• Autistic spectrum disorder;
  • Epilepsy


  1. Top of page
  2. Abstract
  3. Tuberous sclerosis (TS). An important model
  4. Issues of antiepileptic therapy

Summary:  The onset of epilepsy in brain systems involved in social communication and/or recognition of emotions can occasionally be the cause of autistic symptoms or may aggravate preexisting autistic symptoms. Knowing that cognitive and/or behavioral abnormalities can be the presenting and sometimes the only symptom of an epileptic disorder or can even be caused by paroxysmal EEG abnormalities without recognized seizures, the possibility that this may apply to autism has given rise to much debate. Epilepsy and/or epileptic EEG abnormalities are frequently associated with autistic disorders in children but this does not necessarily imply that they are the cause; great caution needs to be exercised before drawing any such conclusions. So far, there is no evidence that typical autism can be attributed to an epileptic disorder, even in those children with a history of regression after normal early development. Nevertheless, there are several early epilepsies (late infantile spasms, partial complex epilepsies, epilepsies with CSWS, early forms of Landau-Kleffner syndrome) and with different etiologies (tuberous sclerosis is an important model of these situations) in which a direct relationship between epilepsy and some features of autism may be suspected. In young children who primarily have language regression (and who may have autistic features) without evident cause, and in whom paroxysmal focal EEG abnormalities are also found, the possible direct role of epilepsy can only be evaluated in longitudinal studies.

After a period of 10–20 years during which there has been a serious hope or belief that some autistic children could have hidden epilepsy as the cause of their handicap, the excess of therapeutic enthusiasm, the disappointing results and growing “absence of evidence” of a direct causal link have definitively closed the issue in the opinion of many clinicians (1,2).

Although it is now clear that epilepsy is not a causal factor in children with typical “idiopathic” or “primary” or “nonsyndromatic” autism even in those who undergo a developmental regression (which is the case in about a third at around 18–24 months) (3–5), there are several situations in which the question of the role of epilepsy may, nevertheless, arise (6).

If one accepts the idea that epilepsy can present at any age with mainly or only cognitive and behavioral manifestations, sometimes with subtle seizures or even bioelectrical epileptic EEG abnormalities only (7), depending on which brain systems or networks are affected (and when the process starts), there is no reason to rule out the possibility that autistic symptoms can occasionally be the predominant epileptic manifestation. This point of view may encourage clinicians and researchers to keep an open mind with regard to the effects of diverse early epilepsies in which this can happen and may encourage them to evaluate the cognitive-behavioral profiles of affected children who appear to be within the autistic spectrum. The dynamics of development, the associated and often fluctuating clinical features (e.g., specific verbal language impairment) can sometimes be correlated with the activity of the epileptic disorder. The assessment of these children should include a study of how the observed symptoms resemble or differ at any point from primary autism.

Autism is a behavioral phenotype with many different possible etiologies, mostly genetic (8), although some early acquired brain disorders affecting the limbic system can sometime be the cause. The underlying brain dysfunction responsible for autism is often also the cause of epilepsy (or epileptiform EEG abnormalities), particularly in children who are autistic and mentally retarded as well (9). The consequence of this might be that, in most cases, epilepsy or epileptic EEG abnormalities are an independent manifestation of the underlying brain disorder (Table 1).

Table 1. Autistic disorder and epilepsy: reasons for cooccurrence
• Both conditions are totally independent
• The same brain pathology is at the origin of an autistic phenotype and epilepsy (e.g., Fragile X-syndrome, duplication 15q11–13 (9)
• An epileptic process occurring in early development interferes with the developing function of specific brain networks involved in communication and social behavior
• A focal (multifocal) brain pathology (i.e., tuberous sclerosis) that affects frontal or mesiotemporal structures (limbic system) can be at the origin of an autistic phenotype as well as the trigger of an epilepsy that aggravates the autistic symptoms
• An epileptic process causes a specific sensory or cognitive dysfunction with “autistic withdrawal” in a vulnerable child

In some children, however, the epilepsy originates or propagates in the abnormal brain systems or networks which are implicated in the autistic disorder. In these situations, epilepsy can aggravate the autistic symptoms or exceptionally, the functional epileptic disturbance in these systems can itself be the main cause of the behavioral regression in a child who was not previously autistic or only had mild symptoms. The autistic behavior can be due to repeated ictal-postictal states from seizures arising subcontinuously in the limbic or frontolimbic system or to other types of epilepsy-related impairment, especially if epilepsy starts or propagates in these systems at an early and crucial developmental phase. From the descriptive point of view, the term “autistic epileptic behavior” should be limited to those cases in which there is a clear contrast between the anomalies of behavior (alteration in communication, play) and other cognitive functions (i.e., visuo-spatial, etc). More often, the child has frequent seizures and fluctuating awareness, apathy, lack of initiative, and absence of social interests. In these situations, the blunting of all perceptual, cognitive, and relational functions usually corresponds to a more global deficit rather than corresponding to a purely autistic picture.

Rarely, autistic symptoms can be shown to develop or disappear in direct relationship to the activity of epilepsy (e.g., when there is remission of seizures accompanied by normalization of the EEG and rapid change in behavior as a result of medical or surgical antiepileptic therapy). When the onset and evolution of symptoms and the improvement is subtle and progressive (occasionally without clinical evidence of epilepsy and only paroxysmal EEG activity), which is the most frequent situation, it is difficult to evaluate the role of epilepsy “per se,” because the natural course of an autistic disorder and its associated deficits (language, etc.) can be very variable.

Despite all these difficulties it is occasionally possible to establish clear correlations between epilepsy and autistic behavior. The examples given below illustrate some of these situations. In each of them, a direct relationship between the autistic behavior and the epileptic activity has occasionally been claimed and in some cases has been well documented in individual longitudinal case studies.

(i) West syndrome (infantile spasms with hypsarhythmia), and particularly the late infantile spasms variant (onset 1–3 years) (10,11). In the latter cases the child may have developed a good level of communication and social behavior before seizure onset so that a regression and recovery in relation to the epilepsy can be directly correlated (Table 2).

Table 2. Epileptic spasms ("late infantile spasms") (1–3 years) and autistic regression
Regression occurs after some basic social and emotional competencies have already developed
The regression, recovery and the evolution of associated autistic characteristics can thus be documented
A mixture of generalized spasms and difficult-to-diagnose complex partial seizures (“epileptic stereotypies,” (12) may be seen often with predominant frontal epileptic foci

Several explanations, which are not mutually exclusive, have been proposed to explain why an autistic disorder might be observed in some of these children.

  • 1
    The constant epileptic activity (hypsarhyrhmia) precludes interpersonal communication and initiative.
  • 2
    A specific dysfunction of posterior cortical areas with complex visual problems (central visual impairment) can prevent the recognition of emotional cues essential for development.
  • 3
    A cortical dysfunction in the brain systems important for emotional and social aspects of cognition (frontotemporal) could lead to an autistic syndrome.

(ii) Complex partial seizures secondary to a developmental pathology (dysplasia) or to a congenital tumor in the temporal lobes. Successful surgical treatment in some of these children has occasionally resulted in a clear and documented postoperative improvement in autistic symptoms (13,14).

(iii) Partial epilepsy of prefrontal origin with continuous spike-waves during sleep on the EEG (CSWS) which occurs in somewhat older children. The disturbances of behavior (“frontal syndrome;”“epileptic psychosis”) have some features in common with autism, but usually without the fundamental communication and emotional disturbances (12,15).

(iv) Children with acquired epileptic aphasia (Landau-Kleffner syndrome) with very early onset in the first 2 years of life (16). In these cases, the children rarely or ever present with a typical “pure” auditory agnosia. They show several autistic features, either at onset or in the worst phase, in addition to absent or limited verbal comprehension and expression. The autistic features disappear as the child grows up or when effective antiepileptic therapy is given. These children may have a combination of language problems and autistic symptoms, with a changing predominance over time, but probably never, or only exceptionally, present the full picture of classical autism.

There are only two large series of children published in the literature (O'Reagan, 27 children (17); and Lewine, 50 children (18) with a regression in language and nonverbal communication considered to be of epileptic origin who improved with antiepileptic therapy, including surgery (18). O'Reagan (17) did not use the word autistic in her series (only “verbal and nonverbal communication”) whereas the children reported by Lewine all had an “autistic regression.” Both authors, however, made reference to the Landau-Kleffner syndrome as the closest clinical analogy with their cases. The children reported in these two studies were quite heterogeneous in terms of etiology and/or type of epilepsy syndromes. Some had paroxysmal EEG abnormalities only.

It does not appear to be justified to state that a child has a form of Landau-Kleffner syndrome, just because he or she has lost verbal language and has autistic symptoms as a consequence of the epilepsy.

Tuberous sclerosis (TS). An important model

  1. Top of page
  2. Abstract
  3. Tuberous sclerosis (TS). An important model
  4. Issues of antiepileptic therapy

Among the well-defined epileptogenic brain disorders frequently associated with autism, tuberous sclerosis deserves special mention. Recent studies have shown that it is not only the location of the tubers in the temporal lobe, but the local epileptic activity in these regions which is associated with autism. Bolton et al. (19) studied 53 children with TS and various neurodevelopmental disorders. Thirty-one had temporal lobe tubers. In those with autistic spectrum disorder (ASD) 17/19 had temporal lobe tubers whereas only 14/34 nonASD subjects had temporal lobe tubers. Of the 17 cases with temporal lobe tubers and ASD, all had a probable or possible temporal EEG focus and several had had a history of infantile spasms. In the literature on autism and tuberous sclerosis, there are several case reports in which the development of autism seemed to correlate with the onset and aggravation of the epilepsy, although this was not studied directly. In a longitudinal study of two children with TS, we demonstrated in one case a direct relationship between the epilepsy and the autistic disorder (20). Important new information could emerge from prospective studies to determine whether there is a direct correlation between clinical symptoms in such children with both the location of the tubers in temporal-limbic areas and epilepsy arising from these areas.

Issues of antiepileptic therapy

  1. Top of page
  2. Abstract
  3. Tuberous sclerosis (TS). An important model
  4. Issues of antiepileptic therapy

Table 3 lists the main clinical situations in which the cooccurrence of an autistic spectrum disorder and epileptic seizures (or paroxysmal EEG abnormalities) raises the possibility of a direct link and, in turn, issues of therapy.

Table 3. Autistic regression and epilepsy. Main clinical situations
(1) Autistic regression is the first symptom in a child in whom epilepsy can be clearly diagnosed when this diagnosis is sought (late spasms, complex stereotypes)
(2) Autistic regression occurs after repeated prolonged seizures (bilateral hippocampal sclerosis, 21)
(3) Epileptic EEG discharges are found during the diagnostic study of a child with an autistic syndrome with/without developmental regression and no history of seizures

(1) In the first instance, the diagnosis of epilepsy is clear and antiepileptic therapy is usually undertaken. In some of these children, seizures not only stop but there is a striking improvement in the autistic symptoms. This is an a posteriori clue of a direct causal role of the epileptic process. (2) In the second instance, the epilepsy has damaged crucial vulnerable areas (medial temporal system) and has created a possible permanent autistic syndrome (21). However, there may be potentially preventable additional damage from further episodes of status epilepticus. Probably only the most severe cases of this situation have been reported; there is probably a spectrum. (3) The last situation is the most difficult and the one which has given rise to much discussion. It should be evident that the mere presence of an abnormal epileptiform EEG is not in itself an indication for antiepileptic therapy because these EEG abnormalities are probably most often simply a marker of the underlying brain disorder responsible for the autism. Occasionally there are clinical hints that the dynamics of the disorder (stagnation, regression, fluctuations, curious developmental dissociations, acquired sleep problems, special EEG patterns such as focal sharp waves and continuous spike-waves during sleep) suggest a direct contributory role of these discharges. This often becomes clear only after a period of observation.


  1. Top of page
  2. Abstract
  3. Tuberous sclerosis (TS). An important model
  4. Issues of antiepileptic therapy
  • 1
    Kanner AM. Commentary: the treatment of seizure disorders and EEG abnormalities in children with autistic spectrum disorders: are we getting ahead ourselves? J. Autism Dev Disord 2000;30:4915.
  • 2
    Tharp BR. Epileptic encephalopathies and their relationship to developmental disorders: do spikes cause autism? Ment Retard Dev Disabil Res Rev 2004;10:132134.
  • 3
    Tuchman RF, Rapin I, Shinnar S. Autistics and dysphasic children. II. Pediatrics 1991;88:12191225.
  • 4
    Tuchman R, Rapin I. Regression in pervasive developmental disorders: seizures and epileptiform electroencephalographic correlates. Pediatrics 1997;99:5606.
  • 5
    Tuchman RF, Rapin I. Epilepsy in autism. Lancet Neurol 2002;1:3528.
  • 6
    McVicar KA, Ballaban-Gil K, Rapin I, et al. Epileptiform EEG abnormalities in children with language regression. Neurology 2005;65:12931.
  • 7
    Deonna T, Roulet-Perez E. Cognitive and behavioural disorders of epileptic origin in children. Clinics in developmental medicine No. 168. Mac Keith Press Cambridge Univ. Press. 2005.
  • 8
    Muhle R, Trentacoste SV, Rapin I. The genetics of autism. Pediatrics 2004;113:47286.
  • 9
    Gurrieri F, Battaglia A, Torrisi L, et al. Pervasive developmental disorder and epilepsy due to maternally derived duplication of 15q11-q13. Neurology 1999;52:16947.
  • 10
    Bednarek N, Motte J, Soufflet C, et al. Evidence for late infantile spasms. Epilepsia 1998;39:5560.
  • 11
    Deonna T, Ziegler AL, Maeder IM, et al. Reversible behavioural autistic-like regression: a manifestation of a special (new?) epileptic syndrome in a 28-month-old child. A 2-year longitudinal study. Neurocase 1995;1:919.
  • 12
    Deonna T, Fohlen M, Jalin C, et al. Epileptic stereotypies in children. In: GuerriniR, AicardiJ, AndermannF, HallettM, eds. Epilepsy and movement disorders. Cambridge: Cambridge University Press 2002;319332.
  • 13
    Neville BGR, Harkness WFJ, Cross JH, et al. Surgical treatment of severe autistic regression in childhood epilepsy. Pediatr Neurol 1997;16:13740.
  • 14
    Gillberg C, Schaumann H. Epilepsy presenting as infantile autism? Two case studies. Neuropediatrics 1983;14:20612.
  • 15
    Roulet-Perez E, Davidoff V, Despland PA, et al. Mental and behavioural deterioration of children with epilepsy and CSWS: acquired epileptic frontal syndrome. Dev Med Child Neurol 1993;2:26977.
  • 16
    Uldall P, Sahlholdt L, Alving J. Landau-Kleffner syndrome with onset at 18 months and an initial diagnosis of pervasive developmental disorder. Eur J Paediatr Neurol 2000;4:8186.
  • 17
    O'Reagan ME. Epileptic aphasia: a consequence of regional hypometabolic encephalopathy? Dev Med Child Neurol 1998;40:50816.
  • 18
    Lewine JD, Andrews R, Chez M, et al. Magnetoencephalographic patterns of epileptiform activity in children with regressive autism spectrum disorders. Pediatrics 1999;104(3 Pt 1):40518.
  • 19
    Bolton PF, Park RJ, Higgins JN, et al. Neuro-epileptic determinants of autism spectrum disorders in tuberous sclerosis complex. Brain 2002;125(Pt 6):124755.
  • 20
    Deonna T, Ziegler AL, Moura-Serra J, et al. Autistic regression in relation to limbic pathology and epilepsy: report of two cases. Dev Med Child Neurol 1993;35:16676.
  • 21
    DeLong GR, Heinz ER. The clinical syndrome of early life bilateral hippocampal sclerosis. Ann Neurol 1997;42:117.