A Brief Survey of Quantitative EEG

Kaushik Majumdar

• Publisher: CRC Press
• Year: 2017
• ISBN:
  9781351643511
• Edition: 1
• Topic: Physical Sciences
• Subtopic: Physics
• Index: Physics

This book covers various quantitative methods for preprocessing and analyzing human EEG signals. It presents a holistic approach to quantitative EEG from its neurological basis to simultaneous EEG and fMRI studies. Equal emphasis is given to major mathematical and statistical theories and computational techniques that have been in use in qEEG and their applications on clinical and laboratory experimental EEG. The book is compact and self-contained, requiring no background in EEG processing or acquisition and quantitative techniques.

1
Neurophysiology of the Human Scalp EEG

Every living brain generates electrical potential spontaneously without break. This was detected by Richard Caton of England in the brains of rabbits and monkeys and published his findings in 1875, which is known to be the first recorded case of brain electrical field potential detection. In the 1920s, Hans Berger in Germany started systematically studying it, as it was detectable on the human scalp. It was he who coined the term electroencephalogram (EEG) in 1934 for the recording of the brain electrical field potential. Generally, the term EEG refers to the recording of the scalp potential. In this book, we will remain focused on human EEG.

The brain is composed of two types of cells namely, glial cells and nerve cells, or neurons. Neurons are electrically excitable, which means that in response to specific inputs each neuron can generate electric potential, called action potential. The human brain contains neurons on the order of 10¹¹, about 85% of which are excitatory pyramidal neurons. These are the cells predominantly responsible for generating the EEG. To be more precise, EEG of the human scalp is an indirect effect of the action potential of the tens of thousands of pyramidal neurons firing simultaneously.

Pyramidal neurons are also long projection neurons, which means they can receive input from one region of the brain and send the output through their long axon to another region up to a few centimeters away from the input zone. The remaining, about 15%, neurons of the brain are mostly inhibitory interneurons (there are excitatory interneurons too). Interneurons are those whose input and output remain confined within a small neighborhood. Pyramidal neurons have only a few varieties, whose differences are mostly regarding their shapes, whereas interneurons have much more versatile variations, and one of their major tasks is to keep in check the firing of action potentials by the pyramidal neurons. When this balance is disturbed, neurological disorders like epileptic seizures may occur. Balanced firing activities at the neuronal network levels in the brain as a combined action of excitatory neurons and inhibitors are key to the normal functioning of the brain. We may consider pyramidal neurons to be electric powerhouses in the brain, which become activated by inputs from the sensory world as well as from within the brain. Interneurons constitute sophisticated electronic control switches, which give rise to a vast number of different output electrical patterns in the generated electric field. These patterns are the representations of the sensory processing and the internal thought processes in the brain. EEG is supposed to capture, at least partly, some of these patterns that are strong enough to be detected on the scalp.

Here are two major caveats: (1) Our brain (or the cortex to be more particular) has six layers, electrical activities in the topmost layer (layer 1) are likely to be captured only in the scalp EEG. Interestingly, layer I does not contain any pyramidal cells. It contains mostly interconnections between the cells. Major inputs come from the pyramidal neurons and branch out to other neurons, including a large number of pyramidal neurons. So, the EEG is generated in the junctions (known as synapses) in between two nerve cells. (2) At any given time our brain is engaged in multiple tasks. So patterns representative of different tasks being executed simultaneously will superimpose, from which separating out a particular pattern very precisely is impossible using the current technologies. Nevertheless, some dominant patterns can be detected with a good degree of certainty, and here exactly lies the usefulness of the EEG. In order to make meaningful observations, computational processing of the EEG signals almost always has to be done under certain neurophysiological constraints. Therefore, a good biological knowledge of the EEG is a must for anyone in the EEG signal-processing community. The current chapter is devoted to this purpose.

1.1Neural Basis of EEG

The human brain is shown in Figure 1.1. The outermost wrinkled layer is called the cortex or the neocortex. This is what is precisely known as the gray matter, responsible for our intelligence. Only mammals have the neocortex. Not all mammals have a folded neocortex like humans. For example, the mouse neocortex is quite flat. In the case of the mammals, the terms neocortex, cortex, and brain are often used i...

Table of contents

• Cover
• Halftitle Page
• Title Page
• Copyright Page
• Contents
• Preface
• Acknowledgments
• 1. Neurophysiology of the Human Scalp EEG
• 2. Preprocessing
• 3. Source Localization
• 4. Event-Related Potential
• 5. Binding Problem
• 6. Epilepsy Research
• 7. Brain–Computer Interface
• 8. An Overview of fMRI
• 9. Simultaneous EEG and fMRI
• Appendix A: Fourier Transformation
• Appendix B: Wavelet Transformation
• Index