EEG in Dementia and Encephalopathy 

EEG has been employed clinically for some time as a measure of brain function in the hope of determining and differentiating certain functional conditions of the brain. It is used in patients with cognitive dysfunction, either a general decline of overall brain function or a localized or lateralized deficit. This article primarily addresses the clinical use of EEG in evaluation of dementias and encephalopathies. In addition, aspects of digital EEG and other newer developments are discussed briefly at the end of the article.

Definition of dementia

Criteria from Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) should be used in the diagnosis of dementia. Clinical dementia is a fairly broad-based decline of brain function; most definitions center on the patient's intellectual decline and memory dysfunction. This is, however, a fairly simplistic approach; dementia encompasses much more than these fundamental deficits. Many of the dementias have distinguishing features. The process that constitutes normal aging is still an ongoing debate. As our understanding and testing procedures develop, more people are being classified as suffering from some type of dementia.

In 1998, Widagdo et al performed a quantitative EEG (QEEG) study of age-related changes during cognitive tasks.^[1] This study revealed no conclusive differences between the young and the elderly. Cognitive decline, unlike normal aging, is associated with alterations in the temporospatial characteristics of EEG. The diagnosis of the initial stages of dementia is based mainly on neuropsychological testing and clinical suspicion. The EEG findings are nonspecific, as shown below).

EEG findings in dementia

In early dementia, the resting alpha frequency declines. Most authors agree that the lower limit of normal alpha frequency is 8 cycles per second. Medications can slow the posterior dominant rhythm; therefore, medication effect should always be excluded. In assessing the frequency of the alpha rhythm, alerting maneuvers are essential in order to ensure that the patient is in the best awake state and not drowsy. Computerized methods, such as EEG spectral analysis, coherence, and complexity (ie, correlation dimension), have been demonstrated to correspond to cognitive function.^[2]

Stevens et al recorded EEGs during 2 resting conditions (eyes closed and eyes opened) and 2 tasks (mental arithmetic and a lexical decision). The goal of the study was to evaluate which temporal and spatial EEG descriptors change with cognitive decline and normal aging. The EEGs were analyzed by using EEG microstates. The primary findings were a significant increase in the number of ultrashort EEG microstates and a reduction in the average duration of EEG microstates in cognitively impaired and demented patients. Cognitive impairment was associated with a reduction or loss of EEG reactivity. In contrast, no alterations in temporal or spatial EEG descriptors were found in normal aging. Cognitive tasks did not add to the information already obtained during the resting states. The reduction in EEG microstate duration correlated with loss of cognitive function.^[3]

Therefore, temporospatial analysis of the EEG record is a useful indicator of cortical dysfunction in dementia and correlates with degree of cognitive impairment. Apparently, temporospatial analysis may be useful in distinguishing patients with dementia from those experiencing normal aging. These data are largely preliminary; whether they contribute additional information to the clinical data in evaluating dementia is unclear.

Definition of encephalopathy

Encephalopathy represents a brain state in which normal functioning of the brain is disturbed temporarily or permanently. Encephalopathy encompasses a number of conditions that lead to cognitive dysfunction. Some of these conditions are multifactorial and some have an established cause, such as hepatic or uremic encephalopathy. Because the EEG patterns in most dementias and encephalopathies demonstrate few specific features, they are discussed together. Some notable exceptions include Creutzfeldt-Jakob disease (CJD) and subacute sclerosing panencephalitis (SSPE); however, no specific patterns exist for most dementias and encephalopathies. Other conditions, such as hepatic and renal encephalopathies, carry distinguishing features; nevertheless, similar patterns may be seen in a fairly wide range of illnesses under certain conditions.

EEG findings in encephalopathy

In general, the most prominent feature of the EEG record in encephalopathies (if there is a change) is slowing of the normal background frequency. Over the course of the disease if serial EEGs are performed, disorganization of the record may develop gradually. Reactivity to photic or other type of external stimulation may be altered. If a QEEG is performed, it may show a frequency shift or decreased interhemispheric coherence of background frequencies. Some conditions are associated with an increase in seizure frequency, and in such cases, epileptic activity may be recorded.^[4]

In a given context, the EEG can play a clinically useful role, especially since functional MRI, positron emission tomography (PET), and single-photon emission computed tomography (SPECT) are either still in an experimental stage or require special settings not widely available.

Use of digital EEG data

Although in the following sections digital EEG data are cited frequently, these data represent primarily digital analysis of clinical EEG recording. The referenced data are presumed to be based on an EEG recording that is read by a clinician; presently, it is recorded by using computerized technology for ease and also for availability for further analysis. A variety of mathematical transforms are available after the initial clinical interpretation—for example, coherence, Fourier transform, wavelets, and microstates (see Digital EEG). These allow for further comparisons with norms and control groups but should be interpreted in conjunction with the primary EEG reading.

Alzheimer disease

EEG is the only clinical diagnostic instrument directly reflecting cortical neuronal functioning. Although the EEG may be normal or minimally disturbed in a number of patients in the initial stages of Alzheimer disease (AD), an abnormal EEG usually is recorded later in the course. A large percentage of patients with moderately severe to severe AD exhibit abnormal EEGs.

In 1981, Stigsby reported diffuse increases of delta and theta frequencies, as well as decreases in the alpha and beta frequency ranges in AD. Frontal slowing was also seen with slowing more prominent anterior to the sylvian fissure, while the blood flow was more decreased posterior to the sylvian fissure. These findings may be explained by the fact that the EEG reflects the functional decline of the anterior structures, while the flow decrease correlates more with the structural damage of the parietal lobe. The frontal slowing probably reflects the loss of functioning of the frontal cholinergic system.^[5]

Wada et al showed that EEG coherence provides a measure of functional correlation between 2 EEG signals. They examined intrahemispheric EEG coherence at rest and during photic stimulation in 10 patients with dementia of the Alzheimer type. In the resting EEG, patients with AD had significantly lower coherence than gender and age-matched healthy control subjects in the alpha-1, alpha-2, and beta-1 frequency bands. EEG analysis during photic stimulation demonstrated that the patients had significantly lower coherence irrespective of the stimulus frequency. The changes in coherence from the resting state to the stimulus condition showed significant group differences in the region of the brain primarily involved in visual functioning. The patients had significantly lower coherence of their EEG reactivity to photic stimulation at 5 and 15 Hz over the posterior head regions.^[6]

These findings suggest that patients with AD may have an impairment of interhemispheric functional connectivity in both nonstimulus and stimulus conditions. This suggests a failure of normal stimulation-related brain activation in AD. Jelic et al found a positive correlation between levels of tau protein in the cerebrospinal fluid (CSF) and EEG alpha/delta ratio. In a subgroup with high CSF tau levels, a strong relationship between EEG alpha/theta and alpha/delta power was found. No such correlation was found in healthy controls and mildly cognitively impaired individuals with elevated CSF tau levels.^[7]

Locatelli et al used EEG coherence to evaluate the functionality of cortical connections and to get information about the synchronization of the regional cortical activity. They studied EEG coherence in patients with suspected AD. Alpha coherence was decreased significantly in 6 patients. Significant delta coherence increase was found in a few patients between frontal and posterior regions. The group with AD demonstrated a significant decrease of alpha-band coherence in the temporo-parieto-occipital areas. This was expressed to a greater extent in patients with more severe cognitive impairment. They theorized that these abnormalities could reflect 2 different pathophysiological changes: (1) the alpha coherence decrease could be related to alterations in corticocortical connections, whereas (2) the delta coherence increase suggests lack of influence of subcortical cholinergic structures on cortical electrical activity.

Strik et al studied EEG microstates in AD. The microstates of the resting EEG of patients presenting with mild or moderately severe dementia of the Alzheimer type demonstrated a significant anteriorization of the microstate fields, and the duration of sustained microstates was reduced. These differences were more important than the diffuse slowing. The measurements of microstates may be useful in the early diagnosis of AD.^[8] Muller et al conducted a study comparing SPECT and QEEG. They concluded that QEEG may be as useful as SPECT brain scanning in staging the disease; however, the correlation with clinical status was weak.^[9]

Akrofi et al used an automated coherence-based pattern recognition system utilizing multiple discriminant analysis (MDA) and k-means clustering coherence features from EEG obtained from 56 subjects. Through statistical analysis, they were able to distinguish the patients with AD from the patients with mild cognitive impairment (MCI) and age-matched controls. This may indicate that patients with AD may have a greater number of damaged cortical neurons than patients with MCI and that MCI may be an intermediate step in the development of AD.^[10]

Siennicki-Lantz et al studied the relation of cerebral white matter lesions to AD.^[11] Cerebral blood flow (CBF) in white matter correlated with systolic blood pressure and multichannel EEG in senile dementia of the Alzheimer type. The presence and functional significance of white matter lesions in the aging brain or in dementia and their relation to blood pressure is an unsettled issue. White matter lesions occur in both cerebrovascular disease and AD. Probably, the white matter lesions in hypertensive patients are not related to but simply are coexistent with the AD. Their influence on overall expression of the degree of dementia is unclear; intuitively, however, the lesions should be causing additional cognitive dysfunction.

They observed significantly lower CBF in the white matter (WMCBF) in patients with AD than in controls. This was more obvious in the posterior cerebral region (ie, parieto-temporo-occipital area). QEEG from the posterior cerebral regions correlated with WMCBF. Systolic blood pressure was significantly lower in the AD group and was correlated positively with WMCBF in the posterior and anterior brain regions.

Epileptiform activity may occur more frequently in patients with AD than in the general population; clinical tonic-clonic seizures can occur. Bilateral synchronous periodic epileptiform discharges (BiPEDs) (see the first image below), such as triphasic waves (TWs) (see the second image below), may be recorded in AD, usually in the late stages (for more information on TWs, see Triphasic Waveforms).

These findings are not specific for AD because they most often are observed in metabolic disorders, particularly hepatic encephalopathy and other degenerative diseases, such as CJD. While good correlation exists between severity of EEG abnormalities and cognitive impairment, epileptiform discharges or TWs are not predictive factors for seizures. EEG often can be useful in evaluating dementia in order to exclude a superimposed reversible metabolic etiology, and to confirm CJD when the dementia is rapidly progressive.

To investigate the relationship between QEEG band powers and CBF, Rodriguez et al studied 42 patients with suspected AD and 18 healthy controls who were elderly. They tried to differentiate patients with AD from the controls by QEEG and CBF measurements. Regional CBF and QEEG were correlated with one another, especially in the right hemisphere. Significant correlations were found between Mini Mental State Examination (MMSE) scores and relative power of the 2- to 6-Hz and the 6.5- to 12-Hz bands on either side and between MMSE scores and left regional CBF, while the correlation between MMSE scores and right regional CBF was less strong.^[12]

Used together, QEEG and regional CBF sensitivity was 88% and specificity 89%, with a total accuracy of 88.3%. QEEG alone showed an accuracy of 77% in the whole group and of 69% in those with mild AD, and regional CBF alone an accuracy of 75% in the whole group and of 71% in those with mild AD. This study suggests that QEEG and regional CBF measurements used together are reasonably accurate in differentiating AD from healthy aging.

Scheriter et al used clinical examinations, QEEG, neuropsychological testing and neuroimaging to see if distinctions could be made between patients with AD, mixed dementia (vascular), and controls. As would be expected, patients with mixed dementias had more subcortical lesions with increased slow frequency power, suggesting subcortical pathology. The QEEG high-frequency power was normal in mixed dementia and decreased in AD, likely reflecting the cortical pathology seen in AD. Hachinski scores and neuropsychological testing showed little difference between mixed dementia and AD. QEEG and neuroimaging may be of great use in diagnosing and differentiating these dementia types.^[13]

ApoE sigma-4 allele is a risk factor for late-onset AD and is proposed to have an impact on cholinergic function in AD. Because the cholinergic system has an important role in modulating EEG, an impairment of the cholinergic system may have a relation to the EEG slowing that is characteristic of AD progression.

Lehtovirta et al studied the relation of apolipoprotein E (ApoE) to EEG changes. The QEEG of 31 patients with AD was recorded at the early stage of the disease and after a 3-year follow-up. Patients with AD were divided into several subgroups according to the ApoE sigma-4 allele (ie, 2 sigma-4, 1 sigma-4, and 0 sigma-4). These subgroups did not differ in clinical severity or duration of dementia. The AD patients carrying the sigma-4 allele had more pronounced slow-wave activity than AD patients without the sigma-4 allele, although the disease progression rate did not change. These differences in EEG may suggest differences in the degree of the cholinergic deficit in these subgroups.^{[14, 15]}

The typical electrophysiological correlates of myoclonus in AD are similar to those of cortical reflex myoclonus, with a focal, contralateral negativity in the EEG preceding the myoclonic jerk. The electrophysiological correlate of polymyoclonus that can be seen in AD and other pathological states is a bifrontal negativity in the EEG that precedes the myoclonic jerk. This new type of electrophysiological correlate of myoclonus may reflect activity of a subcortical generator.

EEG findings in dementia with Lewy bodies and Alzheimer disease

Briel et al found that 17 of the total 19 records from the patients with dementia with Lewy bodies (DLB) were abnormal.^[16] Thirteen showed loss of alpha activity as the dominant rhythm, and half had slow wave transient activity in the temporal lobe areas. This slow wave transient activity correlated with a clinical history of loss of consciousness. The patients with Alzheimer disease were less likely to show transient slow waves and tended to have less marked slowing of dominant rhythm.

The greater slowing of the EEG in DLB than in Alzheimer disease may be related to a greater loss of choline acetyltransferase found in DLB. Temporal slow wave transients may be a useful diagnostic feature in DLB and may help to explain the transient disturbance of consciousness, which is characteristic of dementia with Lewy bodies.

Pick disease

Pick disease, which is a frontotemporal dementia, is much less common than AD. The age of onset is earlier than that of AD. The EEG is less abnormal than in AD, especially in the early stages. Posterior alpha rhythm is more preserved. Theta and delta are increased. Frequency analysis may demonstrate a difference at a time when simple visual reading may not pick up a clear abnormality. The major feature of Pick disease is a decline in judgment and insight with relative early preservation of memory. Because EEG correlates poorly with the clinical symptoms, impressive EEG changes are not observed in this condition. Blood flow measurements correlate with thinking processes; Ingvar demonstrated these changes in 1977.^[17] Stigsby demonstrated a decrease in anterior blood flow in patients with Pick disease.^[5] Because the anterior cholinergic system is relatively preserved in Pick disease, the EEG changes are not prominent frontally.

Gemignani et al studied sleep in Pick disease with a longitudinal polysomnographic and 18-FDG-PET study.^[18] The study showed sleep fragmentation, short REM latency, and a severe reduction of slow wave sleep, with relatively preserved NREM-REM sleep cycles. PET scan revealed severe cerebral glucose metabolic reductions in the frontal and temporal areas. Postmortem study showed severe neuronal loss, spongiosis, and gliosis most marked in cortical layers I, II, V, and VI. In vivo, neurometabolic and postmortem neuropathological data are consistent with and indicative of a severe dysfunction of intra- and trans-hemispheric regional connectivity and of corticothalamic circuits. These findings suggest that the decreased cortical and subcortical connectivity may have been the main pathophysiological mechanism responsible for delta sleep reduction and the cognitive decline.

Huntington disease

Huntington disease is a genetic condition characterized by movement disorder (primarily chorea), cognitive impairment, and psychotic features. The degree of such symptoms varies widely. The EEG changes show gradual and progressive slowing over time. The amplitude also attenuates as the disease progresses. About 30% of the patients have very-low-voltage EEGs with amplitudes below 10 microvolts. Hyperventilation as a rule does not increase the background voltage as it usually does in healthy subjects. About 3% of the patients show epileptiform activity; they tend to be juvenile cases. The EEG has not been proven to be of any predictive value in identifying future affected family members. Genetic testing is far more useful.

Progressive supranuclear palsy

Progressive supranuclear palsy (PSP) causes decreased ocular motility, rigidity, dementia, impaired postural reflexes, and histologically, midbrain atrophy and abnormal tau deposition. Usually, the degree of dementia is not severe. The EEG in PSP may initially be normal but eventually shows increasing delta and theta activity as was the most common finding by Fowler and Harrison in 1986. They found that the delta often was rhythmic with frontal accentuation. Gross et al found a decrease in background frequency of 6-7/s and delta activity over the temporal regions.^[19] Su and Goldstein et al found initial EEG patterns to be normal in 8 of 12 (67%) of patients with PSP.^[20] With disease progression, they found background slowing and frontal intermittent rhythmic delta activity (FIRDA) (see below) in this population.

Through the use of QEEG recordings in 6 patients with PSP as compared with controls, Montplaisir et al found slowing over the frontal lobes in the waking state with neuropsychological testing confirming this frontal lobe dysfunction. Abnormalities of sleep architecture with REM sleep abnormalities were seen as well.

Corticobasal degeneration

Corticobasal degeneration (CBD) is a neurodegenerative disorder and tauopathy with progressive dementia and asymmetrical rigidity and limb apraxia. Tashiro et al found prominent focal slowing on EEG in the anterior and temporal head regions in early CBD in 8 of 10 patients studied.^[21] Frontal intermittent rhythmic delta activity was also observed but was not found to be specific to CBD. Roche et al evaluated 5 patients with CBD with none having REM behavioral disorder (as is often seen in many neurodegenerative disorders) or excessive daytime sleepiness. All 5 patients with CBD had insomnia, 4 had periodic limb movements and/or restless legs syndrome, and 2 had sleep respiratory disorders.^[22]

Parkinson disease

The EEG is frequently normal. In advanced cases, however, marked slowing is noted. Sleep may be markedly abnormal with frequent awakenings, prolonged sleep latency, reduced REM sleep,

periodic leg movements and increased frequency of REM behavioral disorder.

Wszolek et al studied patients with rapidly progressive familial parkinsonism and dementia with pallidopontonigral degeneration (PPND).^[23] The patients had PPND linked to chromosome 17q21-22; 11 EEGs of 9 patients were studied. EEGs revealed normal findings early in the disease and diffuse slowing that became more prominent with disease progression.

Serizawa et al compared patients with Parkinson disease (PD) with age-adjusted controls using QEEG and also found diffuse slowing in the patients with PD.^[24] Electromyograms (EMGs) and nerve conduction studies (NCSs) showed no abnormalities. Visual evoked potentials (VEPs) and sensory evoked potentials (SEPs) were normal. The clinical neurophysiologic study findings were consistent with a cortical and subcortical degenerative process.

With clinical deterioration, progressive decline is seen in the mean parietal frequency and background rhythms. Theta and theta-delta mixture may be recorded bilaterally in the posterior head regions. After stereotactic surgery, focal theta or delta slowing may be observed.

Korsakov syndrome

Obraztsova et al studied 32 patients (21 with reversible and 11 with chronic Korsakov syndrome of traumatic origin) and 20 healthy controls. They found that EEG beta activity (13-20 Hz) in the frontobasal and brainstem locations had negative prognostic significance in Korsakov syndrome. Most typically, patients with Korsakov syndrome have abnormal EEGs with slowing in the theta and delta frequencies.^[25]

Binswanger Disease

Binswanger disease usually demonstrates slowing of background and a nonspecific pattern; however, Kuroda et al reported some other patterns. They described a 72-year-old patient with von Recklinghausen disease exhibiting akinetic mutism within 6 months of the onset of dementia. The EEG demonstrated periodic synchronous discharges (PSDs) suggesting CJD. The CT brain scan findings represented diffuse cerebral atrophy. Autopsy findings revealed diffuse subcortical white matter disease and marked arteriosclerotic changes of the subcortical arterioles.^[26]

The cortex was relatively spared, and the pathologic diagnosis confirmed Binswanger disease. Binswanger disease, therefore, can present with PSD and should be included in the differential diagnosis of dementia. On the other hand, Dzialek et al described a group of 15 patients with Binswanger subcortical atherosclerotic encephalopathy who showed a different EEG appearance. The EEG records were pathological in most cases, with varying degrees of slow activity that was distributed symmetrically.^[27]

Circulatory Encephalopathy

Atherosclerosis

Plachinda et al studied the correlations of cognitive disorders and the EEGs of elderly patients with circulatory encephalopathy. They explored the possibilities of using EEG for evaluating intellectual-mnemonic disorders in elderly patients with cerebral atherosclerosis. Ninety-five patients (aged 60-74 y) with atherosclerotic encephalopathy but without stroke were included in the study. Statistical analysis of the data demonstrated a correlation between psychological test results and EEG readings and computerized EEG data.^[28]

In cerebrovascular disease, focal slowing is far more frequent than in nonvascular dementia; therefore, EEG can be useful in distinguishing the 2 conditions.

Multi-infarct dementia

No specific EEG pattern is associated with multi-infarct dementia. Some background slowing may be observed, especially in advanced disease. These changes are less prominent and do not show the progressive course observed in AD. Iznak et al used QEEG to reveal the specific features of and study amplitude-frequency parameters in patients with mild dementia of different origins compared with healthy elderly individuals. They found that alpha rhythm was suppressed in AD and vascular dementia and that alpha rhythm was slower and theta activity higher in AD. Patients with AD were characterized by desynchronized EEG.^[29]

Transient Global Amnesia

A variety of records have been reported from normal to even epileptiform potentials in transient global amnesia (TGA). Nonepileptiform activity, such as bitemporal delta or bioccipital theta, has been reported. Kushner described patients with normal activity, one with occasional epileptic activity, and one with asymmetric alpha depression, while 2 patients had intermittent rhythmic slowing.^[30] TGA caused by a seizure is uncommon, and is believed to be caused by a vascular etiology or spreading depression.

Action myoclonus

Action myoclonus consists of arrhythmic muscular jerking induced by voluntary movement. It can be made worse by attempts at precise or coordinated movement (ie, intention myoclonus) and may be elicited by sensory stimuli. The effective stimulus for action myoclonus is thought to be feedback from muscle afferents, although it may be related to activity in the motor cortex relayed to the reticular formation preceding or coinciding with voluntary movement. The condition usually is associated with diffuse neuronal diseases, such as posthypoxic encephalopathy, uremia, and the various forms of peripheral neuroepithelioma, although action myoclonus may be limited to one limb in some cases of focal cerebral damage. It is caused by hyperexcitability of the sensorimotor cortex (ie, cortical reflex myoclonus) or reticular formation (ie, reticular reflex myoclonus), or both.

Autopsied cases have failed to reveal a clear pathology. Theories include loss of inhibitory mechanisms involving serotonin and possibly GABA transmitters. Myoclonus may be seen in degenerative disorders of the nervous system. It may be associated with tonic-clonic seizures or dementia. Myoclonus has been described in cases with Lafora inclusion bodies and cerebral storage diseases, as well as system degenerations: cerebellodentatorubral, pyramidal, extrapyramidal, optic, auditory, posterior columns and gracile and cuneate nuclei, spinocerebellar pathways, motor neurons of cranial nerves and anterior horns, and muscle fibers.

Action myoclonus usually responds to sodium valproate or clonazepam, and some patients with posthypoxic action myoclonus may improve with serotonin precursors.

Generalized myoclonus

Generalized myoclonus in comatose survivors after CPR implies a poor prognosis despite improvement of the critical care of patients.^[31] This transient, self-limited phenomenon reflects dysfunction of lethally damaged neurons. Propofol often controls myoclonus but does not change the underlying condition. Clinical, pathological, and EEG findings indicate that these patients die from severe hypoxic-ischemic damage. Abnormal EEG patterns, especially burst suppression EEG (BS-EEG) and alpha-coma-EEG, are seen in these patients. The EEG abnormalities include BS-EEG, generalized epileptiform discharges, alpha-coma-EEG.^[32]

Unverricht-Lundborg disease

Unverricht-Lundborg disease (ULD) (ie, Baltic myoclonus) is an autosomal recessive progressive myoclonic epilepsy syndrome. ULD is found sporadically worldwide, but is common in Finland. The myoclonus is severe and generalized seizures occur that are difficult to control. Progressive background slowing, generalized spike and wave and polyspike and wave complexes, and focal occipital spikes are found on EEG.^[33]

Mitochondrial encephalopathy with lactic acidosis and stroke (MELAS) and myoclonus, epilepsy with ragged red fibers (MERRF)

Isozumi et al described a 50-year-old woman with subacute dementia and myoclonus whose CT scan revealed brain atrophy and EEG revealed PSDs. She initially was thought to be suffering from CJD but dramatically recovered over 5 months. Based on further investigations, the final diagnosis was mitochondrial encephalomyopathy. In general, the EEG changes were described as background slowing, multifocal epileptiform discharges, and photosensitivity.

Poststereotactic surgery

Patients developed EEG slowing of different degrees about 50% of the time.

Alpers' disease

This clinicopathological entity, consisting of progressive neuronal degeneration (ie, Alpers disease) of childhood with liver disease, has been studied by Boyd et al.^[34] The onset is in early childhood and consists of intractable fits, progressive dementia, and brain atrophy. EEG studies have been carried out on 12 children with this condition. The EEGs were similar and demonstrated abnormal patterns with high-amplitude slow activity as well as lower amplitude polyspikes. The flash VEP was usually abnormal and often asymmetrical. In the appropriate clinical setting, the neurophysiologic features may aid the clinician in the diagnosis of this autosomal recessive inherited disorder.

Adrenoleukodystrophy

Multifocal paroxysmal discharges, hypsarrhythmic pattern, and prominent arrhythmic delta are present in temporo-occipital areas. Epileptic discharges usually do not occur in adrenoleukodystrophy.

Zellweger syndrome

This is characterized by diffuse slowing.

Infantile neuroaxonal dystrophy

This condition is characterized by a high-voltage, 14- to 22-Hz activity that is not reactive to environmental stimuli.

Pantothenate kinase-associated neurodegeneration (PKAN)

In PKAN, formerly known as neurodegeneration with brain iron accumulation type I, the EEG is normal to slow.

Neuronal ceroid lipofuscinosis

In the infantile form, the EEG is slow and early, and posterior spikes may be present. Photic response is excessive and evokes high-voltage spikes that are polyphasic. The EEG abnormalities in the juvenile form are not as marked.

Gaucher disease

In patients with type III disease, posterior spikes and sharp waves, diffuse spike and waves, and photomyoclonic and photoparoxysmal responses may be present.

Metachromatic leukodystrophy

Diffuse slowing progresses to high-voltage generalized delta activity. Epileptic activity is rare; however, hypsarrhythmia may be observed.

Tay-Sachs disease

EEG is generally slow. Generalized or multifocal spikes accompany the seizures.

Rett syndrome

This is a progressive encephalopathy observed in girls. Al-Mateen et al reported 15 cases of Rett syndrome.^[35] The course is slowly progressive; it occurs only in girls and is characterized by early deterioration of higher brain function with dementia, autistic behavior, loss of purposeful use of the hands, and deceleration of head growth. When affected girls are aged 2-4 years, epilepsy may develop with minor motor seizures. Additional features may include an extrapyramidal disorder with dystonia and choreoathetosis, and lactic acidemia. A precise biochemical marker of this disorder has not been identified.

According to McIntosh et al, Rett syndrome consists of a progressive encephalopathy and psychomotor deterioration in young girls who have appeared clinically normal until age 6-18 months.^[36] The incidence is similar to phenylketonuria and autism in females. When the child is at least 6 months old, head growth decelerates in association with severe dementia, autism, apraxia, stereotypic "hand washing" movements, and loss of previously acquired skills. Other signs include breathing dysfunction, seizures, EEG abnormalities, and growth retardation. It appears to be sporadic in occurrence.

The EEG may demonstrate slowing, a variety of nonspecific patterns, and epileptiform discharges. The epileptic activity may include multifocal spikes, slow-wave spikes, and paroxysmal delta slowing with spikes that may appear in sleep; in certain cases, however, sleep may attenuate the EEG abnormalities. Background flattening occurs to some degree, corresponding with the stage of dementia and cognitive decline. Rolandic spikes may be elicited by noise.

Creutzfeldt-Jakob disease

The EEG shows a fairly typical repetitive pattern of BiPEDs (see first image below) such as TWs (see second image below) approximately 1-1.5 seconds apart. These usually are present during wakefulness and disappear during sleep.

PSDs seem to be the EEG hallmark of Creutzfeldt-Jakob disease; however, a number of atypical EEG presentations have been reported without these waveforms. Aoki et al reported giant spikes with photic stimulation.^[37] These photic-stimulated giant spikes simultaneously suppressed PSDs. Necropsy exhibited extensive gray and white matter lesions. Both lateral geniculate bodies and pregeniculate bodies were involved preferentially. The superior colliculus, optic nerve, and optic tracts were not affected. The cortices of the occipital lobes were damaged severely. The Gennari line was spared. The lesion of the lateral geniculate body appeared to be associated with the unusual EEG feature.

Their findings indicate that the visual pathway may be involved in the generation of PSD in CJD (see below).

The EEG findings and the evolution of clinical signs were investigated by Hansen et al in 7 patients with CJD who underwent serial EEG recordings. At the onset (mean 8.7 weeks) of periodic slow-wave complexes (PSWC), 5 patients already had progressed to akinetic mutism characterized by loss of verbal contact and movement disorders (ie, myoclonus, exaggerated startle reaction, or focal dyskinesia started in 5 patients). When akinetic mutism commenced (average 7.5 weeks), runs of frontal intermittent rhythmic delta activity (FIRDA), like that shown below, were found in all cases. These were later replaced by PSWC in 6 patients. Occurrence of PSWC often related negatively to external stimuli and sedative medication.^[38]

These data help in the selection of EEG recording dates to detect PSWC in patients in whom CJD is suspected. The survival time is short after the onset of PSWC (average 8 weeks). In earlier disease stages, FIRDA-like EEG activities should be regarded as compatible with the diagnosis of CJD and should encourage further EEG studies for the demonstration of PSWC in a more advanced stage of CJD.

EEG characteristics of CJD and its differential diagnosis were studied by Steinhoff et al. They found some nonspecific EEG findings and also typical PSWC in the course of the disease. They obtained a sensitivity of 67% and a specificity of 86%. With the exception of one familial variant of CJD, PSWC are usually absent in all other human prion diseases. They presented a pathophysiologic hypothesis on the development of PSWC based on the assumption that the specific periodicity of PSWC results from a still functionally active but greatly impaired subcortical-cortical circuit of neuronal excitability. They stressed the use of clinical signs, laboratory data, and EEG correlation and suggested that the clinical diagnosis of CJD should be reconsidered if repeated EEG recordings fail to reveal PSWC under technically adequate conditions. Some patients with CJD presented with visual blurring, diplopia, and visual loss—ie, the Heidenhain 5 variant.^[39]

Focal EEG abnormalities as described in the Heidenhain variant of CJD are uncommon. Lee et al reported a 73-year-old man presenting with visual symptoms, right hemianopia, and rapidly progressive dementia. Myoclonus was synchronous with generalized periodic epileptiform discharges on EEG, as shown below. In addition, periodic focal sharp waves were present at the left occipital region. Diffusion-weighted MRI of the brain showed slightly increased signal intensity in the occipital parasagittal area, left more than right. The 14-3-3 protein was detected in the CSF. The patient died within 5 months of presentation.^[40]

Subacute spongiform encephalopathy

Aguglia et al described 20 patients with subacute spongiform encephalopathy and periodic paroxysmal activities in the EEG. Evolution of clinical and EEG abnormalities were analyzed in all 20 (16 pathologically confirmed). Illness duration was less than 4 months in 65% and greater than 17 months in 10%. The early clinical stage was characterized by gradual gait disturbances, mental deterioration, and sensory or autonomic changes. In 10 EEG recordings from 7 patients examined in the early clinical stage, no periodic discharges were present.^[41]

Early periodic paroxysmal activity appeared within 12 weeks of the onset of the disease in 88% of the patients who underwent EEG recordings. This early periodic paroxysmal activity usually occurred at an intermediary stage, when the patients demonstrated marked worsening of the clinical picture. Focal, segmental, and/or generalized myoclonic jerks were observed in 15%, 53%, and 100% of cases at prodromal, intermediary, and terminal stages, respectively. Different kinds of periodic paroxysmal activity were observed: (1) biphasic or triphasic periodic complexes, (2) periodic complexes with multiphasic configuration, and (3) periodic polyspiking discharges. Abnormal "pacing" by slowly repeated flashes was found in 4 patients presenting with visual hallucinations or cortical blindness. Burst-suppression activity was observed frequently in the terminal stage in decorticate patients.

AIDS dementia

EEG abnormalities usually precede brain atrophy on CT brain scan. Generalized or multifocal slowing may be observed. Computerized EEG is abnormal in most cases. About one half of patients who have normal neurologic findings on physical examination exhibit abnormal EEGs.

Thomas et al described a 40-year-old HIV-positive, right-handed homosexual man who was admitted for progressive mental deterioration coexisting with permanent, segmental, middle-amplitude, arrhythmic, asynchronous, and asymmetrical myoclonic jerks. EEG demonstrated frontocentral bursts of rhythmic triphasic 1.5- to 2-Hz sharp waves similar to the characteristic periodic pattern of CJD. Biological investigations were negative, thus ruling out a metabolic encephalopathy.^[42]

Dramatic neurological improvement occurred shortly after initiation of intravenous and then oral zidovudine, which produced absolute EEG normalization. This unusual electroclinical presentation of the AIDS dementia complex underlines the fact that this condition presents a diagnostic challenge, particularly in individuals in whom HIV infection has not been diagnosed previously.

Canafoglia et al described a case of a HIV-seropositive patient with ataxia and upper limb rhythmic myoclonus.^[43] Electromyograph (EMG) recordings of the forearm muscles correlated with frontocentral electroencephalograph (EEG) rhythmic activity. This movement disorder should be considered a rhythmic variant of cortical myoclonus. HIV infection may have caused a dysfunction in the central nervous system pathways similar to that occurring in genetically determined conditions characterized by cortical myoclonus.

Sinha et al described various electrophysiological abnormalities in HIV encephalopathy.^[44]

Polich et al found greater frontal delta power in HIV cases compared with controls.^[45]

Ferrari et al describe 2 patients with HIV type 1 infection who presented new-onset epilepsia partialis continua (EPC) as an early manifestation of progressive multifocal leukoencephalopathy (PML).^[46] PML represents an increasingly recognized cause of new-onset seizures in both seropositive and seronegative patients.

Diehl et al followed 117 HIV patients with electroencephalography. Serial EEGs on 117 HIV patients without any clinical signs of secondary neuromanifestations were studied in order to document electroencephalographic changes in the course of HIV infection. Clinical signs of HIV-associated encephalopathy presented in 18 patients at the first examination and 23 at reexamination. Significant slowing of background activity occurred in the course of the disease. The results of this study indicated progressive CNS dysfunction with worsening of the immunostatus.^[47]

Chronic rubella encephalitis

This condition is characterized by myoclonus, mental deterioration, ataxia, and chorea, with diffuse slowing on EEG. Intermittent rhythmic delta activity (IRDA) has been described. Periodic activity with spikes and slow-wave spikes may occur.

Viral encephalitis

Viral encephalitis frequently causes EEG abnormalities. If the cortical gray matter involvement is predominant, more polymorphic delta activity is observed, while with subcortical involvement, a rhythmic pattern (IRDA) is more common. In herpes simplex encephalitis (HSE) (see first image below), temporal intermittent rhythmic delta slowing (TIRDA, see second image below), nonrhythmic temporal slowing, and frontotemporal slowing are characteristic; a periodic pattern may develop as the disease evolves.

Hsieh et al found abnormal EEG findings and abnormal neuroimaging in three fifths of children (n=26) with HSE aged 1-6 years correlated with poorer outcomes.^[48]

In a retrospective review of EEG in HSE, Al Shekhlee et al found periodic lateralized epileptiform discharges (PLEDs) (see image below) and/or focal temporal slowing present in 90% of the PCR-positive group (PCR testing for the herpes virus from spinal fluid being the most sensitive and specific test for the diagnosis of HSE) at symptom onset compared with 30% of the PCR-negative group. They found that the sensitivity of the EEG recording for these focal and epileptiform findings decreases after 48 hours. The MRI results were consistent with HSE in 86% of those with HSE-positive PCR results obtained 48 hours from symptom onset. They found the EEG to be of important diagnostic use when obtained within the first 24-48 hours of HSE symptom onset.^[49]

Serial EEGs usually capture PLED activity but in the later stages of the disease course, the EEG may revert to normal.

St Louis encephalitis

Wasay et al studied electroencephalograms and magnetic resonance imaging (MRIs) of patients. Of the 9 patients who were examined with electroencephalography, all 9 had seizures or other abnormalities, and 1 had nonconvulsive status epilepticus. The MRI findings in 2 of the 9 patients showed edema. One of the 9 patients had HIV co-infection.^[50]

Subacute sclerosing panencephalitis

SSPE is a progressive, neurodegenerative disorder caused by defective measles virus replication in the brain as a consequence of measles immunization. The EEG may provide an important clue regarding SSPE and demonstrates bilaterally synchronous, high-amplitude spike or slow-wave bursts that often correlate with clinical myoclonus. As SSPE progresses, the background activity becomes suppressed, resulting in a burst-suppression pattern. Neuroimaging studies demonstrate nonspecific abnormalities or diffuse atrophy, although T-2 prolongation can be detected by MRI symmetrically in the cerebral white matter or multifocally in the subcortical white matter or cortex.

Flaherty et al described a 17-year-old boy with SSPE discovered when he presented with confusion after a mild head injury. The EEG strongly suggested the diagnosis. Repeated CT scans of the head were normal. The boy had a 3-year history of decreased vision, associated with a focal pigmentary retinopathy. On assessment, the patient demonstrated visual agnosia and early dementia. MRI scan demonstrated symmetrical demyelination of the white matter, particularly of the occipital lobes. The typical EEG findings and the presence of measles antibodies in the CSF confirmed the diagnosis of SSPE.^[51]

SSPE should be considered in young patients who have persisting cognitive dysfunction that is not proportional to the severity of the initial trauma. A focal pigmentary retinopathy, especially with macular involvement, should raise the possibility of SSPE, even if neurological symptoms are absent initially. The longest interval (till date) between the visual symptoms and onset of neurological signs of SSPE was reported by the author.

Koppel et al reported on the relation of SSPE and HIV. SSPE had largely disappeared from the United States because of nearly universal measles vaccination; however, it has re-emerged in children infected with HIV. Two children with SSPE were described. The first was HIV positive and presented with seizures and encephalopathy at the age of 21 months. The second developed myoclonus and dementia at 4 years of age; she was not infected with HIV, but her mother had AIDS. MRI findings of the brain were nonspecific. EEG was characteristic of SSPE, showing high-voltage PSWCs and background slowing. Brain biopsy and high measles antibody titers in the CSF confirmed the diagnosis of SSPE.^[52]

Metabolic Disorders

Anoxic encephalopathy

Hypoxia causes diffuse slowing in the EEG. The acute and prolonged anoxia of cardiac arrest exhibits no changes initially. In 7-10 seconds, slow waves appear. This is followed by rhythmic, high-voltage delta activity; subsequently, attenuation and EEG flattening occurs. As a rule, irreversible brain damage results in 4-8 minutes. In some cases, establishing the completeness and duration of anoxia is difficult. Certain patterns carry a poor outcome: flat EEG, burst-suppression patterns, and burst suppression patterns with epileptiform discharges (see the image below) nearly always carry a poor prognosis. Postanoxic EEGs may exhibit a variety of abnormal patterns: triphasic activity, alpha coma pattern, repetitive complexes, and bilateral PLEDs.

Takahashi et al reported a 47-year-old man admitted to the hospital for depression, who suddenly developed cardiopulmonary arrest of unknown etiology and entered a chronic vegetative state as a result of anoxic encephalopathy. PSDs were present for as long as 5 months. The wave pattern, periodicity, and duration of appearance of PSDs were similar to those of PSDs seen in CJD. The PSDs were prolonged gradually, with a course similar to that of the discharges observed in CJD. The mechanism of occurrence is considered to be similar to that of PSDs in CJD.^[53]

Fernandez-Torre et al described the clinical and electroencephalographic features of a comatose patient with severe anoxic encephalopathy who experienced acute reflex myoclonus precipitated by passive eye opening/closure and painful stimulation. Acute stimulus-sensitive postanoxic myoclonus is an underdiagnosed epileptic condition. Shortly after the anoxic insult, the diagnosis should be based on EEG evaluation and various types of stimulation. These should include passive eye opening/closure and painful stimuli.^[54]

Comatose intensive care patients

Young et al investigated the usefulness of continuous EEG monitoring. Twenty percent of patients recorded seizures. The study suggests that CEEG monitoring may be more valuable for detection of seizures in patients with acute structural brain lesions (ASBLs) than in patients with metabolic encephalopathies.^[55]

Hyponatremic encephalopathy

Usually, nonspecific slowing is observed in hyponatremic encephalopathy. A variety of other patterns have been described: TWs; burst of high-voltage rhythmic delta; central, high-voltage, 5- to 7-Hz rhythm; and sensory stimulation-induced, high-voltage delta activity. Epileptic activity is very rare, even in cases of clinical seizure. Kameda et al reported a case of FIRDA in the EEG of a patient with pituitary adenoma, hyponatremic encephalopathy, and major depression. The pituitary adenoma is thought to be a major factor for FIRDA in this case. Complicating factors included diffuse encephalopathy and use of antipsychotic drugs; FIRDA remained in the EEG after these factors diminished. The size of the pituitary adenoma that was proposed to be associated with FIRDA in the EEG recording was not noted. FIRDA may be associated with a small pituitary adenoma less than 10 mm in size.^[56]

Hypocalcemia, hypercalcemia

Paresthesias, tetany, muscle spasm and, rarely, seizures may occur in hypocalcemia. EEG findings include theta and polymorphic delta slowing, polyspikes, sharp waves, and paroxysmal activity. Hypercalcemia is associated with renal failure, neoplasms, bone destruction, parathormone releasing tumors, and hypervitaminosis D. Muscle weakness, polydipsia, polyuria, nausea, anorexia, and coma may develop. EEG changes appear when serum calcium level is approximately 13 mg/dL; slowing and intermittent rhythmic delta activity is seen. Photic driving may be prominent, and TWs may be recorded. When serum calcium is normalized, the EEG usually improves but not immediately. A hypercalcemic condition can be observed in association with hyperthyroidism. Confusional state and EEG alterations, among which diffuse monomorphic delta rhythms were remarkable, were shown by Juvarra.^[57] As soon as normalization of calcium serum level was achieved, rapid clinical and EEG improvement ensued.

Endocrine Conditions

Adrenal disease

EEG pattern is nonspecific.

Cushing disease

EEG changes are uncommon.

Addison disease

Nonspecific slowing and diffuse theta and delta may be seen in a disorganized manner.

Pheochromocytoma

No particular EEG pattern has been noted.

Hypoglycemia

The EEG resembles changes described with hypoxia; hyperventilation response is exaggerated and FIRDA may be observed. If prolonged coma ensues, the EEG changes persist and may become permanent. In most cases of hypoglycemia, a generalized disorganization of record occurs; in patients with long-term diabetes, the EEG is usually mildly to moderately diffusely disorganized and slow.

Hyperglycemia

Similar slowing is the rule; however, epileptic activity may be observed with clinical seizure.

Wang et al described hyperglycemia with occipital seizures. They described acute and follow-up VEP and MRI findings of a patient with hyperglycemia-related visual SE of occipital origin. Occipital seizures and hemianopsia can be caused by hyperglycemia and may be accompanied by special MRI and VEP findings.^[58]

Glutaric aciduria type I

Neurophysiologic abnormalities are frequently seen in organic acidemias. Yalnizoglu et al studied electroencephalogram (EEG), visual evoked potential (VEP), and brain-stem auditory evoked response (BAER) in 7 children with glutaric aciduria type I (GA1).^[59]

Three of the 7 patients showed abnormal EEG findings. Two patients showed asymmetry with intermittent occipital delta slowing in one hemisphere. This finding probably indicates underlying cerebral dysfunction and is not a specific feature. One patient showed high amplitude bursts of beta in the occipital regions with left predominance while on clonazepam and baclofen.

Hyperthyroidism

This has a nonspecific pattern, including slowing and FIRDA. Depending on the severity of the thyroid disfunction, seizures and epiletiform discharges can be seen.^[33]

Hypothyroidism

Low-voltage theta is the rule with reduced photic driving response.

Nutritional Deficiency Syndromes

Pyridoxine deficiency causes severe, and at times, fatal convulsions in infants. The underlying metabolic problem has been suggested to be insufficient GABA synthesis. Thiamine deficiency causes diffuse slowing in Wernicke encephalopathy. Malnutrition results in EEG slowing, proportional and corresponding to the clinical alertness of the patient.

Toxic Agents

Aluminum toxicity

Flaten et al reported a wide range of toxic effects of aluminum. This element has been demonstrated in plants and aquatic animals in nature, in experimental animals by several routes of exposure, and under different clinical conditions in humans.^[60] Aluminum toxicity is a major problem in agriculture, affecting perhaps as much as 40% of arable soil in the world. In fresh waters acidified by acid rain, aluminum toxicity has led to fish extinction. Aluminum is a very potent neurotoxin. Subtle neurocognitive and psychomotor effects and EEG abnormalities have been reported at plasma aluminum levels as low as 50 mcg/L.

Infants and patients with impaired renal function could be particularly susceptible to aluminum accumulation and toxicity. Evidence exists to suggest that aluminum may be the causative agent in the development of dementia in patients with chronic renal failure who are on dialysis (ie, dialysis dementia). The EEG may become abnormal months before the full-blown dementia develops. Aluminum also is associated with dialysis encephalopathy, which often is accompanied by osteomalacia and anemia. Such effects also have been reported in certain patient groups without renal failure.

Aluminum accumulation occurs in the tissues of workers with long-term occupational exposure to aluminum dusts or fumes. Such exposure may cause neurological effects.

In dialysis dementia, the EEG abnormalities usually are diffuse slowing, although TWs may occur. When seizure develops, high-voltage spike and slow-wave complexes and paroxysmal bursts with a frequency of 2-4 Hz have been observed. Polymorphic frontally dominant delta often is observed. The background slowing usually correlates with severity of mental status impairment. Subcortical dysfunction may be present with FIRDA.

Carmofur

Carmofur, an antineoplastic derivative of 5-fluorouracil, has been reported to cause subacute leukoencephalopathy. Kuzuhara described 3 individuals who developed subacute leukoencephalopathy after carmofur (l-hexylcarbamoyl-5-fluorouracil) administration. Initial symptoms were unsteady gait and dementia, developing several weeks or months after administration of carmofur. Symptoms increased gradually even after stopping the drug. Severe encephalopathy with confusion, delirium, or coma developed. Symptoms were usually reversible but occasionally resulted in death. The EEG demonstrated marked slowing. CT scan of the brains of severely intoxicated patients showed marked hypodensity of the entire cerebral white matter.^[61] Carmofur must be discontinued immediately if any psychomotor symptoms develop.

Cefepime

Status epilepticus and encephalopathy have been reported with use of cephalosporins in patients with renal failure. Maganti et al reported the case of a 79-year-old patient with normal renal function who developed subtle mental status changes during cefepime therapy for urinary tract infection. EEG showed nonconvulsive status epilepticus (NCSE).^[62]

Lead

The heavy metal lead is usually absorbed into the body by ingestion and/or inhalation. Symptoms from acute poisoning may range from lethargy, seizures, coma, and death to cognitive impairment, delirium, ataxia and distal motor neuropathy with chronic exposure.^{[63, 64]} Stewart et al found lead to be linked to neurodegeneration with cumulative exposure leading to decreased brain volumes and white matter lesions.^[65] Treatment includes chelation therapy and removing the source of the lead exposure from the patient's environment. See eMedicine article Lead Encephalopathy.

Lithium

Encephalopathy, confusional states, and nonconvulsive status epilepticus due to lithium toxicity and overdose is well documented.^[66] In a recent report by Bellesi et al, a patient presented in nonconvulsive status epilepticus with a normal serum lithium level that resolved with benzodiazepine administration and withdrawal of the lithium. Two months later, the lithium was restarted with therapeutic levels attained. The patient again was found to be in nonconvulsive status epilepticus with 3-4 Hz diffuse spike and wave discharges. The lithium was stopped, never to be readministered.^[67]

Manganese encephalopathy

Excess manganese (Mn) can cause several neurotoxic effects. Herrero Hernandez et al described an epileptic syndrome due to manganese intoxication in a 3-year-old boy. The electroencephalogram (EEG) showed progressive encephalopathy. The patient developed epileptic status. Chelating treatment promptly succeeded in reverting epileptic symptoms.^[68]

Neuroleptic encephalopathy

Treatment of psychiatric patients often necessitates overlapping neuroleptic medication. Lambreva et al report a 60-year-old woman suffering from schizoaffective disorder who temporarily received 3 neuroleptics, together with lithium. She developed neurotoxic encephalopathy with symptoms of a malignant neuroleptic syndrome. They recommend frequent electroencephalographic controls for early detection of neurotoxicity.

Tiagabine

Tiagabine hydrochloride (TGB) (an antiepileptic medication) is a selective gamma-aminobutyric acid (GABA) reuptake inhibitor that is used as an add-on therapy for partial seizures. The risk of nonconvulsive status epilepticus can be elevated by TGB (comedication 7.8% vs TGB alone 2.7%).^[69] Kellinghaus et al reported 3 cases of recent increase in TGB doses causing nonconvulsive status epilepticus with the EEG of one of the patients demonstrating rhythmic delta waves.^[70] Vinton et al reported TGB-induced nonconvulsive status epilepticus in 3 patients with focal lesional epilepsies—1 with initiation of the medication and 2 with recent dose escalation. All of these patients had continuous high amplitude and generalized 2-4 Hz delta activity with intermingled spikes seen with episodes of unresponsiveness. After dose reduction or withdrawal of the TGB and administration of IV clonazepam, EEG and clinical signs normalized.^[71]

Valproate and topiramate encephalopathy

Panda et al reported 2 children with encephalopathy and slowing of EEG background activity, which promptly reverted to normal along with clinical improvement following withdrawal of valproate (VPA).^[72]

According to Segura-Bruna et al, valproate-induced hyperammonemic encephalopathy (VHE) is VPA treatment that results in elevated serum ammonium levels, which leads to a decreased level of consciousness, cognitive slowing, vomiting, drowsiness, lethargy, and increased seizure frequency. If VHE is suspected, serum ammonium levels should be evaluated and the existence of a possible urea cycle enzyme deficiency, such as orthnithine carbamoyltransferase deficiency, should be considered. Generalized slowing in the theta and delta frequencies, frontal intermittent delta activity (FIRDA), and triphasic waves can be found on EEG. These findings and other clinical symptoms usually resolve after VPA is withdrawn.^[73]

Cheung et al proposed a novel term, topiramate-valproate induced hyperammonemic encephalopathy to describe the clinical features of patients on concomitant topiramate and valproate therapy. With this specific encephalopathic syndrome those on the above therapy may display excessive sleepiness or somnolence, increased seizure activity, hyperammonemia, and an absence of triphasic waves on EEG.^[74]

Liver Transplantation

The EEG in hepatic encephalopathy may consist of slow waves and TWs; epileptic activity may be observed. Adams et al studied patients after liver transplantation. Seventeen (33%) of 52 patients who underwent 56 consecutive orthotopic liver transplants had serious postoperative neurological complications. Seizures were described in 13 (25%) patients; of these, 50% had onset of seizures within the first week. In 3 patients, the seizures were associated with postoperative metabolic encephalopathy and fatal progressive neurological deterioration. Cyclosporin was thought to be causing the seizures in some of these patients. In others, electrolyte disturbances, steroid treatment for graft rejection, and cerebral infarction could have contributed to the occurrence of seizures.^[75]

Scleroderma

CNS involvement and psychiatric manifestations can occur in systemic sclerosis (ie, scleroderma). Hietaharju et al evaluated CNS and psychiatric involvement in 32 patients. Severe CNS or psychiatric symptoms were present in 5 patients (16%), including encephalopathy, psychosis, anxiety disorder, grand mal seizures, and transient ischemic attacks. In addition, abnormal VEPs were recorded in 5 of 32 patients (16%), suggesting optic neuropathy. EEGs were mainly normal or showed only slight nonspecific changes.^[76]

Hashimoto Myoclonic Encephalopathy

Ghika-Schmid et al reported 2 patients with subacute diffuse encephalopathy characterized by confusion, myoclonic encephalopathy, and mild akineto-rigid extrapyramidal signs in one case and by apathy, memory deficit, and partial complex seizures in the other. Hashimoto thyroiditis with high titers of antithyroglobulin antibodies was diagnosed in both patients, who were not responsive to anticonvulsant medication but exhibited rapid neurological improvement following steroid treatment. On neuropsychological examination, predominant frontotemporal dysfunction was noted.^[77]

EEG activity was remarkable for its rhythmic delta activity, which was unresponsive to, or even paradoxically increased by, anticonvulsant treatment. Atrophy with temporal predominance was observed on MRI. These observations support the idea that this potentially treatable dementia and movement disorder should be classified as a separate clinical entity.

Kothbauer-Margreiter et al reported 6 patients with Hashimoto thyroiditis and associated encephalopathy and compared with 14 well-documented cases identified in the literature.^[78] Encephalopathy typically affects patients when they are euthyroid and in an appropriate clinical situation; antithyroid autoantibodies are the main indicators of the encephalopathy. Since clinical features of Hashimoto encephalopathy are nonspecific, other etiologies such as infectious, metabolic, toxic, vascular, neoplastic, and paraneoplastic causes need to be considered.

Two types of initial clinical presentation can be differentiated: (1) a vasculitic type with strokelike episodes and mild cognitive impairment and (2) a diffuse progressive type with dementia, seizures, psychotic episodes, or altered consciousness. These types may overlap, particularly over the long-term course in untreated patients. A strong female predominance existed in this study; 18 of the 20 patients were women. The EEG was abnormal in 90% of cases; it showed nonspecific changes. The condition is steroid responsive.

Triphasic waves (TWs), like those shown below, were initially described by Foley et al in hepatic encephalopathy. They later were described in other metabolic states and brain tumors.^[79] Most electroencephalographers now agree that TWs are a relatively nonspecific pattern observed in a number of metabolic conditions, degenerative dementias, and anoxia. In a bipolar montage, TWs usually comprise a high-voltage, positive wave followed by a smaller negative deflection; they usually are bilaterally synchronous and maximal frontally. A fronto-occipital (anteroposterior) phase lag varies from 25-140 ms. This is expressed less in referential montages.

TWs have not been reported in children. Generally, the TW pattern carries a poor prognosis with a high mortality rate if it occurs in association with rapid neurological and clinical deterioration. However, TWs in a psychiatric population described by Blatt and Brenner carried a different prognosis. In a large retrospective study consisting of 15,326 EEGs performed from 1983-1992 in a psychiatric institute, 83 EEGs (62 patients—13 men and 49 women aged 59-90 years, with a mean age of 74 years) had TWs. All 62 patients were awake, though they often were confused. Most (n=56) had dementia, usually severe; 15 also had delirium. Six patients had no dementia. Infrequent etiologies included neuroleptic malignant syndrome (n=1) and hepatic encephalopathy (n=1); in 4, the cause was uncertain, although all were receiving lithium.^[80]

EEG features analyzed included frequency of background rhythms, distribution of the TWs, periodicity, and epileptiform abnormalities. Background rhythms were slow in all but 7 patients (mean 6.2 +1-1.7 [SD] Hz). TWs were maximal posteriorly in 47 patients and anteriorly in 6 patients and were diffuse in 9 patients. Neuroimaging studies demonstrated prominent posterior abnormalities in only 1 individual. Periodicity was prominent in 4 patients; in 2, the TWs were maximal anteriorly. Interictal epileptiform activity was present in 6 patients, a history of seizures in 8, and myoclonus in 4. TWs are uncommon in psychiatric populations; they occur primarily in elderly and severely demented patients.

Aguglia et al discussed nonmetabolic causes of TWs and described 2 patients with TWs on their EEGs in the absence of metabolic disturbances. One patient had coma associated with cerebellar hematoma, the other had mild dementia associated with idiopathic calcifications of the basal ganglia and healthy auditory brainstem responses and subcortical and cortical SEPs. Neurologic examination revealed no asterixis in either patient.^[81]

The literature on nonmetabolic causes of TWs also was reviewed, and the clinical and anatomic reports of 10 patients were analyzed. Seven patients had focal brainstem-diencephalic lesions (craniopharyngioma [2], thalamic gliomas [3], pontine stroke [2]). Three patients suffered from diffuse subcortical or multifocal encephalopathies (Binswanger encephalopathy [1], cerebral carcinomatosis [1], multifocal cerebral lymphoma [1]).

From the clinical point of view, patients with nonmetabolic diseases causing TWs presented with either disturbance of higher cerebral functions with no asterixis or sudden onset of coma. Aguglia et al concluded that TWs may result from focal brainstem/diencephalic lesions or from diffuse subcortical or multifocal encephalopathies in the absence of concomitant metabolic abnormalities. Nonmetabolic causes of TW should be suspected in patients presenting with neurological disturbances not associated with asterixis.

TWs also were evaluated by Sundaram et al, and their clinical correlates and morphology were assessed. Twenty-six (41%) of 63 consecutive patients with TWs had various types of metabolic encephalopathies, while 37 patients (59%) had nonmetabolic encephalopathies, usually senile dementia. TWs were not found to be specific for any single type of metabolic encephalopathy. Etiology was linked more closely to level of consciousness at recording than any morphologic or distributional feature of the TWs themselves. Thus, all 31 alert patients had nonmetabolic encephalopathies, while all 13 comatose patients had metabolic encephalopathies.^[82]

The second, positive component (wave II) most often had the highest voltage, while equally maximal waves I and II occurred next most commonly. In these patients, TWs most often were expressed maximally anteriorly. Among patients with metabolic encephalopathies, a posterior-anterior delay or lag of the wave II peak occurred more commonly than did the better known anterior-posterior lag. Lags occurred with both metabolic and nonmetabolic conditions but were more common with the former. No difference in quantity or mode of appearance existed between the metabolic and nonmetabolic groups when matched for consciousness level.

Prognosis for patients with either metabolic or nonmetabolic encephalopathies was unfavorable. Only 4 of 24 patients with metabolic encephalopathy and 1 of 35 patients with nonmetabolic encephalopathy were well at follow-up more than 2 years later. Forty percent of EEGs with sharp and slow-wave complexes (slow spike waves) had sporadically appearing TWs. The relative amplitudes of the 3 components differed from those of TWs in other conditions; equally maximal waves II and III were the most usual form.

As stated in the assessment report of the American Academy of Neurology and the American Clinical Neurophysiology Society, digital EEG is an established substitute for recording, reviewing, and storing a paper EEG record. In this sense, digital EEG simply replaces and improves the paper record in ways similar to the way a word processor has improved letter writing by hand or even by a typewriter. However, routine digital recording of the clinical EEG by digital means does not add new information that was not present in the paper record.

Once the paper recording is made, a number of options become available for further analysis. Some processing methods, such as different montage displays and digital filtering, simply enhance the visibility of the record. Some other methods, such as calculating the mean band frequencies and different band-energy spectra, may bring into the forefront information that was already there in the paper record but is too tedious and time-consuming to calculate without use of a computer. Spike recognition is an important enhancement and a great time saver but needs careful review by the interpreting physician. Lately, technologic developments have enabled the authors to record long-term monitoring on small storage devices, making the diagnosis of syncope, seizures, and sleep disorders much easier.

EEG brain mapping visualizes some selected electrical event in the brain and maps its geographic distribution. Attempts have been made to standardize some aspects of brain mapping; however, no clear uniform recommendation has yet emerged. Although frequency bands are fairly well standardized, different ways to calculate the data exist. Normative values are being developed; however, most brain maps are not time locked to an event or brain state; therefore, comparisons of frequency bands are difficult to accomplish across groups or disease states. A clear definition for the clinical correlation of the brain maps is still needed; therefore, EEG brain mapping and other advanced QEEG techniques should be used only by physicians highly skilled in clinical EEG and only as an adjunct to traditional EEG interpretation.

These tests may be clinically useful only for patients well selected on the basis of their clinical presentation. Certain QEEG techniques are considered established as an addition to the digital EEG and include screening for possible epileptic spikes or seizures, long-term EEG monitoring or ambulatory recording, and operating room (OR) and ICU monitoring.

Continuous EEG monitoring by frequency trending helps to detect early intracranial processes in the OR or ICU (eg, screening for possible epileptic seizures in high-risk patients in the ICU). QEEG frequency analysis may be a useful adjunct to interpretation of the routine EEG. In a number of conditions, such as postconcussion syndrome, head injury, learning disability, attention disorders, schizophrenia, depression, alcoholism, and drug abuse, the use of QEEG remains investigational. On the basis of available clinical and scientific evidence and expert opinions, QEEG is not currently useful in civil or criminal cases.

QEEG is a derivative of regular EEG. The original data must be evaluated before proceeding to further mathematical translation of this same data set. A thorough understanding and firm knowledge base in clinical EEG diagnosis may help prevent erroneous interpretations of digitally displayed mathematical constructs (eg, brain map, coherence map). Ideally, only physicians properly trained in EEG and, in addition, sufficiently well trained in mathematics and computing science should use these new technologies. A substantial risk of erroneous interpretations exists if any of the elements required is missing. Clinical use of any of the EEG brain mapping or other QEEG techniques by practitioners who are not physicians highly skilled and properly trained in clinical EEG interpretation or without reviewing the original record should be unacceptable.

The EEG is regarded as a fairly nonspecific measure of clinical states, as detailed in this article. A limited number of abnormalities that can be recognized in widely varied disease states exist. On the other hand, an abnormal EEG is a sensitive measure of brain function. When the patient has clinically symptomatic encephalopathy or moderate dementia, the EEG is almost always abnormal. Clinical utility of EEG needs to be appreciated in a different way than some other diagnostic procedures. In medicine, clear correlations and specific answers are desired; however, based on the modality and the underlying principle of measurement, this goal cannot always be achieved.

Besides EEG, the MRI scan is very sensitive in showing various lesions in the brain. But knowing exactly what these lesions represent is difficult on the basis of mere appearance. While interpreting MRI images, the clinician relies to a large extent on other relevant clinical information. EEG measures electrical field variations, and a number of clinical conditions can disturb the normal electrical field of the brain. Simple state or electrolyte changes may alter the appearance and time variation of the brain-generated electrical fields; hence, a large number of conditions cause the EEG to appear abnormal. In EEG practice, the clinician has to rely to a large extent on the clinical history and the neurologic examination findings to make a clinically meaningful conclusion.

In most instances, the correct question may be whether the EEG is normal or abnormal. The next step is to decide how an abnormal EEG would help the clinical diagnosis; therefore, the EEG can be used to confirm clinical observation or suspicion, or to determine the extent of the abnormality for prognostic purposes (ie, attempting to predict outcome of the clinical condition). Sometimes, EEGs serve as a "proof" to families that indeed the brain function is disturbed so greatly that recovery is doubtful. In such cases, the EEG helps resolve anxiety and supports a more correct ethical decision.

The newer EEG techniques offer a number of conveniences and also enhance communication between the electroencephalographer and other clinical specialists; however, they may not make the record more specific but merely easier to understand. Computer analysis on the other hand may offer features that, although present in the regular record, are difficult or time-consuming to extract and display. EEG has a definite role in evaluating changes in mental states. It can confirm or refute nonconvulsive status epilepticus. EEG changes are usually proportional to the degree of metabolic, hepatic, or renal encephalopathy. EEG often is abnormal in subdural hematoma, normal pressure hydrocephalus (NPH), and CJD.

Computer analysis of the EEG may help reveal subtle changes in AD. This is promising and hopefully will soon become clinically useful and available. In cases of clinical dementia, a normal EEG with preserved alpha might help establish the diagnosis of Pick disease, while a slowed and shifted alpha frequency is seen in AD and PSP. Low-voltage, flat EEG and the appropriate clinical presentation may raise the suspicion of Huntington disease. While the EEG is nonspecific, in most cases it is abnormal in altered mental states.

EEG study should be requested if clear clinical indications are present or if the clinician has a reasonable presumption that it may give clinically relevant information. This information is expected to alter the clinical decision-making process. It should help the referring source in diagnosis and treatment. Another role is to help the patient or family to understand the ongoing disease process. EEG frequently is ordered to evaluate patients with different degrees of mental and behavioral changes and encephalopathy or coma. Usually, nonspecific abnormalities are present that do not give definite information about the cause of the underlying process but do provide information on its location and severity; therefore, unless epileptic seizures are a consideration, the EEG does not give direct unequivocal information on the cause of the patient's condition.

Nevertheless, the study may differentiate between a generalized and a focal abnormality. This may guide the clinician to further appropriate imaging studies. On the other hand, if the abnormality is generalized, the EEG can be used to characterize and monitor the disease process. With coma, the EEG may help in predicting the neurological prognosis. EEG is an important diagnostic tool in dementias in which specific morphological lesions are not apparent on imaging studies.

Personal perspective

EEG often is compared with MRI; when this comparison is made, it usually refers to clinical MRI and not functional MRI studies. This comparison is puzzling, since a primarily anatomical test is being compared with a functional one. MRI is good at telling us where the lesion is, while the EEG is pretty good at separating normal and abnormal primarily cortical function. Topological usefulness of EEG is limited, although with computerization it may be improved. The purpose of MRI is to provide precise localization of a lesion, usually one that has passed a certain stage of evolution. The EEG, on the other hand, captures the changing electrical characteristics of a functioning brain, primarily those of the cortex.

Conditions can be identified with EEG that as a rule cannot be seen on the MRI; therefore, the use of these studies is not exclusive but complementary. The EEG may be used for the following:

• To exclude nonconvulsive status epilepticus
• To identify focal interictal epileptiform activity to confirm clinical suspicion that seizures may contribute to the condition in question
• To attempt to record functional disturbance in individuals whose brain MRI is "normal" but brain dysfunction is evident clinically (eg, metabolic encephalopathies)
• To attempt to record disease-specific patterns in the proper clinical setting, such as progressive myoclonic epilepsies, CJD, SSPE
• To help a psychiatrist with the multitude of complex disorders masking as potential epilepsy or encephalopathy (eg, lithium intoxication may present with BiPEDs)
• To identify focal or lateralized changes that suggest a structural cause to the encephalopathy

The truth often is stated that EEG is nonspecific and cannot diagnose etiology or localization well (eg, the cause of coma). However, nonspecificity is often not the question in general medical practice because most of the referrals in general neurology are individuals in whom the cause is pretty well clear, or reasonably suspected, on the basis of clinical history and laboratory chemistry. The question from the clinician is whether the brain is involved and the extent of brain damage, if any. To answer these questions, presently no clinical tool is more useful than the EEG.