting other, particular embodiments of its implementation, which do not go beyond the scope of the claimed scope of legal protection and are obvious to specialists in the relevant field of technology.

Claims (12)

1. A computer-implemented method for detecting depression based on EEG rest signal data, containing a processor and a memory storing instructions executed by the processor, the method comprising two stages: preparatory stage, at which: carry out preprocessing of at least one EEG rest signal; extract informative signs from at least one EEG rest signal, namely, channel synchronization indicators and spectral power indicators; constructing vectors based on informative features extracted from at least one EEG rest signal; build vectors based on the vector of informative features and demographic data; training the neural network is carried out, while at least one vector constructed in the previous step is fed to the input of the neural network, and the trained neural network is obtained at the output; working stage at which: extract informative signs from at least one EEG rest signal; constructing vectors based on informative features extracted from at least one EEG rest signal; build vectors based on the vector of informative features and demographic data; vectors of informative signs of the EEG signal are fed to the input of the trained neural network, and the result of the predicted diagnosis is obtained at the output. 2. The method according to claim 1, characterized in that the machine learning algorithm "logistic regression

Images