The borderland of epilepsy: A clinical and molecular view, 100 years on

The diagnosis of phenomena that mimic epilepsy is anongoingchallengefortheclinicalneurologist.SirWilliamGowers published his classic book The Border-land ofEpilepsyin1907,basinghisanalysisanddifferentialdiag-noses solely on clinical information. He focused on faints,vagal and vasovagal attacks (which he differentiated fromfaints), migraine, vertigo, and some sleep symptoms,particularly narcolepsy. Over the last 100 years, majorinvestigative advances—including electroencephalogra-phy(EEG),imaging,and,morerecently,moleculargenet-ics—have revolutionized the clinical and basic sciencesrelevant to Gowers’ borderland. We have revisited theepilepsy borderland, from the clinical and molecularviewpoints, focusing on syncope, migraine, parasomnias,and some rarer disorders including the paroxysmal dyski-nesias and organic causes of nonepileptic convulsions(Crompton & Berkovic, 2009). Many of the basic clinicaltenets of Gowers’ analysis from a century ago remain,although the importance of vertigo as an epileptic mimichas diminished and the clinical approach to syncope hasbeen considerably modified. Interestingly, nonepilepticseizures (pseudo seizures) were not regarded as a signifi-cantclinicalissuebyGowers.A major advance in scientific thinking, made possibleby molecular genetics, is that the epilepsies are, at least inpart, afamily ofchannelopathies. Ion channeldysfunctionis the major known mechanism of the inherited idiopathicepilepsies, and roles in the acquired epilepsies (‘‘acquiredchannelopathies’’) are also likely. It is of interest to notethat when analyzing the mechanisms of disorders withinthe epilepsy borderland, ion channels also featurestrongly.The clinical distinction of epileptic seizures fromsyncope, particularly convulsive syncope, is a commonclinical problem which, just like in Gowers’ era, remainslargely based on the clinical history. In certain cardiacdisorders causing syncope, such as arrhythmias includingthe long QT syndrome and Brugada syndrome, the majormechanisms are of inherited cardiac ion channel dysfunc-tion. Remarkably, epileptic seizures and cardiac channel-opathies causing syncope rarely coexist. This is becauseion channel subunits present in the heart are generally notexpressedinthebrainandviceversa.However,themolec-ular and cellular mechanisms of electrical instability aresimilar, thus illustrating the parsimony of nature in termsofmechanismsofparoxysmaldisordersaffectingdifferentorgans.Migraine and epilepsy are usually easily distinguished,but a complex clinical overlap of migraine and epilepsydoes occur, as shown by epidemiologic studies, by analy-sis of certain families segregating both disorders, and byspecific but uncommon migraine–epilepsy syndromes.The molecular mechanisms of migraine are slowly beingunravelled, particularly in the autosomal dominantfamilial forms. Again, the genetic culprits are largely ionchannel subunits. Indeed, dysfunction of certain ion chan-nels can cause both migraine and epilepsy, although themolecular mechanisms, cortical-spreading depression inmigraine aura, and paroxysmal high frequency dischargesin epileptic seizures, are quite different. This paradox hasyettobefullyexplainedatamolecularlevel.Theparasomniasprovideanextremelychallengingareaof clinical differential diagnosis from nocturnal epilepticseizures. Moreover, there appears to be a genetic relation-ship between what are regarded as nocturnal seizures andparasomnias.Thediscoveryofacetylcholinereceptorsub-unit mutations in some families with autosomal dominantnocturnal frontal lobe epilepsy has provided a windowinto this issue, although the mechanism of seizure genesishas not been elucidated. Mouse models with humanmutations may show spontaneous epileptic seizures, butapparently nonepileptic phenomena can occur as well,furthercomplicatingthisfascinatingborderland.Paroxysmal dyskinesias are uncommon disorders,often familial, where the underlying molecular defects