Presenting a new index to measure the level of brain fatigue during mental activity from the EEG signal

Master's Thesis in Computer Engineering (Artificial Intelligence)

Abstract

In recent years, many methods have tried to evaluate mental fatigue with different criteria. These methods have used different scales for this task including performance and electrophysiological based measurements. Among these tools, electroencephalogram (EEG) seems to work better and more accurately than other tools. However, most of the research findings on EEG changes related to fatigue have different limitations and sometimes conflicting results. As a result, in order to better detect fatigue from the EEG signal, it is necessary to investigate it more.

Most of these methods have a high extracted feature dimension, and as a result, they need dimension reduction methods to reduce this dimension and also increase accuracy. Also, the accuracy of these methods depends on the dimension reduction method used and the number of features used in them. In addition, these methods have mostly investigated mental fatigue in a few limited cases. As a result, we have investigated the effect of fatigue on the strength and location of brain resources in order to increase the speed and accuracy of mental fatigue investigation. With the help of this method, an attempt has been made to reduce the computational complexity of previous methods. We have also continuously investigated mental fatigue. In addition, in this thesis, both the real signals recorded from different people and the simulated signal have been used to demonstrate the correctness of the proposed method, and we have shown that this method performs better than the previous methods.  

Chapter One

1-1-

Fatigue is a common phenomenon in our daily life. A common definition of fatigue is that fatigue is a state that follows a period of mental or physical activity characterized by a reduction in the ability to work. The concept of mental fatigue was first introduced by Grandjean [1], who clearly differentiated mental fatigue from physical fatigue. He defined physical fatigue as a result of decreased muscle system performance and mental fatigue as a decrease in mental performance and feeling tired. Fatigue has major consequences in road fatalities and is currently one of the major issues in the transportation industry. According to the preliminary works in this case, driver fatigue accounts for 35-45% of road accidents [2]. In addition to this, fatigue causes a decrease in mental efficiency, especially in professional people who have very high mental activity during work (professional computer programmers and industrial system designers who work in the R&D departments of companies) and also increases the response time in people. As a result, in addition to the effects that mental fatigue has on lowering the efficiency of people in various job situations, it can be an important factor in road accidents and measuring the efficiency of people in factories. As a result, since by being tired, a person has difficulty in performing work with sufficient performance power, and considering the relationship that fatigue has in increasing the probability of accidents on roads and factories [3 and 4], it seems necessary to determine the level of fatigue of a person, in reducing such accidents and also increasing the performance power of people. As a result, we have investigated mental fatigue in this research (in the following, fatigue is used to mean mental fatigue). Among these methods, it seems that the recorded signal of electrical activity of the brain [2] (EEG) is a better indicator of the degree of fatigue and has more predictive power in the diagnosis of brain fatigue [5]. The first recording of brain electrical activity from rabbit and monkey brains was reported by Caton[3] in 1875 [6], but it was in 1929 that Hans Berger[4] [7] reported the first measurement of brain electrical activity in humans. After that, this signal was used in practical diagnostics, especially for various diseases. Since it is widely accepted that characteristic changes in the EEG waveform and its power bands can be used to characterize the transition from wakefulness to sleep and different stages of sleep [8], EEG has been observed as a standard for measuring the level of wakefulness and sleepiness.

However, there are significant differences among the current fatigue detection algorithms based on EEG. Previous studies have shown that the relationship between EEG changes and the degree of fatigue depends on the type of work and the state of the person. These studies differ both in the nature of the algorithm for fatigue detection and the location and number of electrodes for signal recording [9]. In addition, all these algorithms face different limitations. For example, many of these methods need methods to reduce the dimension of the extracted feature space to increase the accuracy of their methods. As a result, the purpose of this thesis is to detect the level of fatigue with the help of a method that does not need to reduce the dimension of the data and also observe the effect of fatigue on brain activities. As a result, we have used source location methods to achieve this goal.

There are different methods in the field of locating foci in the brain. Among these methods, we can mention radiation [5] [10], in which, by filtering the data obtained from different electrodes, we try to find the direction and location of the foci that produce these signals. To test the proposed method, we use both the signals recorded from different people and the EEG signal produced during fatigue according to the existing characteristics. As a result, the objectives of this thesis can be summarized in the following.

Achieving an algorithm that can continuously determine the level of fatigue.

Increasing the accuracy and speed of detecting the level of fatigue

In addition, considering the relationship between fatigue and sleep, if this relationship is determined, it may be used in the treatment of diseases such as sleep disorders and other similar diseases.

1-3- Overview To the thesis chapters

The topics mentioned in this thesis are presented in the form of five chapters. The continuation of the mentioned contents can be summarized in the following cases.

Chapter Two. Research background

In this chapter, the most important previous works that have been done so far to investigate mental fatigue along with their characteristics have been studied and investigated.

Chapter three. Research method

In this chapter, first, the noise target method is explained from the recorded data. Then one of the location methods and its problems are explained. After that, a method to improve resource location is proposed. Finally, we will examine the method of determining mental fatigue.

Chapter Four. Experiments and results

In this chapter, the investigated signals are explained at the beginning. Then the different steps explained in the previous chapter and common fatigue detection methods are applied to these signals. Also, the results of applying these methods are explained.

The fifth chapter. Conclusions and suggestions

In the last chapter, the contents mentioned in this thesis are summarized and discussed. Then suggestions and ways to continue and expand this research in future research are presented. Chapter Two Research Background 2 Research Background Since 200 years ago, neuroscientists have sought to determine the tasks and activities performed in the human brain. It was believed that different brain activities involve different areas of the brain. As a result, their primary goal was to define the areas of the brain involved in the most basic tasks humans can perform, such as hearing. With the help of methods such as anatomy, many areas related to these activities were discovered [11].

Only knowing the responsibility of the brain areas did not satisfy the researchers, and researchers are currently more interested in investigating what is happening in the brain or what is its mental state. One of these cases is the concept of mental fatigue, which was introduced by Grandjean in 1981.

Fatigue is a natural phenomenon in our daily life. In general, fatigue is divided into two groups: physical fatigue [6] and mental fatigue [7].