Seize the moment that is thine: how should we define seizures?

‘You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time.’ – John Lydgate

This can be aptly applied to definitions and classifications in neurology and perhaps no more so than the definition and classification of the epilepsies. The International League Against Epilepsy (ILAE) has recently revised the definition of epilepsy (Fisher et al., 2014). Epilepsy remains the propensity to have seizures but the criteria for a ‘propensity’ are now carefully defined. The ILAE has also proposed a reclassification of the epilepsies, and again the general scheme is not dissimilar to previous classifications (Berg et al., 2010). Although these recent revisions build upon previously accepted classifications, they have generated a considerable amount of debate, illustrating the emotive impact of classification and definition in epilepsy.

Underpinning all of the above is a generally accepted definition of seizures as a ‘transient occurrence of signs and/or symptoms due to abnormal excessive or synchronous neuronal activity of the brain’ (Fisher et al., 2005). Despite advances in our understanding of network dynamics underlying seizures, this definition has changed little from the 1873 definition of Hughlings Jackson who referred to seizures as ‘the name for occasional, sudden, excessive, rapid, and local discharges of grey matter’. The lack of change is even more surprising given that Hughlings Jackson’s definition predated EEG; indeed, Hughlings Jackson envisaged seizures as a chemical ‘explosion’ rather than an electrical storm.

However, that the definition has not changed greatly does not imply that it has some intrinsic fault. So how useful and accurate is the present definition of seizures? We propose that the present definition may be inadequate to define seizures in humans, but is even more problematic with regards to defining seizures in basic science research.

The essential elements of the present definition are: (i) a transient state; (ii) the occurrence of signs and/or symptoms; and (iii) excessive or synchronous neuronal activity of the brain. Seizures are transient but so are almost all brain states, although over differing timescales. We will therefore first consider the occurrence of signs and/or symptoms. Symptoms are, by definition, peculiar to humans so in order to broaden the definition to all animals we should add behavioural change. Notwithstanding this, the concept of some clinical change remains problematic. To illustrate these conceptual difficulties, we will consider the following thought experiment. Imagine someone is having simple motor seizures manifesting only as right hand twitching. If we now amputate the person’s hand, then there is no clinical sign or symptom and so by definition the person is no longer having seizures despite having identical abnormal brain activity. If we were to accept that the abnormal electrical activity that continues to occur despite removing the hand is seizure activity, then would this still be seizure activity if we resected that part of the brain and kept the abnormal neurons and networks operational through adequate oxygenation and nutrition? It is a simple step from this form of reductionism to propose that similar activity generated in a brain slice in vitro is indeed seizure activity.

A separate but related challenge is that the detection of a clinical change depends on how hard we look. For example, careful monitoring of reaction time during a spike-wave discharge may detect a hesitation that would not normally be apparent to patient or observer. Similarly, polyspike-wave that occurs in motor cortex may elicit a jerk, whereas the same electrical activity in prefrontal cortex may elicit no clinical change and so be classified as interictal. This also illustrates the gamut of EEG activities that can be considered ictal from the typical seizure with a build-up of fast activity evolving to repetitive spikes to a sudden burst of spikes/polyspikes.

Furthermore, subclinical seizure activity could, in the longer time, have effects on the brain, perhaps manifesting as such problems as cognitive decline or psychiatric disease—that this activity does not have immediate symptoms or signs would mean that it would not be defined as seizure activity, yet it has some clinical impact. The necessity of having symptoms and/or signs in the definition is indispensable for defining a clinical seizure, but leaves subclinical seizures poorly defined.

The other aspect of the definition that is troublesome is the attribution of clinical change to ‘abnormal excessive or synchronous neuronal activity of the brain’. First, the clinical change may be due to inhibition of a brain area rather than activation. Second, the dynamics of neuronal activity during a seizure are complex with some areas showing heterogeneous changes in neuronal firing (some neurons having decreased firing) (Truccolo et al., 2011); though, to what extent these areas represent ‘penumbrial’ regions is still a matter for debate (Weiss et al., 2013).

Undoubtedly, with advances in our understanding of brain networks and mathematical descriptions of brain states and activity (Jirsa et al., 2014), we should somehow improve our definition of seizures and seizure activity. However, the desire to have a unifying definition that satisfies all could be detrimental to having a satisfactory definition for anyone and is, to our mind, a pipe dream. For example, removing the clinical aspect from the definition would, for most people, remove the essence of a seizure, yet keeping that aspect as a sine qua non hinders basic science research, which abounds with seizure activity without clinical change. We could try to be inclusive by stipulating that a seizure is ‘transient abnormal network/neuronal activity that can result in symptoms, signs or behavioural change’. Using this definition, abnormal activity in a brain slice could be considered a seizure because if it were to occur in a specific area of the brain in situ then a clinical change would occur. However, such compromise diminishes the definition, and would probably be unacceptable clinically. In the end, we have to accept that the definition cannot ‘please all of the people all of the time’, that separate and careful definitions will probably be needed, and that these definitions will likely evolve as our understanding of seizures and seizure activity advances.