To record a typical EEG from humans, multiple surface recording electrodes are glued to the scalp and one ► reference electrode is positioned preferably over an isoelectric part (linked earlobes, nose, and mastoid bone). The signals of interest are in the microvolt (millionths of a volt) range and sensitive to extracerebral
Electroencephalography. Fig. 1. Diagram summarizing common straightforward relationships in brain physiology, behavior, and psychopharmacology research and applications.
noise-sources. Although there exists no ideal zero-reference, scalp electrode recordings are improved by subtracting the common reference signal. Artifacts due to muscle activity or movement occur in EEG recordings, and they must be removed. Such artifacts are frequently caused by eye movements, blinks, heart beating/blood circulation, facial/scalp muscle activity, breathing, and swallowing. To help with artifact removal, polygraphic recordings are often performed alongside the EEG recording in both humans as well as animals, by the placement of additional sensors for monitoring, e.g., eye movements (EOG), postural/neck muscles (EMG), heart activity (EKG), thoracic expansion, and blood pressure. In practice though, the eye of the expert, who examines the entire multichannel EEG recording in addition to the accessory records, is the best means for artifact removal and epoch selection for the analyses in both wakefulness and ► polysomnography.
Sleep pattern characterization is usually carried out by visual inspection (see scoring rules in Iber et al. 2007) on cleaned traces in the temporal domain. Digital EEGs in animals and humans are nowadays nearly always obtained (as recommended, e.g., Penzel and Conradt 2000) with commercially available equipment. This has promoted the field of quantified wake EEG or ► qEEG which uses the fast fourier transform to calculate power spectra of the EEG segments and thus to quantify changes in the various frequency bands. Multivariate EEG-analysis by way of ► spectrograms provides valuable biomarkers for the endogenous modulation (by speeding or slowing)
of the level of arousal, for the continuum in the change of patterns from adolescence, adulthood to normal brain aging, or for occasional EEG-changes related to life style, but also pathological states. Importantly, multivariate EEG-analysis also allows the study of perturbations induced by drugs.
Pharmaco-EEG is a technique to evaluate the effect of drugs on electrical brain activity in a defined context and implicitly makes use of qEEG. A large number of currently available therapeutics act on receptors/channels/^ neurotransmitter transporters/enzymes in the brain, which take part in or influence synaptic transmission. Many of these drugs also influence the EEG power spectra, thus pharmaco-EEG has become a useful research tool in psychopharma-cology; it is most commonly (but not exclusively) used in preclinical and early phase I ► randomized controlled trials for novel drugs. This technique allows to evaluate the cerebral ► bioavailability of compounds in man in comparison to animals (► translational research). qEEG provides a high-dimensional biomarker with the potential to perform systematic ► classification of psychoactive drugs, drug "finger-printing" (Saletu 1987), and risk evaluation. The versatility of EEG parameters lies in the fact that it provides a metric which is objective and can be expressed as (single subject or group) "raw" data, either as predose-adjusted value or as placebo-adjusted value. EEG data with good test-retest reliability are obtained when the volunteer is awake but with the eyes closed, whereas in freely-moving animals, the exact behavioral state and other confounding factors are more difficult to control.
Neurons firing action potentials generate extracellular electromagnetic gradients (► extracellular recording) in the cortex. When the timing is sufficiently coherent, synchronous activity gives rise to the microvolt signals named ► field potentials that make up the EEG by volu-mic summation of these elementary (neuronal) events; optimal geometry is an arrangement in the form of a "palisade-like" array. The macroscopic EEG signals switch polarity in a cyclic manner, and zero-crossings for the alternating waveforms can be as fast as 20 ms for a particular structure, or up to >1,000 ms for a different structure. The technique does not allow to reconstruct the individual ► firing patterns of neurons. Exactly which type of neurons and interactions contribute to the EEG and from where in the brain the different characteristic rhythms originate are the subjects of intensive research. The present view of the cellular aspects of generators of EEG is that particular neuron types, which are located in different parts of the CNS, have membrane properties that
Was this article helpful?