Introduction to QEEG and EEG

by Frank H. Duffy, MD, Clinical Neurophysiology Laboratory, Boston Children’s Hospital. This article may be copied and reproduced freely.

WHAT IS EEG?

To understand QEEG one must first understand EEG. EEG is the abbreviation for electroencephalography. Small, non-invasive electrodes (usually 16 to 32 in number) are placed upon a patient’s scalp, after careful measurement by a trained technologist, with paste or a glue like substance to hold them in place. Low voltage signals (5-500 microvolts) are amplified by the EEG machine and results are typically written by ink-fed pens on a moving paper strip chart. The resulting polygraphic strip chart, looking much like a multiple channel seismograph, is typically read by unaided visual inspection. The physician interpreting such a tracing is usually a neurologist with special training in EEG. Such an individual is often referred to as a neurophysiologist or electroencephalographer. Psychiatrists, neurosurgeons, and psychologists may also interpret EEGs but to do so, like neurologists, they require special EEG training. Board certification is available in EEG and other aspects of neurophysiology from several organizations. Similarly EEG technologists should have special training in EEG and may become “registered”.

Techniques for interpretation of EEG by visual inspection have changed little since EEG’s discovery in the 1920s by Berger and its extension to clinical issues in the 30s and 40s by Gibbs, Lennox, Lombroso. Typically the BEG is screened for features that stand out (transient responses) like the spike or spike and wave associated with epilepsy. Next the frequency or spectral content of the remaining EEG background is visually evaluated. There are four broad spectral band of clinical interest: delta (0-4 Hz), theta (4-8 Hz), alpha-(8-12 Hz), and beta (above 12 Hz). Not everyone agrees on the exact boundaries of these rhythms and many subdivide these bands, especially beta. Pathology typically increases slow activity (delta, theta) and diminishes fast activity (alpha, beta). Thus overlying a localized brain tumor one would expect increased slowing and decreased fast activity. Similarly following a global brain insult resulting in a global encephalopathy one might expect globally increased slowing and decreased fact activity. However, there are many exceptions to this oversimplified explanation. EEG interpretation requires considerable skill and often years of clinical experience. The mere determination of whether an EEG spectral band is normal, increased, or decreased may require years of experience. Some have likened BEG reading to the grading of equine or canine confotmation by judges who have spent their careers learning what to look for. EEG interpretation is as much an art as a science.

Modern advances in EEG have included what is referred to as digital EEG or dEEG. Here brain signals are similarly collected from the scalp and amplified but are fed into a computer (i.e., digitized) and then interpreted by viewing them not on paper but on the computer screen. Important advantages include storage of efficient digital media rather than on bulky paper. Another advantage is the ability to view the same EEG signals from different perspectives – paper affords only one view of a time period. A draw-back is that the computer screen may not afford the same clarity of image that is available on paper. Another advance is the more speedy placement of electrodes by using an elastic cap with electrodes already imbedded. Careless use of this technology may result in improperly positioned electrode or poor electrode contact.

EEG has survived the advent of all the modem neuroimaging techniques including pneumoencephalography, arteriography, CT scanning, MRI, fMRI, SPECT and PET and remains the number one diagnostic test for epilepsy. Its advantages, among other measures of brain function, is that demonstrates a nearly diagnostic finding in epilepsy and it is the most sensitive functional test to changes in brain function over short time periods. It lacks primarily in ability to localize exactly where in the brain abnormalities arise. Clinically, therefore, EEG is often combined with other neuroimaging tests. Training in EEG is also very demanding with the value of a given EEG to a patient often determined by who interprets it. This is very true in pediatric EEG and especially true for newborn EEGs. The child and neonatal EEGs are not simply smaller versions of adult EEG. Pediatric EEG is a most demanding specialty.

Good solid texts in EEG are provided by Hughes and also by Neidermeyer.

WHAT IS QEEG?

Introduction

To understand QEEG one must first understand EEG. QEEG, or quantitative EEG, began in the 1970s and early 80s as an attempt to extract from brain electrical activity more than what could be readily appreciated by simple, unaided visual inspection of EEG. In that sense qEEG should be viewed as an extension of and not a replacement for traditional EEG. Clinically, as now used, qEEG should always follow the preparation and analysis of the classic EEG (or dEEG). The human eye is still superior to the computer in many aspects of brain signal analysis. Pioneers in qEEG include names such as Bickford, Duffy, Harner, John, Lehmann, Ueno, and many others.

Spectral Analysis and Mapping

To assist in the estimation of EEG spectral content (one of the most difficult tasks by visual inspection), EEG data are entered into a computer, as for dEEG, and spectral content is rigorously determined by the use of techniques of mathematical signal analysis (typically by the FFT or Fast Fourier Transform algorithm). One of the first problems was how to visualize results since qEEG typically uses more channels than EEG. The solution was to map the results using colored grey scaling on schematic maps of the head. To some, such brain electrical activity mapping or simply “mapping” is taken as synonymous with qEEG. However mapping is only a display technique and only the first step. The heart of qEEG lies with the underlying computerized analytic and statistical techniques.

Spectral Coherence

A special results of spectral analysis is the measure of coherence between two electrodes. It assesses the similarity of spectral content of two electrodes over time and is usually taken to reflect a measure of “coupling” between brain regions. It is virtually impossible to estimate coherence by visual EEG inspection. Some illnesses may begin with abnormalities of cortical coupling. Leuchter has reported such abnormalities in Alzheimer’s disease and Thatcher found abnormality of coherence as the best discriminator of mild closed head injury.

The SPM (statistical probability maps)

Such spectral maps provide excellent displays of the spatial distribution of EEG spectral content and are clinically useful as such. However, it soon became evident that in some way it would be necessary to estimate when such data were outside of normal bounds for a patients age. This lead, first, to the need for and the development of normative data bases of brain electrical activity at all ages. Second, it lead to the development of the technique of mapping not just a patient’s brain activity but also the degree of statistical deviancy of the patient from the normal data base (in units of standard deviation of Z-scores). Such images of deviancy are referred to as SPM (statistical or significance probability maps). Thus a neurophysiologist may look at a SPM and locate regions of possible clinical abnormality by deviant regions on the SPM. The term “encephalopathic” often refers to brains with excessive EEG slowing A typical application would be to determine whether behavioral disturbance in an adult is due to early dementia (increased slowing) or otherwise uncomplicated depression (no increase of slowing). QEEG techniques add significant power to the search for subtle encephalopathic change. Although developed first for qEEG analyses, the SPM technique has been widely adapted for use with other neuroimaging techniques.

Long Latency Evoked potentials

Another area where qEEG techniques have been applied is to the long latency sensory evoked potentials. EEG represents the brain’s ambient, spontaneously ongoing electrical activity. Evoked potentials (EPs) are the brain’s transient response to externally applied stimuli – such as light flashes, auditory clicks, and mild electrical shocks. These stimuli form, respectively, the visual evoked response or potential (VEP), auditory EP (AEP) and somatosensory EP (SEP). Since the EEG is much higher amplitude than the EP, it is necessary to apply a stimulus repetitively at random times and average the result so as to effectively remove the random background EEG and visualize the EP. This computerized technique is often referred to as signal averaging. Classic neurophysiology employs a few EP channels and evaluates the short latency response (e.g., under 30 msec). When obtained these signals are seen to arise from specific deep brain structures and allow for assessment of structures within the brain stem and thalamus. When longer latencies (longer times from stimulation) are evaluated, signals appear to be coming from the cortical mantle. Unfortunately the complex waveform morphologies from a large set of such long latency EPs can be very difficult to analyze by unaided visual inspection. However with the use of normative data bases and the SPM technique, regions of clinically important abnormality can be delineated within the complex combined spatial-temporal information within long latency EP data sets. Many qEEG laboratories incorporate the long latency EP along with spectral analyzed EEG signals and traditional EEG as part of their routine clinical studies. Such EP data tend to be sensitive to clinical conditions where cortical dysfunction is hypothesized (e.g., dyslexia, schizophrenia, Alzheimer’s disease) although they are also often found to be abnormal in epilepsy.

Discriminant Analysis

Discriminant analysis refers to the established “multivariate” statistical technique whereby a multiplicity of gathered data (multiple variables) are combined into a single number (the discriminant function) in such a way that this new variable (the discriminant) maximally separates two patient populations. John, Duffy, and Thatcher have all demonstrated that when discriminant analysis is applied to qEEG data, resulting discriminant functions are accurate in classifying individual subjects into clinically relevant diagnostic groups (e.g., head injured or not, dyslexia or not,

bipolar vs. monopolar depression, etc). Such discriminants are more widely used for psychiatric than neurologic issues.

Epileptic Source Analysis

A major goal in the neurophysiologic investigation of patients with epilepsy is to locate the epileptic focus. This involves determining where inside the three dimensional brain, the abnormal signals are generated using only data gathered from the intact scalp. This is a key prelude to removal of the epileptic focus by neurosurgical procedure. Considerable progress has been made in our ability to calculate, from simple scalp recorded segments containing epileptic spikes, where these signals arise. Scherg has been a leader in the development of brain electrical source analysis or besa. It involves calculation of a source assuming a multi-sphere brain model. Other techniques (using boundary or finite element analyses) such as that pioneered by Fuchs use MRI constructed realistic head models. Multisphere calculations permit better separation of multiple epileptic sources, whereas, realistic head models allow for better representation of results with the patient’s own brain structure. This technology is rapidly improving and it is likely to shown increasing use and value in the combined neurological and neurosurgical investigation of epileptic patients.


Post Trauma Treatment Associates: Advanced NeuroTherapy is BCIA Certified #1408 We use the Lexicore medical technology to conduct QEEG.

Duffy FH, Hughes JR , Miranda F, Bernad P, Cook P. (1994) The status of Quantitative EEG (qEEG) in Clinical Practice. Clinical Electroencephalography, 25, 6-22.

Hoffman DA, Stockdale S. (1995) Neurofeedback in the treatment of mild closed head injury. Paper presented at the 3rd Annual Meeting of the Society for the Study of Neuronal Regulation.

Lubar JO, Lubar JF. Electroencephalographic biofeedback of SMR and beta for treatment of attention deficit disorders in a clinical setting. Biofeedback and Self-Regulation, 1984, 9, 1-23.

Mann CA, Lubar JF, Simmerman AW, Miller CA, Muenchen RA. (1992) Quantitative analysis of EEG in boys with attention deficit hyperactivity disorder: Controlled study with clinical Implications. Pediatric Neurology, 8, 30-36.

Nledermeyer E, Da Siiva FL. Electroencephalography. (1994), 3rd Edition, Williams & Wilkins, Baltimore, 608-610.

Nuwer MR, Jordan SE, Ahn SS. Evaluation of stroke using EEG frequency analysis and topographic mapping. Neurology, 1987, 37, 153-1159.

Oken BS, Chiappa KH, Sallnsky M. (1989) Computerized EEG frequency analysis: Sensitivity and specificity in patients with focal lesions. Neurology, 72, 16-30.

Packard, RC & Ham, LP. (1994) Promising Techniques in the Assessment of Mild Head Injury. Seminars in Neurology, 14(1), 74-79.

Thatcher, R.W., Walker, R.A. and Guidice, S. (1987) Human cerebral hemispheres develop at different rates and ages. Science, 236:1110-1113.

Thatcher, R.W., Walker, R.A., Gerson, I. and Geisler, F. (1989) EEG discriminant analyses of mild head trauma. EEG and Clinical Neurophysiology, 73: 93-106.

Thatcher, R.W. (1991) Maturation of the human frontal lobes: Physiological evidence for staging. Developmental Neuropsychology, 7(3):370-394.

Thatcher, RW, Cantor DS, McAlaster A, Geisier F, Krause P. (1991)Comprehensive predictions of outcome in closed head injured patients: The development of prognostic equations. Annals of the New York Academy of Sciences, 620, 82-101.

Thatcher, R.W. (1992) Cyclic cortical reorganization during early childhood. Brain and Cognition, 20: 24-50.

Silver JM, Yudofsky. SC & Hales RE (Eds.) Neuropsychlatry of Traumatic Brain Injury, American Psychiatric Press, Washington, D.C., pp. 119-122, 718-719. A thorough, excellent book. Sections on diagnostics, types of symptoms, conventional treatments.

Sterman MG, MacDonald LR. (1978) Effects of central cortical EEG feedback training on seizure incidence in poorly controlled epileptics. Epilepsla, 159, 207-222.