EEG Signal Processing and Machine Learning. Saeid Sanei

and artificial intelligence have become popular research focuses.

1.2 History

The first understanding of the brain was in 1700 BCE when Imhotep lived in Egypt (in Edwin Smith Surgical Papyrus). At that time the hieroglyphic for ‘brain’ was presented as that in Figure 1.1. Then, during 460–379 BCE, Hippocrates discussed and introduced epilepsy as a disturbance of the brain. Since then many physicians, clinicians, and philosophers from around the world, particularly from Greece (Roman) and Iran (Persian), have encountered various brain diseases. In his Canon of Medicine (Al‐Qanun fi al‐Tibb), Avicenna (980–1037, also known as Abu Ali Sina), a Persian physician and philosopher, categorizes the causes of epilepsy into two main groups: those caused by brain diseases and those associated with the abnormalities and diseases of other organs.

      However, the history of EEG, as an instrument to record the brain activity, goes back to when for the first time some activity of the brain was recorded or displayed. Carlo Matteucci (1811–1868, Pisa, Italy) and Emil Du Bois‐Reymond (1818–1896, Berlin, Germany) were the first people who registered the electrical signals emitted from muscle nerves using a galvanometer and established the concept of neurophysiology [1, 2]. However, the concept of action current introduced by Hermann Von Helmholtz (1821–1894, Potsdam Germany) [3] clarified and confirmed the negative variations occurring during muscle contraction via measuring the speed of frog nerve impulses in 1849.

      Richard Caton (British, 1842–1926) measured the brain activities of rabbits and monkeys from over the cortex in 1875. He discovered the electrical nature of the brain and laid the groundwork for Hans Berger to discover alpha wave activity in the human brain. He also placed two electrodes over the human scalp to record for the first time the brain activity in the form of electrical signals in 1875. Since then, the concepts of electro‐ (referring to registration of brain electrical activities) encephal‐ (referring to emitting the signals from head) and gram (or graphy), which means drawing or writing, were combined so that the term EEG was henceforth used to denote electrical neural activity of the brain.

Figure 1.1 Hieroglyphic symbol for the ancient Egyptian word for ‘brain’.

Figure 1.2 Physiologists Adolf Beck (Polish, 1863–1942) on the left and Vladimir Pravdich‐Neminsky (Ukrainian, 1879–1952) on the right who performed the first recording of brain activities from over the skull.

Fritsch (1838–1927) and Hitzig (1838–1907) discovered that the human cerebral can be electrically stimulated. Vasily Yakovlevich Danilevsky (1852–1939) followed Caton's work and finished his PhD thesis in the investigation of brain physiology in 1877 [4]. In this work he investigated the brain activity following electrical stimulation as well as spontaneous electrical activity in the brain of animals.

      The cerebral electrical activity observed over the visual cortex of different species of animals was reported by Ernst Fleischl von Marxow (1845–1891). Napoleon Cybulski (1854–1919) provided EEG evidence of an epileptic seizure in a dog caused by electrical stimulation.

      The idea of the association of epileptic attacks with abnormal electrical discharges was expressed by Kaufman [5].

Adolf Beck (Polish, 1863–1942) and Vladimir Pravdich‐Neminsky (Ukrainian, 1879–1952) measured the EEG from over the skull of dogs. Therefore, these two scientists are indeed the pioneers in scalp EEG recording (Figure 1.2).

      Pravdich‐Neminsky recorded EEG from the brain, termed the dura, and the intact skull of a dog in 1912. He observed a 12–14 cycles s⁻¹ rhythm under normal conditions which slowed under asphyxia and later called it the electrocerebrogram.

      Although much research work on EEG principles and measurements has been performed by the above scientists, Hans Berger (1873–1941, Germany) was credited and named the first one for discovering and measuring human EEG signals. He began his study of human EEGs in 1920 [6]. Berger is well known by almost all electroencephalographers. He started working with a string galvanometer in 1910, then migrated to a smaller Edelmann model, and after 1924, to a larger Edelmann model. In 1926, Berger started to use the more powerful Siemens double coil galvanometer (attaining a sensitivity of 130 μV cm⁻¹) [7]. His first report of human EEG recordings of one to three minutes duration on photographic paper was in 1929. In this recording he only used a one channel bipolar method with fronto‐occipital leads. Recording of the EEG became popular in 1924. The first report of 1929 by Berger included the alpha rhythm, as the major component of the EEG signals as described later in this chapter, and the alpha blocking response.

During the 1930s, the first EEG recording of sleep spindles was undertaken by Berger. He then reported the effect of hypoxia on the human brain, the nature of several diffuse and localized brain disorders, and gave an inkling of epileptic discharges [8]. During this time another group established in Berlin‐Buch and led by Kornmüller, provided more precise recording of the EEG [9]. Berger was also interested in cerebral localization and particularly in the localization of brain tumours. He also found some correlation between mental activities and the changes in the EEG signals.

      Toennies (1902–1970) from the group in Berlin built the first biological amplifier for the recording of brain potentials. A differential amplifier for recording of EEGs was later produced by the Rockefeller foundation in 1932.

      The importance of multichannel recordings and using a large number of electrodes to cover a wider brain region was recognized by Kornmüller [10]. The first EEG work focusing on epileptic manifestation, and the first demonstration of epileptic spikes was presented by Fischer and Lowenbach [11, 12].

      In England, W. Gray Walter became the pioneer of clinical EEG. He discovered the foci of slow brain activity (delta waves), which initiated enormous clinical interest in the diagnosis of brain abnormalities. In Brussels, Fredric Bremer (1892–1982) discovered the influence of afferent signals on the state of vigilance [13].

      Research activities related to EEGs started in North America in around 1934. In this year, Hallowell Davis illustrated a good alpha rhythm for himself. A cathode ray oscilloscope was used around this date by the group in St. Louis University in Washington, in the study of peripheral nerve potentials. The work on human EEGs started at Harvard in Boston and the University of Iowa in the 1930s. The study of epileptic seizure developed by Fredric Gibbs was the major work on EEGs during these years, as the realm of epileptic seizure disorders was the domain of their greatest effectiveness. Epileptology may be divided historically into two periods [14]: before and after the advent of EEG. Gibbs and Lennox applied the idea of Fischer based on his studies about picrotoxin and its effect on the cortical EEG in animals to human epileptology. Berger [15] showed a few examples of paroxysmal EEG discharges in a case of presumed petit mal attacks and during a focal motor seizure in a patient with general paresis.

      As the other great pioneers of EEG in North America, Hallowell Davis, Herbert H. Jasper, Frederic A. Gibbs, William Lennox, and Alfred L. Loomis were the earliest investigators of the nature of EEG during human sleep. Alfred L. Loomis, E. Newton Harvey, and Garret A. Hobart were the first who mathematically studied the human sleep EEG patterns and the stages of sleep. At McGill University, Herbert Jasper studied the related behavioural disorder before he found his niche in basic and clinical epileptology [16].

      The American EEG Society was founded in 1947 and the first international EEG Congress was held in London, United Kingdom around this time. While the EEG studies in Germany were still limited to Berlin, Japan gained attention by the