MEG/EEG in the Study of Brain Function

This chapter summarizes some relative advantages and disadvantages of MEG and EEG, most of which have been previously elaborated. MEG and EEG are the two sides of the same coin and provide complementary information about the human brain’s neurodynamics. The combined use of MEG or EEG together and with other noninvasive methods used to study human brain function is advocated to be important for future research in systems and cognitive/social neuroscience. This chapter also examines combined use and interpretation of MEG/EEG with MRI/fMRI, and performing EEG recordings during non-invasive brain stimulation.

Auditory Responses

This chapter briefly describes the various types of evoked and event-related responses that can be recorded in response to auditory stimulation, such as clicks and tones, and speech. Transient auditory-evoked responses are generally grouped into three major categories according to their latencies: (a) brainstem auditory evoked potentials occur within the first 10 ms, typically with 5–7 deflections, (b) middle-latency auditory-evoked potentials occur within 12 to 50 ms, and (c) long-latency auditory-evoked potentials range from about 50 to 250 ms with generators in the supratemporal auditory cortex. Steady-state auditory responses can be elicited by periodic stimuli, They can be used in frequency-tagging experiments, for example in following inputs from the left and right ear to the auditory cortices of both hemispheres.

Artifacts

This chapter focuses on different types of biological and nonbiological artifacts in MEG and EEG recordings, and discusses methods for their recognition and removal. Examples are given of various physiological artifacts, including eye movements, eyeblinks, saccades, muscle, and cardiac activity. Nonbiological artifacts, such as power-line noise, are also demonstrated. Some examples are given to illustrate how these unwanted signals can be identified and removed from MEG and EEG signals with methods such as independent component analysis (as applied to EEG data) and temporal signal-space separation (applied to MEG data). However, prevention of artifacts is always preferable to removing or compensating for them post hoc during data analysis. The chapter concludes with a discussion of how to ensure that signals are emanating from the brain and not from other sources.

An Overview of EEG and MEG

This chapter discusses the rather different histories of MEG and EEG. EEG has been used as a tool of clinical diagnostics since the 1930s with visual inspection of spontaneous EEG activity in patients suffering from various brain disorders, particularly epilepsy. For MEG, the evolution of applications has been the opposite: the first recordings were made in research laboratories and the recordings started by averaging a very high number of single trials, and it took some time before low-noise MEG equipment became available to allow single-trial analyses. Only with recent technical developments, especially with the advent of whole scalp–covering devices, has MEG become increasingly popular in the clinical environment. We discuss the evolution of both techniques with respect to the measurement of brain rhythms and evoked and event-related responses.

Introduction

Neuronal communication in the brain is associated with minute electrical currents that give rise to both electrical potentials on the scalp (measurable by means of electroencephalography [EEG]) and magnetic fields outside the head (measurable by magnetoencephalography [MEG]). Both MEG and EEG are noninvasive neurophysiological methods used to study brain dynamics, that is temporal changes in the activation patterns, and sequences in signal progression. Differences between MEG and EEG mainly reflect differences in the spread of electric and magnetic fields generated by the same electric currents in the human brain. This chapter provides an overall description of the main principles of MEG and EEG and provides background for the following chapters in this and subsequent sections.

Brain Disorders

This chapter discusses clinical applications of spontaneous EEG and MEG as well as evoked responses in epileptic and stroke patients, and in presurgical mapping (including identification of the central sulcus). EEG and MEG remain, due to their excellent temporal resolution, the methods par excellence for diagnosing and identifying epileptic syndromes. Timing of discharges can differentiate primary and secondary (mirror) epileptic foci. In contrast to studies of clinical populations for research purposes, clinical assessment is always based on findings in individual subjects, and the tests thus have to be reliable with high specificity and high sensitivity, showing statistically significant differences compared with normative values of healthy subjects with similar attributes, including age. The chapter ends by discussing EEG monitoring in coma or brain death.

Other Sensory Responses and Multisensory Interactions

This chapter discusses olfactory and visceral responses as well as the MEG/EEG signature of multisensory interaction. Olfactory stimuli can be embedded in a continuous humified airflow where the stimuli are presented at intervals of tens of seconds to avoid short-term habituation. Visceral stimulation typically requires purpose-built stimulating electrodes for direct access to the viscera. Studies of multisensory interaction are necessary because our everyday experiences involve inputs from multiple senses, the temporal coincidence of which allows the brain to construct representations of unique objects or events. Detection and correct interpretation of the nonlinear multisensory interactions call for careful considerations of both the sites of interaction and the changes in the amplitudes of evoked responses and spontaneous activity.

Practicalities of Data Collection

This chapter provides a number of suggestions about optimization of MEG/EEG recording sessions to guarantee as good signal quality and as high signal-to-noise ratio as possible. It also advices for performing replicability checks on the data. The practical aspects of preparing and performing an EEG recording (skin preparation, electrode-impedance measurement) and postrecording infection control are presented. Similarly, practicalities of performing MEG recordings are discussed. Measurement of MEG sensor-array location with respect to landmarks on the scalp and different methods for the measurement of EEG electrode positions are presented. The chapter ends with a discussion on electrical safety in the MEG/EEG laboratory.

Brain Signals Related to Change Detection

This chapter discusses, in the context of the predictive-coding framework, evoked responses to various changes in the environment and describes how the responses are related to variations in stimulus probability and the subject’s expectations. The focus is on three well-known responses: (a) the mismatch negativity peaking at 100 to 250 ms and elicited to changes in stimulus attributes, even when the stimuli are not attended to, (b) the P300 response peaking about 300 ms after attended low-probability “oddball” stimuli, and (c) the N400 peaking about 400 ms after semantic or lexical violations of sentences presented either visually or auditorily. Continent negative variation and error-related negativity are introduced as well.

Motor Function

Voluntary movements are preceded by slow brain activity, visible in EEG as the Bereitschaftspotential (the readiness potential), and in MEG as the readiness field. These slow shifts can begin a few seconds before movement onset in the primary motor cortex and in the premotor areas. Cortex–muscle coherence refers to coupling between MEG/EEG signals and the surface EMG of a steadily contracted muscle; it typically occurs at around 20 Hz and implies an efferent drive from the cortex to the muscle. Corticokinematic coherence can be measured as the coupling between MEG/EEG signals and the acceleration or velocity of a rhythmically moving limb; it typically occurs are the movement frequency and its first harmonic. Coherence of MEG/EEG signals can be computed also with respect to other peripheral signals, such as the fundamental frequency of the voice measured with an accerometer above the subject’s throat.