Clinical manifestations

Presentation and course

Reading epilepsy with jaw myoclonus. Reading-induced seizures usually begin in neurodevelopmentally normal adolescents and consist of clusters of single and brief (instantaneous) myoclonic jerks mainly restricted to the masticatory, oral, and perioral muscles (06; 62; 47; 31; 49; 04; 30; 41).

The most characteristic symptom is abnormal sensation or movements involving the jaw. This is described as a clicking sensation, jerking, tightness, or numbness and occurs after a variable latency (usually 3 to 20 minutes) from the onset of reading, particularly loud reading of texts difficult or unusual to read. Its severity and frequency varies even in individual patients. One quarter of the patients may have also similar jaw jerks provoked by talking (particularly if this is fast or argumentative; see case 1 of clinical vignettes), writing, reading music, or chewing. Text messages on mobile phones have also been reported as provoking seizures in reading epilepsy (60), and there is also an unusual case of a woman whose reading-induced seizures were initiated by specific letter fonts (61).

As a rule, there is no impairment of consciousness during the jaw myoclonus. This is expected because the jaw myoclonus is brief and instantaneous without massive cortical involvement. However, many patients complain about a feeling of unrest, discomfort, anxiety, or confusion. Some patients describe themselves as “sticking to the word” or, more frequently, report “loss of track of the reading text” when a jaw jerk occurred, which made them restart reading from the beginning of the sentence. The phenomenon seems to correlate with the severity of motor manifestations, though some patients may also experience dyslexia or alexia after the jaw myoclonus (01). Ictal stammering may also occur, probably because the brief ictal jaw or perioral clonic or tonic activity impedes the flow of speech.

Generalized tonic clonic seizures. The majority of patients have only 1 generalized tonic clonic seizure (GTCS), which is usually self-inflicted because of their curiosity to see what will happen if they continue reading despite jaw jerks or other manifestations. In these occasions of continued reading, the jaw jerks may become more violent and spread to trunk and limb muscles before a GTCS develops. This is usually the first and last GTCS because the condition is effectively treated and the patient learns to stop reading or talking when oral or perioral jerks occur. It is extremely rare for patients with reading epilepsy to have more than 1 to 5 generalized tonic clonic seizures while reading or spontaneously, not related to reading.

See video EEG in (22).

Other types of reading-induced seizures. It is rare for reading epilepsy to present with other types of ictal manifestations (mainly visual hallucinations) in addition to the jaw myoclonic seizures. Spontaneous myoclonic jerks are rare and may occur in those with a more widespread cortical hyperexcitability than in pure reading-induced epilepsy. Hand myoclonic jerking is common in those with writing precipitation of seizures.

Absence seizures induced by reading are exceptional (36; 55).

An overlap of reading epilepsy with juvenile myoclonic epilepsy was suggested on the basis of clinical findings, such as the age at onset, bilateral myoclonic jerks, strong influence of heredity, progression to generalized tonic-clonic seizures, response to sodium valproate, and persistence through life (47; 64). The conclusion of Wolf and colleagues is that “perioral reflex myoclonus” is not a specific finding in primary reading epilepsy, but “it seems to be a reflex epileptic trait which can be observed, like photosensitivity, in conjunction with many epileptic syndromes, both generalized and focal. Juvenile myoclonic epilepsy seems to be the syndrome in which it is second most common (with talking as the predominant precipitating mechanism, and reading only in a minority of patients)” (38; 64).

Reading epilepsy with prolonged focal seizures manifested with alexia. This variant of primary reading epilepsy manifests with prolonged (longer than 1 to 3 minutes) focal seizures always triggered by reading and characterized by ictal alexia associated with a posterior temporal ictal discharge in the language-dominant hemisphere (see case 2 of clinical vignettes).

Prolonged alexia is the prevailing ictal manifestation in this form of reading epilepsy, and it is strikingly different from the transitory “loss of track in the text” of the myoclonic variant. Visual symptoms (mainly illusions and palinopsia) are also described by some patients.

Reading-induced focal seizures manifested with alexia may be apparently primary because of the normal neurologic status, resting EEG activity, and brain MRI in most patients. Only 1 definite symptomatic case has been described (09); an arachnoid cyst found in another patient (31) is of uncertain causative significance.

There are nearly 10 reported cases, but the possibility of under-recognition or misdiagnosis may contribute to its rarity. It is possible that the lack of the typical jaw jerk may have led to the misdiagnosis of some patients as having temporal epilepsy or to the misplacement of others into the group of the poorly defined “secondary” reading epilepsy, which has received little attention in recent years.

The case reported by Chavany and colleagues is usually cited as symptomatic reading epilepsia with alexia (09). This was a rather unusual 30-year-old patient with relatively frequent generalized convulsions from the age of 10 months to 8 years. From 8 years of age to the time of the report, the attacks were less severe. The majority of them were evoked by reading. There would be a prolonged warning of various subjective sensations. Before each attack, words would appear to be “out of place” and letters interchanged prior to a secondarily generalized tonic-clonic seizure of right-sided onset. On regaining consciousness, he was sometimes unable, for several minutes, to articulate although he knew exactly what he wanted to say and could express himself by gesture. There were structural abnormalities in the left occipital region. The resting EEG was abnormal, and overbreathing caused small isolated sharp waves in the left occipital region. On 2 occasions, reading provoked an attack after a short time. Ictal symptoms consisted of the letters changing place vertically and horizontally and becoming entirely incomprehensible. This was associated with a rhythmical theta activity of left occipital predominance, which was replaced 20 seconds later by delta activity at 1.5 to 2 Hz, the whole electrical event lasting for approximately 2 minutes.

Gastaut and Tassinari described a 15-year-old boy who, since the age of 6 years, had experienced several epileptic seizures exclusively during prolonged reading (19). On EEG, 2 seizures were evoked, characterized by the patient seeing the last word in a foggy way; this word became fixed before him even though he kept his eyes closed; then the letters changed place, and the word became distorted. In the first recorded seizure, which occurred while he was doing a Latin translation, the word invicta obtruded like a true hallucination, then was transformed into victa and finally into victoria (victory). EEG consisted of a fast rhythmic discharge lasting for 50 seconds and remained localized at the left parieto-occipital region. A second seizure in the same subject was preceded by a slowing of the reading speed and a laryngeal sensation that coincided with a brief generalized burst of spike-and-wave activity of right predominance. At this moment, the subject interrupted the reading and, a few seconds afterwards said, “I am going to have a seizure.” He was persuaded to continue reading; the laryngeal sensation reappeared, and the reading became irregular. T patient repeated the same words several times; at the same time on the EEG there appeared 2 sporadic bursts of spikes and waves, predominant in the parietal region of the right hemisphere, followed in that region by the appearance of low-voltage fast activity, intermixed afterwards with spikes and spikes and waves, while the subject presented a deviation of the body and head towards the left followed by a tonic-clonic seizure.

Patient 12 of Radhakrishnan and colleagues gave a history of inability to comprehend the meaning of the words without any jaw myoclonus before 3 generalized tonic-clonic seizures induced by reading. This patient had a left posterior temporal-occipital seizure during pentylenetetrazol activation while reading. In addition to the focal seizures, and independent of these, he had reading-induced jaw jerks associated with generalized spike wave discharges (47).

Koutroumanidis and associates described 2 cases (patients 16 and 17) with video-EEG-documented reading-induced prolonged seizures of alexia (31). Both described an associated dizziness and an ill-defined feeling of discomfort. No other ictal or postictal symptoms or clinical changes were elicited by history or observed on analysis of video tapes. Patient 16 had reading-induced focal seizures that evolved into generalized tonic-clonic seizures on several occasions. In Patient 17’s reading-induced alexia never evolved into generalized tonic-clonic seizures (see case 2 of clinical vignettes). He never had spontaneous daytime seizures but suffered infrequent nocturnal convulsions coinciding with the onset of the reading-provoked seizures. The routine EEG was normal in both patients. Reading characteristically provoked a subclinical focal EEG activation over the left posterior temporo-parieto-occipital area. Ictal EEG changes were prolonged and clearly focal over the left posterior temporo-parieto-occipital region. Treatment with phenobarbital, diphenylhydantoin, and clonazepam over a period of 7 years was ineffective in patient 16; patient 17 received sodium valproate for 2 years, also without apparent benefit. Introduction of carbamazepine resulted in considerable improvement in both patients, with less frequent reading-induced focal seizures and rare generalized tonic-clonic seizures associated with poor compliance. Brain MRI was normal in patient 17. An arachnoid cyst at the left temporal pole was found in patient 16.

Maillard and associates described 2 cases of video-EEG-documented reading-induced focal seizures manifested with alexia. Onset was in adolescence; there were no spontaneous seizures and no other triggers other than reading. Brain MRI was normal (34). Case 1 was a right-handed 41-year-old man who had his first seizures at the age of 14 years. Seizures were provoked by both silent and aloud reading and began with a feeling of dizziness and inability to continue reading, followed by a sensation of oropharyngeal discomfort. Attempts to continue reading provoked generalized tonic-clonic seizures. The duration of reading necessary to elicit a seizure was variable, either brief (such as reading a road sign when stressed) or extended up to 10 minutes. In 1 of his EEG-recorded seizures provoked by reading:

Case 2 was a 40-year-old right-handed male patient who had his first seizures at the age of 14 years while reading a German text in school. Seizures began with alexia followed by loss of contact and chewing and occurred after several minutes of silent reading. He never had any spontaneous seizures. Generalized tonic-clonic seizures occurred once or twice a year:

Osei-Lah and colleagues described a 19-year-old right-handed student with reading-induced focal seizures of alexia progressing to generalized tonic-clonic seizures (44). “While reading very late at night his reading suddenly became disturbed; he could not make sense of what he was reading and felt confused. He then lost awareness and recovered in the ambulance en route to a hospital.” He had never had jaw or facial myoclonic jerks associated with his seizures. All of his seizures were associated with reading at the onset. He reported that the seizures were more likely to occur when reading difficult material or printed material with small font and when reading during late afternoon or evening or when he was tired. He frequently read at other times without any problems. Writing, talking, solving mathematical problems, or listening to material being read to him did not trigger a seizure. He had not attempted reading texts in foreign languages. Brain MRI was normal. The interictal EEG was practically normal. Reading (both silently and aloud) consistently evoked sharp and slow wave complexes over the left mid- and posterior temporal region, which ceased when the patient stopped reading. He reported no symptoms during these. Prolonged reading resulted in a 6-minute focal seizure of left posterior temporal onset with secondarily generalization. The patient reported difficulty making sense of the text at the onset of the seizure. No facial, jaw, or limb myoclonic jerks were seen.

Gavaret and colleagues studied a 31-year-old right-handed woman with focal seizures that always occurred during silent reading (20). Seizures began at the age of 28 years. During the 3 years prior to assessment, she had experienced a total of 8 seizures. Physical and neurologic examinations were normal. Seizures began during silent reading with the feeling of no longer being able to understand what she was reading. After looking up from the page, she then continued to see letters and words despite actual disappearance of that image from either visual field (palinopsia). She had a feeling of strangeness. She could then have right hemi-body jerks and secondarily generalization. Seizures usually occurred soon after the onset of reading (less than 10 minutes). She had not abandoned reading altogether but had developed a distinct style of reading to try to avoid the onset of seizures, in that she read only for short periods and tended to scan the page diagonally. She was initially treated by lamotrigine (200 mg/day), but seizures persisted with secondarily generalization. Replacement of lamotrigine with carbamazepine (1000 mg/day) led to a considerable improvement. Brain MRI was normal. Interictal EEG showed isolated left posterior temporal spikes. Interictal [18 F] fluorodeoxyglucose-PET showed bilateral occipito-temporal hypometabolism with left-sided predominance. During video-EEG monitoring, a seizure was recorded 5 minutes after silent reading of a women’s magazine. Clinically, the patient experienced habitual subjective signs (alexia, dyslexia, palinopsia). She stopped reading at the beginning of the seizure and signaled to the nurses. She was able to explain that a seizure had started. Then, she had right-sided convulsions progressing to a generalized tonic-clonic seizure. Ictal EEG showed a rhythmic discharge in the left temporo-parieto-occipital junction electrode. Interictal high-resolution EEG highlighted the left occipito-temporal junction. Interictal PET demonstrated bilateral occipito-temporal hypometabolism with left-sided predominance. MRI fusion of the co-registered subtraction between ictal and interictal SPECT showed relative hyperperfusion affecting (1) the left occipito-parietal junction area, (2) the left lateral middle and inferior temporal gyri, and (3) the left inferior frontal area (see also pathophysiology).

A small number of reported cases do not fit the typical descriptions of either the myoclonic or the focal variant of reading epilepsy (07; 56; 12).

Prognosis and complications

The seizures in primary reading epilepsy are usually mild, and there is no deterioration in neurologic status or in seizure control. Furthermore, the response to medication is usually excellent, and seizures may also improve with modification of the triggering factors. However, primary reading epilepsy is probably lifelong (though some improvement may occur after the age of 40 to 50 years), and although the overall quality of life is not significantly influenced, reading-induced seizures may be a significant disability in those with continuing seizures in whom reading is an essential part of their life and profession.

Clinical vignette

Case 1: Reading epilepsy with jaw myoclonus. A 34-year-old woman with reading epilepsy manifested with jaw myoclonus starting at 15 years of age. Jaw myoclonus was consistently provoked by reading. Only once had a generalized tonic-clonic seizure occurred when she continued reading despite increasing jaw myoclonic jerks because she wanted to see what would happen. She did not have any other spontaneous or reflex seizures. The diagnosis of primary reading epilepsy was made at 26 years of age, and this was confirmed with EEG.

No additional seizures occurred in the next 8 years after treatment with clonazepam 0.5 mg nightly was initiated. Previous treatment with phenytoin was entirely ineffective. Her older sister also had similar symptoms of jaw myoclonus when involved in argumentative and fast talk. She has never asked medical advice for her condition and never had any other type of seizures. She gave the following written account of her symptoms: “I had jaw jerking but I can’t remember having one recently. It used to happen quite frequently when I was about 14, at school, and always when I was talking. I was talking quite rapidly at the time, and it was like a very quick uncontrollable spasm. It didn’t last long enough for anyone else to notice, but my flow of speech was disrupted.”

Case 2: Reading epilepsy with focal seizures of alexia. A 29-year-old right-handed man was referred at the age of 24 years with a 2-year history of infrequent nocturnal convulsive seizures for which he was treated with sodium valproate, without apparent therapeutic effect. He had also experienced episodes of alexia since the age of 22 years. These usually occurred after prolonged reading and invariably consisted of gradual loss of ability to recognize, first, the infrequently encountered letters, then the commonest ones like the letter “A,” and finally the numbers. These episodes would last for 1 to 2 minutes, and the ability to understand the reading material would resume in the reverse order. During the seizure, he could understand other people talking to him, but he was only able to respond basically. He described an associated slight dizziness and a feeling of discomfort, but he never experienced these symptoms or any other symptom, suggesting epileptic seizure activity when he was not reading. Episodes of paroxysmal alexia occurred almost daily and were worse when he was tired. None of them ever evolved into a generalized tonic-clonic seizure. Reading figures, talking, writing, playing cards or chess, and solving mathematical problems never provoked any symptoms. The patient had prolonged video-EEG studies during the alert resting state, multiple reading sessions, and all-night natural sleep. The EEG during the resting state, hyperventilation, and photic stimulation was normal. Reading consistently activated the EEG, resulting in frequent asymptomatic brief small spike-and-wave discharges over the left temporoparietal area, alternating with runs of low voltage regular slow activity at 3 to 4 Hz, lasting up to 4 seconds. One of the patient’s habitual seizures was recorded on video-EEG and lasted for 75 seconds. During the subsequent all-night natural sleep, spike-and-wave discharges were apparent over the left temporal area. Replacement of sodium valproate with carbamazepine 600 mg/day led to a considerable improvement, with only occasional episodes of reading-induced alexia over a follow-up period of 5 years. A single nocturnal generalized tonic-clonic seizure occurred due to poor compliance. Follow-up video-EEGs showed only occasional bursts of left-sided slow waves evoked by reading. Talking, writing, and solving mathematical problems never exerted any influence on the EEG, and generalized discharges were never recorded.

Biological basis

Etiology and pathogenesis

Primary reading epilepsy with jaw myoclonus. Primary reading epilepsy with jaw myoclonus is probably genetically determined though the mode of inheritance has not been elucidated. A family history of epilepsy is common, and familial reading epilepsy may occur. Autosomal dominant inheritance with incomplete penetrance overlapping with a genetic background for IGE was proposed for some families with primary reading epilepsy (25). Reading-induced jaw jerking has been described in paired cases of parent and child, suggesting dominant inheritance (37; 47), as well as between monozygotic twin pairs (17). Furthermore, a clear inheritance of spike-and-wave reflex activation during reading has been well documented in asymptomatic family members of patients with reading epilepsy (17). In the important meta-analysis by Wolf, 18 of 69 index patients (with reading epilepsy and available information on family history) had first-degree relatives with epilepsy or epileptic seizures of known type; 11 of them had reading epilepsy, 3 had idiopathic generalized epilepsies, 2 had febrile seizures, 1 had generalized tonic-clonic seizures since the age of 3 years, and 1 had symptomatic localization-related epilepsy (62). Radhakrishnan and colleagues found that 4 of the 18 patients with a known family history had first-degree relatives with epileptic seizures, and 1 had reading epilepsy (47). In the report of Koutroumanidis and colleagues, 3 of 15 patients with primary reading epilepsy and jaw myoclonus had siblings with reading epilepsy (31).

Valenti and colleagues studied a 3-generation family with 6 members having stuttering associated with EEG abnormalities triggered by language, 2 with clear juvenile myoclonic epilepsy (58). The family also had idiopathic generalized epilepsy, apparently inherited in a Mendelian pattern and associated with language-induced epilepsy and ictal stuttering related to the facial myoclonia typical of reading epilepsy. Monitoring showed both generalized EEG abnormalities and focal spikes with the facial myoclonia. The authors referred to the theory of nonuniform cortical hyperexcitability reviewed by Ferlazzo and colleagues (14) and concluded that language-induced epilepsy “could have more similarities with generalized epilepsies than with focal ones.”

Exceptional cases of symptomatic reading epilepsy due to cerebral lesions have been described (33; 50; 47; 08). However, these cases may have significant differences from the pure primary reading epilepsy as demonstrated by the case of Canevini and colleagues. Many years after a traumatic severe left frontal lesion, this patient developed nonfluent aphasia and facial myoclonic jerks triggered by reading, speaking, and listening to spoken language (08). This patient first noted at 57 years of age that he would begin to stutter when delivering lectures at conferences; the stuttering would worsen if he continued talking. The video-polygraphic EEG recording showed brief paroxysms of spikes and polyspikes, followed by a slow wave, more evident in the left fronto-temporal region. The myoclonic jerks originating from the submental area correlate with EEG abnormalities.

One of the 2 cases that Bickford and colleagues regarded as examples of secondary reading epilepsy was that of a woman whose troubles followed a probable cerebral infarct (06). The acts of calculation or of reading--or even simple recall--would all result in a jerking of the right arm. This was associated with bilaterally synchronous spike and wave discharge on the EEG.

The common precipitating mechanism in all patients with reading epilepsy is the formal act of reading, that is the transformation of the graphically displayed linguistic material into phonetic speech (audible or internal), which needs to be semantically understood (62; 64; 24). Cognitive neuropsychology has demonstrated the existence of at least 2 pathways for reading, the sublexical pathway involved in converting graphemes to phonemes, and the lexical pathway used when meaning is conveyed. Pegna and colleagues suggested that seizures triggered by sublexical (nonword) reading involve the left hemisphere preferentially, as compared to those triggered by lexical reading, which have bilateral EEG expression (46). The authors reported the case of a patient suffering from primary reading epilepsy in which the 2 routes were distinguished on the basis of the reading material employed. Significantly fewer epileptic discharges were observed when the patient read non-words than words. In view of their findings, they tentatively contrast a lexical form of primary reading epilepsy, triggered by the activation of semantic knowledge structures, with a sublexical form, triggered by non-word reading.

Primary focal reading epilepsy with alexia. The pathophysiological process of the primary focal reading epilepsy with alexia is probably straightforward. This is a pure reflex focal epilepsy with the stimulus (reading) directly acting on and disturbing the localized brain region subserving reading in the dominant occipito-temporal junction, which coordinates information that is gathered from visual and auditory processing and assigns meaning to the reading. These include the dominant supramarginal gyrus, which seems to be involved in phonological and articulatory processing of words, and the angular gyrus, which (together with the posterior cingulate gyrus) seems more involved in semantic processing. In this sense, primary reading epilepsy with alexia has a similar pathophysiology to the idiopathic occipital photosensitivity with the light activating the visual cortex. Gavaret and colleagues explained well and in detail that the pathophysiological process of primary reading epilepsy with alexia electively involves the brain network sustaining physiological reading (20). In all reported patients, the ictal discharge starts from the dominant posterior temporo-occipital junction, implicating the dominant angular gyrus and its connections with the associative visual cortex. In their patient, high-resolution EEG and PET revealed interictal abnormalities in the left temporo-occipital junction (20). Ictal video-EEG recording and ictal SPECT argued in favor of an initial left temporo-parieto-occipital junction implication with occipitotemporal predominance, propagating to more anterior areas of the left hemisphere. Ictal SPECT demonstrated a set of hyperperfused structures, including the left occipito-parietal junction area, left lateral middle and inferior temporal gyri, and a left middle frontal area. The hyperperfused structures highlighted by the ictal SPECT are close to the anatomy of reading in normal subjects. Furthermore, electrical stimulation of the posterior fusiform and inferior temporal gyri of the dominant hemisphere produced alexia, reading difficulties involving sentences and words, sparing letter by letter reading (35).

Myoclonic type of primary reading epilepsy. The pathophysiology of the myoclonic type of primary reading epilepsy is not clear. The relationship between the stimulus (reading) and the response (jaw myoclonus) is not immediately evident. How does specific cognitive stimulus (reading) evoke a motor (jaw myoclonic) response? Although the proprioceptive bombardment from the jaw or laryngeal muscles in consequence of reading has been shown in some patients to facilitate the occurrence of the myoclonic jerk (06), the sensorimotor cortex is not hyperexcitable as spontaneous jaw jerks never occur in this type of epilepsy.

Radhakrishnan and associates proposed that the interaction between hyperexcitable focal cortical areas and corticoreticular systems, under provocation by linguistic or other higher cognitive activities, may give rise to primarily generalized myoclonus and explain the EEG variability (47).

Ramani suggested that reading may elicit an abnormal response from a vulnerable link in the underlying neural systems and that the location of the defect at the input (visual registration), processing (meaning extraction), or output (motor expression) stage may determine the diverse clinical subtypes accordingly (49).

Koutroumanidis and colleagues suggested that the hyperexcitability in the myoclonic type of reading epilepsy should be conceptualized as a network or as regional, involving multiple, noncontiguous anatomically cortical areas in both hemispheres, rather than localized (31). They propose that ictogenesis in this type of reading (or language-induced) epilepsy is based on the reflex activation of a hyperexcitable network that subserves the function of speech and extends over multiple cerebral areas on both hemispheres. The parts of this network responding to the stimulus may, secondarily, drive the relative motor areas producing the typical regional myoclonus. This network hyperexcitability can be genetically determined, and its clinical expression is age-related. Thus, according to Koutroumanidis and colleagues, the myoclonic type of reading epilepsy is a system (language) epilepsy as also proposed by other authors (57; 65). This concept can explain the fact that precipitating stimuli include all aspects of language function. Evidence for the distribution of this network and for the abnormal activity within it in reading epilepsy has been obtained from functional imaging.

“System epilepsy” is a relatively new proposed concept with the hypothesis that some types of epilepsy reflect pathological expression (ictogenesis) of a specific brain system, which normally integrates activity that subserves normal physiological functions (03; 63).

Functional neuroimaging in the myoclonic type of reading epilepsy. Ictal SPECT with [99Tc]hexamethylpropylene amine oxime showed focal hyperperfusion of the frontal lobes bilaterally and of the left temporal area in a 14-year-old boy with a myoclonic type of reading epilepsy (40). Jaw jerks occurred every time he read an English textbook and sometimes during prolonged reading of a Japanese textbook. The jaw jerking evolved to generalized tonic-clonic seizures on only 2 occasions during prolonged reading aloud. Routine EEGs showed no abnormality. Ictal EEG showed bilateral 2-Hz, 150-µV spike-wave complexes with left frontotemporal accentuation. Interictal EEG, interictal SPECT, and MRI were normal.

One patient with reading epilepsy and normal brain MRI had EEG multifocal seizure onset bilaterally in temporal and frontocentral regions. 11C-diprenorphine (DPN) PET revealed peri-ictal opioid binding decreases in both temporal lobes and the left frontal lobe (28). The same group of authors investigated the release of endogenous opioids in 5 patients with reading-induced seizures compared with healthy controls (29). All subjects had 11C-diprenorphine PET scans while reading a string of symbols (baseline) or a scientific paper (activation). Statistical parametric mapping was used to find areas with differences in opioid-receptor binding. It was found that on activation scans mean 11C-DPN binding to opioid receptors was significantly lower (p< 0.05 corrected for multiple non-independent comparisons) in the left parieto-temporo-occipital cortex (Brodmann area 37) in patients with reading-epilepsy compared with controls. These findings suggest that opioid-like substances are involved in the termination of reading-induced seizures.

A right-handed 28-year-old woman started having spontaneous focal tonic seizures of the left leg, sometimes evolving to a generalized convulsion, at 17 years of age (32). She was seizure-free for 5 years while on primidone. At the 24 years of age, she began to have jaw jerking and groaning when reading regardless of the design of the letters that progressed to a generalized convulsion on 2 occasions. She did not have any history of epilepsy with any other precipitants, such as music or pictures. Neurologic examination, cranial CT and MRI were normal. The baseline EEG was unremarkable, whereas bilateral paroxysms of short spike and wave complexes appeared in association with typical jaw jerking, after about 30 minutes of reading a text. The ictal Tc-99m HMPAO brain SPECT showed hyperperfusion of the right superior temporal region whereas normal perfusion was seen in the same region in the interictal phase.

Archer and colleagues studied 2 patients with reading epilepsy with spike-triggered fMRI (02). The EEG was recorded inside the 3-tesla MRI while subjects read silently. Spike-triggered fMRI images were compared to baseline. In a second fMRI study, 30 seconds of silent reading was compared to visual fixation. Morphometry of these areas was assessed using curvilinear surface reconstruction. Left central sulcal patterns in 3 subjects with reading epilepsy were compared to 3 subjects with idiopathic generalized epilepsy and 12 normal controls. One subject with reading epilepsy showed spike-related activity (17 spikes) in the left precentral gyrus and bilaterally in the central sulcus and globus pallidus. The other showed no definite activation owing to low spike numbers (4 spikes). In both subjects, the block reading task recruited normal visual and language areas including the left posterior middle frontal gyrus. Two subjects with reading epilepsy showed an unusual gyrus branching anteriorly off the left central sulcus. A similar sulcal pattern was seen in none of the subjects with idiopathic generalized epilepsy and only 1 of 12 controls. The authors concluded that the spike activity overlapped with reading activity in the left middle frontal gyrus, a structure recruited during working memory cognitive tasks. They postulated that, because of a local structural anomaly, the spikes of reading epilepsy spread from working memory areas into adjacent motor cortex, activating a cortical subcortical circuit.

Salek-Haddadi and associates used combined EEG/EMG-fMRI methods to characterize the spatial relationship between activations related to language-induced seizure activity, language processing, and motor control in patients with reading epilepsy (53). They recorded and simultaneously monitored several physiological parameters (voice-recording, electromyography, ECG, EEG) during BOLD fMRI in 9 patients with reading epilepsy. Individually tailored language paradigms were used to induce and record habitual seizures inside the MRI scanner. Voxel-based morphometry was used for structural brain analysis. Reading-induced seizures occurred in 6 of 9 patients. One patient experienced abundant orofacial reflex myoclonia during silent reading in association with bilateral frontal or generalized epileptiform discharges. In an additional 5 patients, symptoms were only elicited while reading aloud with self-indicated events. Consistent activation patterns in response to reading-induced myoclonic seizures were observed within left motor and premotor areas in 5 of these 6 patients, in the left striatum (n = 4), in mesiotemporal/limbic areas (n = 4), in Brodmann area 47 (n = 3), and thalamus (n = 2). These BOLD activations were overlapping or adjacent to areas physiologically activated during language and facial motor tasks. No subtle structural abnormalities common to all patients were identified using VBM, but 1 patient had a left temporal ischemic lesion. Based on the findings, the authors hypothesize that reflex seizures occur in reading epilepsy when a critical mass of neurons is activated through a provoking stimulus within corticoreticular and corticocortical circuitry subserving normal functions.

A 45-year-old patient with reading epilepsy was studied using magnetic source imaging (43). The patient underwent 2 whole-head MEG recordings within 2 months while reading aloud. Forty-two spike wave discharges associated with jaw jerks were recorded and averaged with respect to spike wave discharges peak power. Epileptic discharges were then reconstructed using conventional equivalent current dipoles (ECDs) modeling, distributed sources sLORETA modeling, and beamforming approach. These methods identified 2 brain sources located in the left supplementary motor cortex and the left primary sensorimotor face area. The spatiotemporal pattern of the sources was characterized by a cross-talk between these 2 brain regions, with an initial source in the left supplementary motor cortex. This magnetic source imaging investigation suggests that reading epilepsy-induced jaw myoclonic jerks are associated with reading-induced activation of hyperexcitable neurons in the left supplementary motor cortex followed by secondarily propagation to the left primary sensorimotor face area producing the myoclonus.

Vaudano and colleagues analyzed ictal magnetoencephalography and EEG-correlated fMRI data in a unique patient with reading epilepsy affected by frequent perioral reflex myocloni triggered by reading silently (59). Ictal MEG corroborated EEG localization and revealed activity extending precentrally into Brodmann area 6. fMRI blood oxygen level-dependent (BOLD) signal changes in the left deep piriform cortex and left Brodmann area 6 preceded seizures and occurred before BOLD changes were observed in the thalamus and right inferior frontal gyrus (BA44). Dynamic causal modeling provided evidence of a causal link between hemodynamic changes in the left piriform cortex and reading-evoked seizures. The authors concluded that their findings support the important role of deep cortical and subcortical structures, in particular the frontal piriform cortex, as key regions in initiating and modulating seizure activity. In their patient with reading epilepsy, Brodmann area 6 appeared to be the area linking cognitive activation and seizure activity.

Safi and colleagues described a 42-year-old male patient with reading epilepsy evaluated with continuous video/EEG recordings through verbal, nonverbal, and reading tasks. Reading-induced seizures were significantly more numerous than those observed during verbal and nonverbal tasks. In reading tasks, spike frequency significantly increased with involvement of the phonological reading route. Spikes were recorded predominantly in left parasagittal regions (52). In a subsequent study the same patient was asked to read irregular words and nonwords presented in a block-design paradigm during magnetoencephalography, functional near-infrared spectroscopy, and functional magnetic resonance imaging recordings, all combined with EEG. Spike analyses from all these modalities data revealed an epileptic focus in the left precentral gyrus, and spike localization did not differ in lexical and phonological reading. The authors concluded that the somatosensory and motor control functions of the left precentral gyrus that are congruently involved in lexical as well as phonological reading can explain the identical spike localization in both reading pathways. The concurrence between findings in this study and those from the previous one supports the role of the left precentral gyrus in phonological output computation as well as seizure activity in a case of reading epilepsy (51).

Network hyperexcitability in a patient with focal reading epilepsy. In a report, the spatial-temporal characteristics of reading-induced epileptic spikes and hemodynamic activation in a Japanese patient with focal reading epilepsy have been delineated (18). MEG was recorded during silent letter-by-letter reading, and the source of reading-induced spikes was estimated using equivalent current dipole (ECD) analysis. Diffusion tractography was employed to determine if the white matter pathway connected spike initiation and termination sites. fMRI was employed to determine the spatial pattern of hemodynamic activation elicited by reading. In 91 spike events, ECDs were clustered in the left posterior basal temporal area during Katakana reading. In 8 of these 91 events, when the patient continued to read more than 30 minutes, another ECD cluster appeared in the left ventral precentral gyrus/frontal operculum with a time-difference of about 24 minutes. Probabilistic diffusion tractography revealed that the long segment of the arcuate fasciculus connected these 2 regions. fMRI conjunction analysis indicated that both Katakana and Kanji reading activated the left posterior basal temporal area, but Katakana activated the left lateral frontal areas more extensively than Kanji. Thus, prolonged reading of Katakana induced hyper-activation of the cortical network involved in normal language function, concurrently serving as the seizure onset and symptomatogenic zones. The authors concluded that reading epilepsy in their patient arose from a hyperexcitable network of cortical regions. Physiological activation of this network can have cumulative effects, resulting in greater reciprocal network propagation and electroclinical seizures (18).

In reading epilepsy, as in all reflex epilepsies, it remains a challenge to identify the tipping point, when normal physiological activities or sensory stimuli lead to recurrent extreme events (27; 26).

Epidemiology

Reading epilepsy is unusual but may be underdiagnosed. The incidence and prevalence are unknown but may be very low. In my experience of thousands of patients with epilepsies of all ages, I have seen 4 patients with the myoclonic and 3 patients with the alexia type of primary reading epilepsy.

Age at onset is usually 12 to 19 years with a peak in the late teens (15 to 18 years) that is long after reading skills have been acquired. Onset before the age of 12 years or after the age of 25 years is exceptional (62). There is a male preponderance of around 1.8:1 male to female ratio.

Differential diagnosis

Primary reading epilepsy may be underdiagnosed. If isolated jaw jerks occur without leading to limb myoclonic jerks, ictal disturbance of spoken language, or a generalized convulsion, the condition may not be recognized as a form of epilepsy. Isolated jaw jerks may be dismissed as a meaningless tic, and associated ictal language disturbance may be ascribed to stuttering (39) or to a movement disorder. Dyslexia or alexia progressing to generalized tonic-clonic seizures may be considered temporal lobe epilepsy if the association with reading is not appreciated. Secondary reading epilepsy, typically without jaw jerks, occurs in patients with triggered and spontaneous seizures and must be considered if there are abnormalities on examination or imaging or on the interictal EEG.

Diagnostic workup

By definition of an idiopathic epilepsy syndrome, all tests other than the EEG are normal (30). Rarely, reading-induced seizures may be caused by brain lesions (see symptomatic reading epilepsy).

For functional neuroimaging studies see the Pathophysiology section.

Interictal EEG. Interictal EEG is normal in more than half of patients. However, focal spikes or generalized spike-wave abnormalities likely have been reported in the other half. In a meta-analysis of 104 patients with reading epilepsy 11 (11%) had bilateral spike-and-wave discharges, 5 (5%) had paroxysmal temporal abnormalities, and 11 (11%) had photoparoxysmal discharges (62). Radhakrishnan and associates reported 7 patients (35%) with spontaneous bilateral spike-and-wave discharges and none with focal abnormalities (47). Koutroumanidis and associates found left-sided focal interictal spontaneous epileptiform abnormalities in 27% of their patients and symmetric generalized spike-and-wave discharges in 6%; no photoparoxysmal responses were observed (31).

Ictal EEG. Ictal EEG manifestations may be inconspicuous and difficult to detect in the myoclonic type of reading epilepsy because of muscle activity from the jaw muscles and head.

Ictal EEG changes show considerable heterogeneity in terms of discharge morphology and scalp topography. In the myoclonic variant, discharges are brief and consist of sharp waves, which are bilateral with left side preponderance in the temporoparietal regions.

In a meta-analysis of 73 ictal EEGs in primary reading epilepsy, 32% showed bilateral symmetrical discharges, 38% bilateral lateralized discharged, and 30% unilateral discharges (62). Lateralization to the dominant hemisphere was noticed in 78% and to the nondominant side in 10%, and it switched between sides in 12% of patients. Unilateral or bilateral regional discharges were more often temporal-parietal (80%) than frontocentral (20%).

Radhakrishnan and associates reported a significantly higher proportion of patients with generalized symmetrical ictal discharges (75%). Discharges were strictly lateralized to the dominant side in only 10% and were generalized with dominant side preponderance in 15% of patients (47).

In the study of Koutroumanidis and associates, brief spike/sharp-wave discharges or sharp theta wave discharges associated with reading-induced regional jerking were observed in 12 of their 15 patients with the myoclonic form of primary reading epilepsy (31). Five of these patients had a number of video-recorded or directly observed jaw jerks without noticeable electrographic changes. None of the jaw jerks in 3 patients were associated with EEG changes. In total, 8 of 15 patients had myoclonic jerks without EEG changes on at least 1 occasion. Regarding topography, clearly unilateral, ictal focal discharges were noted in 4 patients; they were all lateralized to the left side. In 1 of them, a previous EEG during reading was reported as showing brief generalized spike-wave discharges associated with jaw jerks. Bilateral synchronous discharges were observed in the remaining 8 patients. These were generalized in 4 patients, of fronto-temporo-central distribution in 2 patients, and both bisynchronous frontocentral and generalized in 2 patients. Bilateral synchronous EEG findings were symmetrical in 5 patients and lateralized to the left side in 2 patients and to the right side in 1 patient. Ictal EEG changes were not invariably associated with muscle potentials in surface electromyography from the jaw and submental region (31).

In the variant of reading epilepsy with alexia, ictal discharges are prolonged and entirely focal in the language-dominant temporo-occipital regions

(see case reports in Clinical manifestations section).

A patient with the myoclonic variant of reading epilepsy had ictal EEG with focal spike-slow wave in the left temporo-occipital location. A PET scan showed asymmetric FDG-uptake in 2 different locations in the left hemisphere; temporo-occipital and orbito-frontal. High resolution MRI was normal (42).

Clinical note. Studies of reading epilepsy suggest that increased task difficulty and increased complexity of different types or duration increase the chance of EEG or clinical activation. Therefore, the video-EEG should be performed while the patient is reading for a sufficient time, loudly and silently, certain texts that are likely, by history, to provoke the habitual reflex seizures of the patient. Otherwise the test may fail to provoke seizures.

Management

Modification of reading and talking habits may be successful (see Case 1 of the Clinical vignette section). Avoidance of prolonged reading and maneuvers that briefly disrupt comprehension or increase arousal may be helpful, but social and educational handicap may arise from these. Stopping reading when jaw jerking or alexia happen prevents generalized tonic-clonic seizures. Audiotaped texts may be useful.

Clonazepam 0.5 mg to 1 mg before sleep is highly effective in primary reading epilepsy with jaw myoclonus (45). Sodium valproate is also recommended by many authors, but this should be used in the unlikely event of clonazepam ineffectiveness, and it should be avoided in women of child bearing age. Carbamazepine may worsen jaw myoclonus. Conversely, carbamazepine is effective in reading epilepsy with alexia.

Levetiracetam has been found beneficial in reading epilepsy (58; 23).