Search Results

As long as valid assumptions can be made about a focal source, MEG can transform the challenge of EEG based 2D inference of lateralization or regional localization to 3D sublobar indication of epilepsy-related spike generators. As such MEG spike source imaging provides a unique tool for targeting epileptogenic tissue for the surgical treatment of epilepsy. This information can be particularly valuable for patients with neocortical epilepsy in whom intracranial EEG (icEEG) investigations are commonly necessary. MEG localization of spikes may provide more accurate electrode sampling of the cortex responsible for seizures, and as a result, increase epilepsy localization and surgical resection accuracy. Combined with mapping of eloquent cortical function, MEG can play a role in multiple aspects of the preoperative (non-invasive) decision-making—potential to improve (1) patient selection, (2) ICEEG yield, and (3) increase the net number of seizure-free outcomes. Work remains to determine the validity of various analysis methods (stratified on different types of spike sources), and the cost effectiveness of MEG in epilepsy surgery, but it can be concluded that any patients able to proced to surgery that otherwise would not without MEG would contribute evidence to added clinical utility even if the cure rate is unchanged.

The highly interdisciplinary nature of EEG is apparently the main reason why many fallacies have appeared in EEG and, in some cases, persisted over long periods. Common EEG fallacies occur on both side of the normal division between the physical and biological sciences. This chapter presents a summary of fallacies with minimal supporting arguments, which are considered in more detail throughout the book. Topics include: the chauvinism of spatial scale (the attitude that data recorded at one scale is more scientific than others), the myth of the quiet reference, use and misuse of mathematical models, the EEG folklore, appropriate and inappropriate methods of EEG data analysis, the often-adopted mantra “artifact-free” data, the extreme non-uniqueness and (often) unreliability of source localization, advantages and limitations of high resolution EEG, over-promotion of brain magnetic field recordings (MEG), and “pacemaker” icons adopted as a psychological crutch to avoid genuine scientific issues.

Functional Imaging with MRI using BOLD techniques plays an increasing role in clinical practice. fMRI may be used for interictal source localization but is more commonly used to identify eloquent areas to be spared during epilepsy surgery. fMRI is used to identify primary sensory cortex –visual, sensory, and auditory and to identify motor cortex. There are several language paradigms that may be used to identify frontal (“expressive”) and temporal (“receptive”) speech. These methods have excellent agreement with Wada and predict post operative naming outcomes. There are circumstances in which fMRI is not helpful as the BOLD response may be disrupted by several pathological processes. fMRI memory tasks have good agreement with Wada, demonstrate material specificity, show functional capacity to be more important than functional reserve, and may predict outcomes of specific memory tasks.

Every EEG measurement depends on the locations of the recording and so-called reference electrode. The measured potential difference is a property of the path between electrode pairs. This applies to any reference including the linked-ears (or linked-mastoid) reference. The physical linked-ears reference provides an unbalanced “random” reference. The mathematical linked-ears reference offers no obvious advantages over other references. The average reference can be used to approximate reference independent potentials if used with a large number of electrodes, but is still biased by the limited sampling of potentials over the lower surface of the head. Regardless of source, the EEG is a low-pass spatially filtered signal, making discrete sampling of the potential distribution feasible without spatial aliasing. Modern EEG systems have facilitated spatial mapping of EEG potentials. Any potential distribution on the scalp can be fit (in a least-squared sense) by an equivalent dipole distribution by solving the inverse problem.

To understand cognition and emotion in the real world, it is critical to investigate the phenomena of interest within the rich context of moment-to-moment variations in the real world, which we assume is at least in part encoded in the high-dimensional state of the brain. Here the chapter reviews empirical evidence from a series of novel validation studies that demonstrate the technical capabilities of one blind source separation (BSS) algorithm—second-order blind identification (SOBI)—in enabling neuronscientists and clinicians to investigate human brain functions, cognition, and behavior using the electroencephalography (EEG). The chapter concludes that by shifting from an EEG-sensor-based to a neuronal-source-based characterization of brain states, one may better capture the rich context of moment-to-moment variations in the real world.

Since its adoption in clinical practice, analysis and interpretation of magnetoencephalography (MEG) recordings in patients with epilepsy have evolved as a result of multidisciplinary input, with the aim of developing formalized criteria that can yield reliable results and that are complementary to other modalities used for diagnostic assessment in this cohort. The purpose of this chapter is to familiarize the interested practitioner with the process of analyzing interictal MEG recordings, using examples from clinical cases to illustrate the utility of this modality and acknowledging some of its limitations. To achieve this, the reader will first be presented with consideration of basic quality assessment measures when interpreting clinical MEG recordings. Subsequently, a description will be provided on a clinically valid approach that is standard of practice for performing source localization of epileptiform transients. Specifically, the latter is discussed in the broader context of identifying various interictal epileptiform transients in MEG recordings and determining their anatomical extent in order to better characterize the irritative zone. Finally, a review is provided for emerging methods in MEG connectivity analysis and their potential utility in clinical practice for elucidating epileptogenic networks.

This chapter on analytical approaches discusses a variety of data analysis methods commonly employed in consumer neuroscience. The emphasis is on the foundations of methods, such as time–frequency representations, phase space reconstruction, source localization, and dynamic causal modelling. In addition, a number of relevant statistical approaches are addressed, such as Bayesian inference and stochastic processes. The issue of so-called reverse inference is discussed in the context of machine learning algorithms that are becoming increasingly important for the analysis of complex data. The goal of the chapter is to equip the reader with a robust overview of what is an active field of research, where new algorithms are being developed at a staggering pace. The chapter concludes with a brief note on issues revolving around artefact rejection, data pre-processing, and analysis pipelines.