The recurrence of unprovoked epileptic seizures is the decisive factor in the definition of epilepsy (1). However, due to their relatively unpredictable occurrence, recording them in a routine electroencephalogram (EEG) is not trivial and often reserved for a long-term EEG or a stay at an EEG monitoring unit. Interictal epileptiform discharges (IED), on the other hand, are more common and therefore easier to detect transients of electrical cortical activity of different duration and configuration. Their strong association with epilepsy and their occurrence in the interictal interval, i.e., between two successive seizures, have made IEDs a biomarker of epileptic activity and a powerful screening instrument for identifying and classifying epilepsies (2). The brief variants, the so-called “spikes,” are defined as transients with a generally negative main component, a characteristic “sharp peak at conventional paper speed with a duration of 20 to less than 70 ms, and a variable amplitude of >50 μV,” which if followed by a slower wave of the same polarity, are called “spike–wave complexes.” Because of the arbitrary and mostly descriptive distinction between spikes and transients of similar characteristics with a slower time course up to 200 ms, the so-called “sharp waves,” these transients have been grouped together as IEDs (3, 4). Beyond their diagnostic and localizing value, it remains unclear what their exact role is within epilepsies, epileptic seizures, and epileptogenesis, and more fundamental, if their presence represents an excitatory or inhibitory activity within an epileptic network (5).

The significance of their frequency remains unclear. In addition to diurnal fluctuations, conflicting findings on IED frequency in the pre- and postictal interval have been reported (5, 6). While single IED and short sequences of epileptic activity can occur without any recognizable clinical symptoms, short series of generalized spike-and-wave complexes can be associated with impaired awareness and reaction as in absence epilepsy. Furthermore, IED are not exclusively an electrographic phenomenon, but are associated with impaired cerebral perfusion, metabolic changes and can affect cognitive performance (4).

Typically, IEDs are analyzed in terms of their shape, distribution, and frequency. While the first two usually require the averaging of many potentials of similar shape and distribution, the latter compare, e.g., their relative incidence in different states of vigilance, i.e., wakefulness and sleep, in pre- and postictal episodes, and under different drug conditions. Concerning their association with the efficacy of anti-seizure medication, conflicting findings have been reported with both, an increase and decrease of IED frequency under medication (5). Therefore, we designed this study to investigate the association of IED frequency in continuous EEG and the serum concentration of levetiracetam (LEV) and lamotrigine (LTG), two widely used newer agents of anti-seizure medication (ASM) with sequential serum samples over the course of several days.

We included 56 patients who were referred to our center for a one-week video EEG monitoring between 2018 and 2020. All patients received a continuous EEG with scalp electrodes according to the 10–20 system (7) for either diagnostic reasons or for a presurgical workup. We included people with confirmed epilepsy if they were at least 18 years old, took either ASM and were admitted to our monitoring unit with the intention of recording epileptic seizures and/or IED for diagnostic purposes, sometimes by reducing or discontinuing medication. Exclusion criteria encompassed pregnant or breastfeeding women, minors and people under legal guardianship or limited legal capacity.

For both ASM, LEV and LTG, the course of IED frequency on up to 9 days of consecutive recording showed a clear increase to 4.5 and 15.5-fold, respectively between at the day of dose reduction or ASM cessation, and a decrease after re-dosing or reintroduction of the drugs. In fact, IED frequency and serum concentration showed a negative correlation for both substances. This suggests suppressive properties of LEV and LTG against IED. But a review of the literature on this association points to a more complex interaction with evidence for a possible underlying inhibitory activity.

The scientific evidence on IED gathered over the past decades is heterogeneous, with conflicting findings regarding their frequency and association with epileptic seizures and epileptogenesis. Gotman reported that the frequency of IED did not change systematically leading up to epileptic seizures (8). Moreover, he observed an increase in their frequency in the hours and days following epileptic seizures but not after ASM reduction (6). The latter might be explained in parts by the fact, that patients in this similarly sized cohort were not recorded throughout the day, nor following a similar protocol, and did not receive neither LEV nor LTG. In contrast, we obtained the IED frequency from continuous 24-h recordings for all patients and registered an increase of IED frequency following ASM reduction or discontinuation, regardless of the occurrence of epileptic seizures. Our findings are consistent with a more recent study based on continuous recordings, which showed an increase after ASM withdrawal (9). Goncharova et al. confirmed Gotman’s observation of increased frequency with the occurrence of an epileptic seizure in continuous recordings averaging 9 days, which we did not observe in our data (10). However, they reported a decrease in IED frequency with ASM reduction leading up to epileptic seizures. This observation differs from our results and those of De Stefano and may be explained in part by the fact that Goncharova et al. recorded a higher median IED frequency of more than 600 per day. In contrast, we registered a median IED frequency of 367 and 275 per day for patients under LEV and LTG, respectively (Supplementary Figure S2), which may indicate a better seizure control in our cohort or a relevant influence of other dynamics.

Interestingly, ASM serum concentration not dose showed a significant negative correlation with IED frequency for both drugs (Figures 1B,C, 2B,C), although dose correlated significantly with serum concentration for both substances (Figures 1D, 2D). In our study, LTG dose did not correlate significantly with IED frequency, but serum concentration (Figures 2B,C), which could be explained by the slower absorption of LTG (5–6 days to steady-state vs. 2 days LEV). However, it remains unclear if the changes of IED frequency following reduced ASM concentration are of pathophysiological significance. While the suppressive effect of ASM on seizures is well established, the effect of ASM on IED is heterogeneous. Oftentimes the cortical region that generates IED (the so-called irritative zone) does not coincide with the ictal onset zone, and it is subject to an ongoing discussion whether IED originate from an excitatory or inhibitory activity. Although, it is believed that interictal spiking sets a condition that preludes the onset of ictal discharges, they should not be interpreted as pre-ictal phenomena since they are generated by different populations of neurons through different cellular and network mechanisms (5). Furthermore, the strong after-inhibition produced by IED maintains a low level of excitation after an initial neuronal bursting (4), which seems to protect against the occurrence of ictal discharges in a general condition of hyperexcitability determined by the primary epileptogenic dysfunction (5). Combined with the early observations of Gotman, that IED frequency decreases in the run-up to epileptic seizures, the periodicity of IED can be interpreted as sign of focal inhibitory processes opposing the epileptogenic activity (Lebovitz 1979). More interestingly, Goncharova found a higher frequency of IED to be associated with a longer time to the first seizure, which suggests that IED are a marker of inhibition, rather than excitation (Goncharova 2016). In this sense, our finding of an increased IED frequency after reduced or discontinued ASM can be interpreted as an increase of inhibitory network activity (rescue attempt) due to an increased epileptogenic activity. However, this might be different for generalized epilepsies, where LEV reduced not only spike–wave density (spikes per hour, i.e., frequency) and but also duration of generalized IED in refractory IGE (13). Due to the low number of generalized epilepsies, we could not observe a difference for type of epilepsy in our study. Furthermore, the fact that subclinical seizure patterns are frequently observed in continuous invasive EEG blurs the boundaries between interictal and ictal events and might explain the heterogenous observations of the last decades.

Furthermore, it is well established that the IED frequency is also subject to circadian fluctuations with a robust and uniform increase during sleep (14). More precisely, IED frequency depends on the degree of wakefulness or rather sleep stages. In a meta analysis Ng and Pavlova showed that relative to REM sleep (1.0), the IED rate differed significantly between wakefulness (1.11) and NREM, with an increase of 75 percent in N1, 69 percent in N2 and 2.46 times higher rates in N3 (11). In addition to circadian changes, analyzing ultralong recording durations of up to 10 years, revealed that IED frequencies fluctuate with ultradian (12 h) but especially with even slower multidian (10 and 26d) periods (12). However, the changes we observed in our study following dose reductions were of a different order of magnitude (mean increase for LEV ~4 and LTG 15 times higher) and consistent over several days of dose reduction or drug discontinuation (Figures 1A, 2A) and are therefore not sufficiently explained by fluctuations alone. Thus, our study highlights the importance of continuous EEG monitoring in capturing fluctuations in IED frequency, which may not be fully explained by circadian or ultradian rhythms alone. These insights shed light on the complex interplay between ASM treatment, IED dynamics, and epileptic activity, and potentially paving the way for further research into more personalized epilepsy management strategies.

In the light of other valuable research our findings contribute to the understanding of ASM effects on IED frequency dynamics. The negative correlation suggests suppressive properties of both medications. However, considering their well-documented suppressive effect against epileptic seizures, and the indications of an inhibitory property of IED, the reduction of IED frequency following ASM administration may reflect their seizure-suppressive effect in epilepsy, which would render an inhibitory network activity less important, rather than a suppressive effect against IED themselves.