Epilepsy: A Comprehensive Textbook
2nd Edition

Data are collected following the standard procedure for EEG recording. The number of recording channels and the montage used depend on the characteristics of the clinical situation being investigated.

Surface electromyogram (EMG) activity is obtained by applying two electrodes over the muscular belly of the recorded muscles. To allow collection of high-frequency EMG signals, a short time constant (0.03 second or less) to reduce artifacts caused by movement or sweating and minimum filtering are necessary. Simultaneous EEG and EMG recording can provide useful information to verify the existence of correlations between cortical and muscular events. Analysis of latency between EEG and EMG events, which was once difficult and imprecise by visual inspection of data on paper, can be performed more precisely and reliably by using computerized polygraphic systems that collect the neurophysiologic signals in a digital format and allow off-line processing and analysis of the data. Recording of EMG activity from antigravity muscles can be useful for investigating modification of the muscular tone associated with paroxysmal discharges. Disorders such as tremor or myoclonia require the monitoring of muscular activity of both agonist and antagonist muscle groups.

The inherent limitations of surface EMG, including difficulty recording from a deep or a single muscle without the interference of nearby muscles, loss of high-frequency muscular activity, and inability to obtain single-motor-unit activity, mean that there is only a gross correlation between the EEG and the compound activity of a single muscle. For analysis of the effect of epileptic activity on the single motor unit, needle EMG recording is indicated.

Eye movement recording is usually performed by placing two electrodes at the outer canthi to record horizontal movements and one electrode above and the other below one eye to record vertical movements.

When only one electrooculogram (EOG) channel is available, two electrodes can be used, one placed above the outer canthus of one eye and the other placed below the outer canthus of the opposite eye. Other systems have been proposed, including using infrared-detecting cells mounted on a spectacle frame together with an infrared-emitting diod⁴³ or employing piezoelectric sensors.⁵³ Electrooculography is particularly useful for detecting artifacts induced by eye movements, investigating lambda waves, studying ocular manifestations correlated with EEG paroxysms, such as in cases with epileptic nystagmus,^58,103 and scoring sleep stages in polysomnography. Computerized methods for removing the EOG artifacts using the EOG signal have been extensively described by Barlow.⁵

Electrocardiograms (ECGs) can be easily recorded with an EEG device by placing two electrodes on the chest wall or—to obtain a lead I derivation—on the right and left arms. Sensitivity must be adjusted by switching from microvolts per millimeter to millivolts per millimeter; bandwidth ranges from 0.8 to 60 Hz. Simultaneous recording of EEG and ECG allows the detection of pulse and ECG artifacts contaminating the EEG tracing.

Respiratory activity can be monitored either by measuring the changes in thoracic volume due to respiratory movements or by detecting the flow of air through the nostrils and the mouth. Modifications of chest volume can be measured using a strain gauge attached to a piece of elastic forming a band around the chest. Variations of chest circumference due to respiration stretch the elastic band, causing a change in resistance in the strain gauge. Change in resistance can be measured by means of a DC-excited Wheatstone bridge circuit. This system requires that the patient remain still because chest size in moving individuals can change independently from respiration.

Another method for monitoring respiratory activity is to detect the air flow through nostrils and mouth by means of a thermocouple, which generates a current whenever warmed, or by a thermistor, which requires a constant-current source and reproduces changes of temperature as variations of resistance. Both thermocouples and thermistors must be placed between the nostrils and the mouth to detect differences in temperature between inspired and expired air. All of these methods require a long time constant (0.6 or 1 second) and filtering of high frequencies.

Electrodermal (EDG) response, also referred to as electrical skin resistance, galvanic skin reflex or response, and psychogalvanic reflex, is a change in resistance and the generation of a potential between areas containing many sweat glands and areas almost devoid of them, due to modifications of sweat gland activity. Change in resistance is also called the Fere effect; the generation of a potential is called the Tarchanoff response. The Fere effect and the Tarchanoff response are probably two measures of the same phenomenon, both being affected by cholinergic modulation.⁷⁰

Measurement of EDG response is performed by placing one electrode in the palm of the hand or on the sole of a foot and another in an area without sweat glands (such as the back of the hand). Another method is to place both electrodes on the dorsal surface of the forearm 5 to 6 cm apart. Recording must be performed using the longest time constant available with the maximum high-frequency filtering. Measurement of the skin potential, which appears as a DC potential (Tarchanoff response), is quite easy; however, it is unstable due to the possible occurrence of an offset potential at the electrode–skin interface and is difficult to control and distinguish from the EDG. For these reasons, measurement of change in skin resistance is more appropriate.

In normal individuals, resistance recorded from two electrodes on the hand ranges from 20 kΩ to perhaps 0.2 MΩ. If the autonomic nervous system is malfunctioning, resistance can reach values of >1 MΩ. Variations of EDG by measuring the Tarchanoff response during sleep in normal individuals and in patients with epilepsy have been described by Broughton et al.¹⁶ Arousal is associated with a decrease in resistance and relaxation with an increase in resistance. Evident EDG modifications were observed—especially during sleep—and were associated with temporal lobe interictal discharges and brief generalized discharges of polyspikes (Fig. 1).

Blood pressure can be monitored invasively by introducing a microcatheter in the radial artery. The microcatheter is connected with a pressure transducer to polygraph preamplifiers. DC recording is required. It is possible, via microcatheter, to obtain blood samples for blood gas measurements. A preferable method for monitoring blood pressure is a noninvasive assessment based on the computerized processing of instantaneous blood pressure modifications, which are collected by a plethysmographic system applied, usually, to the middle finger of the hand.

A micturition recorder is a device that allows the detection of enuresis or loss of urine due to a convulsive seizure. Broughton³⁷ designed a system consisting of two long electrodes, in separate nylon sheaths, placed parallel with and close to one another in a serpentine shape close to the urinary meatus. The signal recorded, using a high upper-frequency response and a short time constant, is a 50-Hz artifact (60 Hz in North America). As soon as the patient urinates, the first drops of urine create a conductive path between the two electrodes that causes the disappearance of the artifact.

Systems for signaling body movements are usually based on displacement transducers, accelerometers, or actigraphy monitors. These systems can be applied to the bed of the patient—especially when the aim is to detect seizures during sleep¹³¹—or to body parts. A wrist accelerometer can be extremely useful for monitoring movements during sleep,⁶⁰ identifying different types of movements (tremors, myoclonia),⁷² and differentiating EEG activity from movement artifacts.¹⁸ Actigraphs are small, wrist-worn devices (usually about the size of a wristwatch) that measure movements; they are equipped with a microprocessor and an on-board memory that can allow off-line analysis and display. Actigraphy monitoring is particularly useful in the study of sleep and circadian rhythms.^2,64

Motility of the gastrointestinal tract was studied by means of EMG recording of the duodenal activity¹¹⁶ using bipolar platinum electrodes. This study investigated modifications of duodenal motility in relation to wakefulness, sleeping (Fig. 2), and ictal activity (Fig. 3). Cherubini et al.²⁰ polygraphic techniques, including monitoring of EMG activity of intestines in cats, to investigate the effects of pharmacologically induced generalized epileptic seizures on intestinal motility.

A positive myoclonus is characterized by a massive, shock-like muscular contraction involving one or more body segments. It appears in the EMG as a brief burst of muscular potentials (Fig. 4), synchronous on agonist and antagonist muscles, with or without an EEG correlate. The opposite phenomenon, a “negative” myoclonus, is a brief interruption of a tonic muscular contraction (Fig. 4), sometimes clinically indistinguishable from the positive myoclonus.^94,123 When associated with a paroxysmal EEG event, a negative myoclonus is defined as “epileptic negative myoclonus” (Fig. 5).^48,106,122 Negative myoclonus, defined as an interruption of tonic muscular activity for <500 msec without evidence of preceding myoclonia, was recognized in 2001 as a seizure type by the Task Force of the International League Against Epilepsy on Classification and Terminology.¹³

A spasm, which appears as a massive but slow contraction reaching a climax and progressively decreasing, more often than not involves axial and proximal muscular groups. Indeed, polygraphic recordings with multiple EMG leads have demonstrated that epileptic spasms can occur with a complex pattern of muscular activation, with the involvement of cranial as well as limb and axial muscles. This complex pattern of contraction can appear clinically similar in different spasms, although with variable sequences of muscular recruitment (Fig. 6).^10,134 On the EEG, spasms can be devoid of any modification of tracing or they can be associated with flattening, diffuse low-voltage fast activity,⁴² or a slow wave.^34,46

Tonic contractions consist of slow, sustained contractions maintained over time and involving several muscular groups. They are usually associated with fast recruiting EEG activity.

Clonic contractions are characterized by a series of jerks, appearing on the EMG as hypersynchronous muscular potentials. They can vary in amplitude, symmetry, frequency, and topography and are often related to contralateral EEG spikes.⁴²

Atonic phenomena can appear as a sudden global or focal loss of muscular tone, characterized polygraphically by an abrupt flattening of the EMG activity. They are associated with different types of EEG paroxysmal discharges (generalized spike- or polyspike-and-wave, diffuse fast rhythmic spikes, bilateral synchronous fast waves intermixed with slow waves) or with no EEG changes at all.⁴²

Polygraphic recordings make it possible to characterize the complex manifestations of generalized tonic–clonic seizures, which are composed of a more or less stereotyped sequence of motor phenomena.^9,29,38,40 This sequence consists of an initial tonic phase of sustained muscular contraction lasting 10 to 20 seconds and involving all skeletal muscles. This phase is responsible for the characteristic body attitude (a flexion followed by a longer extension phase). A diffuse vibratory contraction follows this initial phase, and clonic manifestations follow next. The clonic manifestations consist of brief, maximal flexor contractions of the whole body, with the interval between the jerks becoming progressively longer. It is interesting that polygraphic recordings have demonstrated that, just a few seconds after the last clonic jerk, a new tonic phase usually reappears, as intense as the first one but with different topographic distribution, involving mainly facial and masticatory muscles.⁴²

The EEG correlates of a tonic–clonic seizure are usually represented by an initial desynchronization, sometimes preceded by generalized bursts of polyspike-and-waves, followed by a recruiting rhythm. These are intermixed with slow waves of decreasing frequency and increasing amplitude that correspond, at a certain point, with the interruption of the tonic massive contraction and with the onset of the clonic manifestations. A flattening of the EEG activity follows, which represents the phase of “cortical extinction” and lasts several seconds (see Chapters 47 and 74).

Clinical features of tonic seizures consist of a brief (5 to 20 seconds) muscular tonic contraction accompanied by impairment of consciousness and involving, with varying degrees of intensity and extension, distinct muscular groups (i.e., the head and trunk or the limb girdles and, to a lesser extent, the legs) or the whole body.^19,40,41,42 The associated polygraphic findings are represented by different EEG patterns such as (a) flattening of the tracing, (b) fast activity (around 20 Hz) of progressively increasing amplitude, and (c) recruiting rhythmic discharge at about 10 Hz, sometimes of high amplitude from the onset. The EMG leads show an interference pattern in all involved muscles corresponding to the tonic contraction (Fig. 7). Different
patterns of muscular activation can be observed according to the duration of the seizure: When it lasts a short time, the contraction is maximal at the onset, then decreases; when it is of longer duration, a progressive increment of the muscular activity reflects the increasing intensity of the contraction. In global tonic seizures, an axial preponderance of the muscular activation is evident. Asymmetry of the tonic contraction or the occurrence of myoclonia at the end of the tonic manifestation can be observed. Tonic seizures are often associated with autonomic changes, such as modifications of heart and respiration rate, mydriasis, vasomotor phenomena, increase in intravesicular pressure, and positive electrodermogram responses (see Chapter 52).^19,40

In myoclonic absences, the appearance of a generalized 3-c/s spike-and-wave discharge lasting up to several seconds is associated with myoclonic jerks (see Chapter 240). These jerks are particularly evident in the upper limbs and are polygraphically characterized by myoclonic potentials that, a few seconds after the beginning of the discharge, are superimposed on a progressively increasing tonic muscular contraction, which mainly involves the shoulders and the deltoids and results in the abduction and elevation of the upper limbs (Fig. 8).^109,118 With extra leads, facial muscles and the mentalis and orbiculus oris muscles can be observed for twitchings synchronous with the spike-and-wave complex. Oscilloscopic analysis of the EEG/EMG correlation has demonstrated that the positive transient of the 3-c/s spike-and-wave complex¹³⁷ is correlated with the appearance of the myoclonic jerks. Latency between the myoclonic jerk and the spike on the EEG is 15 to 40 msec for the upper limbs and 50 to 70 msec for the lower limbs. The myoclonic bursts are followed by a brief (60 to 120 msec) silent period that interrupts the tonic contraction (Fig. 9).¹⁰⁹

Polygraphy can be extremely useful for supporting a diagnosis of juvenile myoclonic epilepsy, in which brief myo-clonic jerks—particularly in the upper limbs—are associated with polyspike-and-wave complexes at a frequency of 3 to


3.5 c/s (see Chapter 244).^54,99 Computerized polygraphic recordings have shown that myoclonic potentials in juvenile myoclonic epilepsy are related to a cortical positive potential encompassed in the polyspike discharge of the polyspike-wave complex.⁷³

In cases of epilepsia partialis continua, polygraphic study allows the analysis of the rhythm, intensity, rate, and distribution of the muscle jerks.¹²⁷ Myoclonic EMG potentials can be associated on the EEG with slow focal abnormalities, focal paroxysmal discharges, or no evident paroxysmal activity (Fig. 10) (see Chapters 61 and 243).

Some patients with epilepsy can present with a clinical picture resembling epilepsia partialis continua, characterized by frequent, subcontinuous jerks, evident when the patient maintains a posture, or by a tonic contraction (Fig. 11). Polygraphic recordings can be crucial in defining the exact nature of this motor disorder, which can be caused by subcontinuous epileptic negative myoclonia (i.e., brief lapses of the muscular activity time-locked to paroxysmal EEG activity).^48,106,122 In epileptic negative myoclonus, the interruption of the muscular activity occurs synchronously on agonist and antagonist muscles; when epileptic negative myoclonus is focal, involving one or both limbs on the same side of the body, it is usually associated
with a contralateral EEG spike located in the centroparietal region. Epileptic negative myoclonus can be found in a wide spectrum of epileptic conditions—idiopathic, cryptogenic, and symptomatic (see also Chapter 277).¹²²

The polygraphic features of progressive myoclonus epilepsies, such as Unverricht-Lundborg disease, consist of constantly present action myoclonus, characterized by high-amplitude EMG potentials of short duration (20 to 30 msec), which are synchronous on agonist and antagonist muscles. These are followed by an EMG-silent period lasting 40 to 120 msec (rarely up to 300 msec) (Fig. 12). The myoclonic bursts and the silent periods are inconstantly related to EEG spike-and-waves and polyspike-and-waves.¹¹³ Myoclonic seizures in Unverricht-Lundborg patients are characterized by generalized myoclonia, predominant proximally in the upper limbs, with varying rhythm and associated with generalized, symmetric polyspikes or polyspike-and-waves. Lafora disease is characterized polygraphically by the presence of abundant, asymmetric, asynchronous, subtle myoclonia at rest, diffuse to all muscular groups, usually without an EEG correlate.^81,110 In the progressive myoclonus epilepsies, intermittent photic stimulation is very effective in eliciting fast polyspikes and polyspike-and-waves associated with massive myoclonic jerks (see also Chapter 252).^81,112

In patients presenting with Lennox-Gastaut syndrome (see Chapter 241),⁴¹ nocturnal tonic seizures can recur so frequently during non-REM sleep that they produce a tonic status epilepticus, which is characterized polygraphically by a progressive attenuation of tonic motor manifestations, paralleled by a worsening of the vegetative parameters and by an impairment of consciousness between seizures. The situation can deteriorate into a coma-like state, with autonomic derangement and mild motor phenomena, such as raising of the hands and slow eye movements. An adequate polygraphic monitoring can be of utmost importance for the diagnosis and treatment of such a condition. Patients with Lennox-Gastaut syndrome can also show minor nocturnal seizures, which are associated with eye-opening spells, head nodding, and changes in cardiac and respiratory rate. Sometimes these seizures are uncovered only by means of polygraphic recording, represented by generalized polyspikes, fast recruiting rhythms, or slow spike-and-wave discharges.^4,19 A disruption of sleep architecture has been reported in patients with Lennox-Gastaut syndrome.^1,4

Polygraphic studies during sleep have demonstrated modifications of myoclonic phenomena in different pathologic conditions. Epilepsia partialis continua is usually characterized by the persistence of focal myoclonic jerks, which modulate only in intensity and frequency, across all sleep stages (see Fig. 10).⁶⁵

In Unverricht-Lundborg disease, sleep studies demonstrate (a) lack of activation of generalized paroxysmal discharges and (b) appearance of focal multiple fast spikes occurring in repetitive bursts, localized over the midline and centroparietal regions, and occurring more frequently during REM sleep, particularly when eye movements are abundant. These fast spikes can be time-locked to myoclonic jerks, particularly in those muscles that show a striking action myoclonus during wakefulness (Fig. 13).^111,113,120

In Lafora disease, sleep organization is extremely altered, with the different stages barely recognizable: paroxysmal activity does not seem to increase during sleep; diffuse multiple fast spikes show variable amplitude and topography and can be intermixed with fast activity, and posterior spikes persist during slow sleep and can show an enhancement during REM sleep.^112,120

Myoclonic encephalopathies can be properly investigated by means of sleep polygraphy to detect myoclonus during sleep when other associated abnormal movements that occur in wakefulness disappear.²⁷ In patients presenting with myo-clonic epilepsy with ragged red fibers (MERRF), sleep polygraphic studies showed the occurrence of vertex and central spikes, absence of physiologic EEG elements of stage II sleep, and persistence of generalized paroxysmal activity as in wakefulness.⁸² Sleep polygraphic investigations can be relevant in the study of aminoacidopathies, in which multifocal
paroxysmal abnormalities and suppression burst patterns are inapparent during wakefulness but are unveiled by sleep.

One such technique, a computerized EEG averaging technique called jerk-locked averaging, was applied to polygraphic data by Shibasaki and Kuroiwa.⁹⁵ Jerk-locked averaging is triggered
by an EMG pulse, which is represented by a myoclonic potential, and permits cortical correlates associated with muscular events to be extracted from background EEG activity (Fig. 14). This allows different types of myoclonus to be characterized in terms of presence or absence of a cortical correlate and according to the latency between the cortical correlate and the myoclonic jerk.^96,97

A similar technique, triggered not by a myoclonic jerk but by the onset of an EMG silent period on a background of a tonic contraction, was introduced by Ugawa et al.¹³⁰ Known as silent-period locked averaging, this technique allows different types of negative myoclonus, including asterixis, to be identified and associated with cortical events.

Spike activity associated with different muscular events can be investigated by means of EEG computerized systems with an extra channel for polygraphic signals and displaying the topographic distribution over the scalp of the EMG related-cortical potentials. These techniques have been adopted to investigate paroxysmal activities associated with epileptic negative myoclonus.^57,88 Through the finding of a frontal spike component that suggests the involvement of inhibitory frontal areas in the generation of negative motor phenomena (as subsequently demonstrated by intracerebral electrical stimulation⁸⁶), spike-averaging procedures have demonstrated that spikes associated with epileptic negative myoclonus differ from spikes without epileptic negative myoclonus (Fig. 15).⁸⁸

A quantitative analysis of EMG signals can be performed by rectification and integration of EMG wave forms. These techniques allow the area between the EMG tracing and the baseline to be calculated. The area varies according to modification of amplitude, frequency, and duration of the potentials, which provides a quantification of EMG activity.

Multichannel computerized EMG recordings with only a few EEG channels can analyze the pattern of sequential activation of different muscular groups that participate in a sudden motor event (i.e., a massive myoclonic jerk) (Fig. 16) or in the complex motor manifestations characterizing focal seizures involving motor areas⁶⁷ (Fig. 17). Analysis of the temporal
sequence of the onset of contraction in different cranial and limb muscles has allowed a distinction to be proposed between cortical and reticular myoclonus.^50,51,85

Investigations on the modulation of the motor system during ictal and interictal paroxysmal discharges can be performed by combining polygraphic recordings with techniques able to assess cortical (e.g., transcranial magnetic stimulation) and spinal (e.g., H reflex, stretch reflex, F wave) excitability. During epileptic negative myoclonus, an inability of transcranial magnetic stimulation to evoke a motor response, associated with the preservation of the F wave, suggests a cortical origin for this motor disorder (Fig. 18).¹²² In patients with severe partial epilepsies and falling seizures, H-reflex and motor-evoked potentials are unmodified during slow spike-and-wave complexes, whereas they appear to increase during subclinical fast polyspike discharges. A similar facilitation can be observed at the beginning of tonic seizures, when the EMG is silent. During the ensuing tonic contraction, H-reflex and motor-evoked potentials are further enhanced, decreasing consistently only in the postictal phase. In contrast, 3-c/s spike-and-wave discharges of typical absences or evoked by intermittent photic stimulation do not alter motor-evoked potentials or the stretch reflex.¹²⁴