EEG Signal Processing and Machine Learning. Saeid Sanei
Читать онлайн книгу.
of EEG rhythms [17]. During these years the neurophysiologists demonstrated the thalamocortical relationship through anatomical methods. This leads to the development of the concept of centrencephalic epilepsy [18, 30].
Throughout the 1950s the work on EEGs expanded in many different places. During this time surgical operation for removing the epileptic foci became popular and the book entitled Epilepsy and the Functional Anatomy of the Human Brain (Jasper and Penfield) was published. During this time microelectrodes were invented. They were made of metals such as tungsten or glass, filled with electrolytes such as potassium chloride, with diameters of less than 3 μm.
Depth EEG of a human was first obtained with implanted intracerebral electrodes by Mayer and Hayne (1948). Invention of intracellular microelectrode technology revolutionized this method and was used in the spinal cord by Brock et al. in 1952, and in the cortex by Phillips in 1961.
Analysis of EEG signals started during the early days of EEG measurement. Berger assisted by Dietch (1932) applied Fourier analysis to EEG sequences which was rapidly developed during the 1950s. Analysis of sleep disorders with EEGs started its development in the 1950s through the work of Kleitman at the University of Chicago.
In the 1960s the analysis of EEGs of full‐term and premature newborns began its development [19]. Investigation of evoked potentials (EPs), especially visual EPs, as commonly used for monitoring mental illnesses, progressed during the 1970s.
The history of EEG however has been a continuous process which started from the early 1300s and has brought daily development of clinical, experimental, and computational studies for discovery, recognition, diagnosis, and treatment of a vast number of neurological and physiological brain abnormalities as well as the rest of human central nervous system (CNS). At this time, EEGs are recorded invasively and noninvasively using fully computerized systems. The EEG machines are equipped with many signal processing tools, delicate and accurate measurement electrodes, and enough memory for very‐long‐term recordings of several hours. EEG or MEG machines may be integrated with other neuroimaging system such as fMRI. Very delicate needle‐type electrodes can also be used for recording the EEGs from over the cortex (electrocorticogram), and thereby avoid the attenuation and nonlinearity effects induced by the skull. We next proceed to describe the nature of neural activities within the human brain.
1.3 Neural Activities
The CNS generally consists of nerve cells and glia cells, which are located between neurons. Each nerve cell consists of axons, dendrites, and cell bodies. Nerve cells respond to stimuli and transmit information over long distances. A nerve cell body has a single nucleus and contains most of the nerve cell metabolism especially that related to protein synthesis. The proteins created in the cell body are delivered to other parts of the nerve. An axon is a long cylinder, which transmits an electrical impulse and can be several metres long in vertebrates (giraffe axons go from the head to the tip of spine). In humans the length can be a percentage of a millimetre to more than a metre. An axonal transport system for delivering proteins to the ends of the cell exists and the transport system has ‘molecular motors’ which ride upon tubulin rails.
Dendrites are connected to either the axons or dendrites of other cells and receive impulses from other nerves or relay the signals to other nerves. In the human brain each nerve is connected to approximately 10 000 other nerves, mostly through dendritic connections.
The activities in the CNS are mainly related to the synaptic currents transferred between the junctions (called synapses) of axons and dendrites, or dendrites and dendrites of cells. A potential of 60–70 mV with negative polarity may be recorded under the membrane of the cell body. This potential changes with variations in synaptic activities. If an action potential (AP) travels along the fibre, which ends in an excitatory synapse, an excitatory post‐synaptic potential (EPSP) occurs in the following neuron. If two APs travel along the same fibre over a short distance, there will be a summation of EPSPs producing an AP on the post‐synaptic neuron providing a certain threshold of membrane potential is reached. If the fibre ends in an inhibitory synapse, then hyperpolarization will occur, indicating an inhibitory post‐synaptic potential (IPSP) [20, 21]. Figure 1.3 shows the above activities schematically.
Figure 1.3 The neuron membrane potential changes and current flow during synaptic activation recorded by means of intracellular microelectrodes. APs in the excitatory and inhibitory presynaptic fibre respectively lead to EPSP and IPSP in the post‐synaptic neuron.
Following the generation of an IPSP, there is an overflow of cations from the nerve cell or an inflow of anions into the nerve cell. This flow ultimately causes a change in potential along the nerve cell membrane. Primary transmembranous currents generate secondary ional currents along the cell membranes in the intracellular and extracellular space. The portion of these currents that flow through the extracellular space is directly responsible for the generation of field potentials. These field potentials, usually with less than 100 Hz frequency, are called EEGs when there are no changes in the signal average and called DC potential if there are slow drifts in the average signals, which may mask the actual EEG signals. A combination of EEG and DC potentials is often observed for some abnormalities in the brain such as seizure (induced by pentylenetetrazol), hypercapnia, and asphyxia [22]. We next focus on the nature of APs.
1.4 Action Potentials
AP is actually the information transmitted by a nerve. These potentials are caused by an exchange of ions across the neuron membrane and an AP is a temporary change in the membrane potential that is transmitted along the axon. It is usually initiated in the cell body and normally travels in one direction. The membrane potential depolarizes (becomes more positive) producing a spike. After the peak of the spike the membrane repolarizes (becomes more negative). The potential becomes more negative than the resting potential and then returns to normal. The APs of most nerves last between 5 and 10 ms. Figure 1.4 shows an example of an AP.
Figure 1.4 An example of an AP.
The conduction velocity of APs lies between 1 and 100 m s−1. APs are initiated by many different types of stimuli; sensory nerves respond to many types of stimuli, such as: chemical, light, electricity, pressure, touch, and stretching. Conversely, the nerves within the CNS (brain and spinal cord) are mostly stimulated by chemical activity at synapses.
A stimulus must be above a threshold level to set off an AP. Very weak stimuli cause a small local electrical disturbance, but do not produce a transmitted AP. As soon as the stimulus strength goes above the threshold, an AP appears and travels down the nerve.
The spike of the AP is mainly caused by opening of Na (sodium) channels. The Na pump produces gradients of both Na and K (potassium) ions, both are used to produce the AP; Na is high outside the cell and low inside. Excitable cells have special Na and K channels with gates that open and close in response to the membrane voltage (voltage‐gated channels). Opening the gates of Na channels allows Na to rush into the cell, carrying a +ve charge. This makes the membrane potential positive (depolarization), producing the spike. Figure 1.5 shows the stages of the process during evolution of an AP for a giant squid.
For a human being the amplitude