If you feel a symptom of an event such as dizziness, light-headedness, faintness, seizures, shaking you will be instructed to push the event marker and give a detailed description of the events. Diary entries should be made every hour indicating activities such as eating, taking medication, sleeping, watching TV, etc.

During the testing period: You may do many quiet normal daily activities such as:. All equipment including EEG headbox, event marker, carrying case or backpack, and completed diaries must be returned. Your tech will go over all instructions at the time of your appointment. These pathophysiological issues are beyond the aim of this paper, in which we would rather focus on several controversial issues that make the interpretation of findings obtained by SD-EEG difficult even after more than 40 years of its use.

The results of published SD-EEG studies testing this method are difficult to compare to each other, mainly because of the protocols used and the population of patients assessed. Concerning the first issue, the protocol of SD total or partial , the length of the SD-EEG recording, the recording of drug-induced sleep, and the time of the day of the recording morning or afternoon constitute the main variables. Moreover, the inclusion and exclusion criteria in the published papers are very different concerning age children versus adults , seizure and epilepsy types, absence of abnormalities on basal EEG, neuroimaging, and treatment with antiepileptic drugs AEDs.

Again, even the definition of epileptic IIAs is not homogenous and, last but not least, most of the studies are retrospective, using Epilepsy Center databases.

The protocol of SD i. The time of the day of the recording morning, afternoon, or night is related to different circadian rhythms and mechanisms of sleep see as a review [ 18 ] and possibly to a different risk of occurrence of IIAs in different types of epilepsy. A residual postictal activation seems to be likely only when performing SD-EEG within days after seizure [ 6 , 23 ], but in the routine clinical practice this rarely occurs, and thus it is not as important as expected [ 9 ].

Another major issue varying from one study to another is the population studied. Even though the majority of the studies recruited patients undergoing a complete evaluation for suspected epilepsy, other inclusion and exclusion criteria are often not comparable, and in many cases it is not even possible to separate and analyze single variables. In fact, only few prospective series have been published [ 11 , 14 , 22 , 24 — 27 ], while most studies are retrospective using data from Epilepsy Center databases.

Even though age is considered critical for SD-EEG outcome, in some reports both adults and children have been included in the same group [ 5 , 6 , 8 , 10 , 16 , 21 , 24 , 25 , 28 — 34 ].

Moreover, although occurrence of IIA during the first routine wake EEG could be considered as an exclusion criterion in studies testing SD-EEG sensitivity, this aspect is sometimes not evaluated or not considered as a bias see, e. Concerning the classification of epilepsy , the oldest studies included patients with different types of seizure or syndrome, which were often only roughly classified [ 6 , 10 , 22 , 24 , 25 , 28 , 31 , 32 , 34 , 36 — 39 ], while some of the newer ones included populations more homogeneous concerning those aspects [ 16 , 17 , 40 — 44 ].

Therapy with AEDs is another aspect varying from one study to another. Only in a few studies SD-EEGs were performed in de novo patients which had never been treated with AEDs [ 6 , 8 , 9 , 14 , 20 , 23 ], while in most of the remaining ones, also patients taking AEDs were included, and therapy was left unchanged or at least AEDs tapering was performed thus probably significantly affecting occurrence of IIA and, thus, sensitivity and specificity of SD-EEG [ 47 — 49 ]. Furthermore, the number and type of AEDs were not described in detail in most papers see, e.

Another critical issue is the definition of IIAs, since usually EEG activation was defined by the occurrence of specific epileptic IIAs, but in some studies the authors considered also occurrence of slow waves [ 11 , 12 , 20 , 23 , 29 ], and, in a surprisingly high number of studies, the types of EEG abnormalities were not even specified [ 13 , 21 , 26 , 32 , 33 , 38 , 39 ].

Unfortunately, even in the few SD-EEG studies including such a protocol, a detailed analysis of the results was not reported, or, when IIAs occurrence during photic stimulation was listed, a correlation with seizure type s or syndrome was not provided see, e. Two important questions about the role of SD in inducing IIAs and thus in the diagnostics process of epilepsy are 1 whether the increased sensitivity of SD-EEG is due just to an effect of sampling or length of the recording, and 2 whether sleep per se or rather SD, indeed, induces activation of the EEG.

Some data suggested that EEG activation is already present during the waking phases of EEG recorded after SD [ 15 , 24 , 25 , 30 — 32 , 37 , 40 , 42 ]. However, in most cases, epileptiform discharges occurred more frequently during sleep [ 11 , 14 , 46 ], in particular during light sleep stages [ 8 , 9 , 19 , 22 , 23 ]. Some exceptions exist in which IIAs occurred both during wakefulness and sleep [ 5 , 34 , 35 , 37 ].

When directly compared to each other, spontaneous sleep seemed to increase significantly generalized discharges, while sleep occurring after SD might increase more focal discharges [ 39 ]. On the other hand, the comparison between SD-induced sleep and drug-induced sleep gave discordant results. Four studies [ 16 , 32 , 35 , 40 ] showed a similar yield in IIAs occurrence between the two approaches, while the other three ones [ 20 — 22 ] found a significantly higher activation rate in SD-EEG.

Given the dramatic differences highlighted above, the variability of sensitivity and specificity in single studies is not surprising. Some features of epilepsy seem to be more likely to be associated with IIAs occurrence during SD: activation seems to be greater shortly after epilepsy onset or seizures occurrence [ 23 , 25 , 31 , 40 ], in patients with an earlier seizure onset [ 19 ], and in those with a history of recurrent seizures [ 15 , 23 ].

Concerning the role of different types of syndrome or seizures, the first observation, by Mattson et al. The main limit of these studies was the lack, at that time, of an unanimously accepted seizure classification. In more recent observations, even though data are not consistent and numerous, we could state that, according to R. Degen and H. Degen [ 16 ], activation is more frequent in patients with complex partial seizures only, as compared to complex partial seizures plus other seizures types.

Concerning syndrome classification, a higher activation in idiopathic generalized epilepsy [ 9 , 11 ] and in particular in awakening grand mal and childhood absence epilepsy [ 19 , 32 ], has been reported. Neuroimaging data were available for few patients only in few recent SD EEG casistics [ 6 , 9 , 10 , 17 , 23 , 27 , 39 , 41 , 44 ]. Also, in our recent study [ 9 ], the diagnostic power for SD-EEG is not statistically different among subgroups of focal epilepsies.

Even more recently, in a retrospective study assessing the role of partial SD EEG in a wide population of patients assessed for suspected seizure, bearing a normal basal EEG and with a prolonged followup, we confirmed a high specificity rate Epilepsy is a complex disease, whose diagnosis is the results of the combination of anamnesis data and clinical history with diagnostic techniques, among which neuroimaging and EEG play a pivotal role.

The predictive value of the diagnostic exam is particularly crucial for epilepsy, especially in light of the burden of potential side effects of AEDs, which often need to be taken for years by patients, and of the stigma still surrounding the diagnosis of epilepsy. Thus, such a test should be very specific for epilepsy, and as sensitive as possible, in order to avoid the potential risk of not treating epileptic.

As described above, SD-EEG has been generally shown to bear a high specificity, and seems to be, thus, a good diagnostic tool. However, in the previous paragraphs we underscored the main difficulties in getting the full-blown potentiality of SD-EEG recording: these are mainly related to the huge methodological variability among the different studies in the field. Among them, the most relevant ones are represented by the striking differences in patients population and the SD-EEG protocols themselves, which often varies significantly from one centre to another.

Furthermore, many of the most important studies on SD-EEG and epilepsy date back to several decades ago, when neuroimaging data on the patients were lacking especially the nowadays routinely performed MRI data and some epilepsy syndromes subtypes had not been detailed yet. Of course, a prospective approach and an adequate amount of patients would be preferable; in any case, the detailed analysis of EEG IIAs occurrence should be performed by investigators rigidly blinded to the final diagnosis i.

The latter is a key requisite for avoiding the potential bias of over interpreting the role of a technique which, by itself, is considered crucial for the diagnosis itself, and is difficult to be ruled out in most of the existing studies on SD-EEG, which are retrospective in nature. A study bearing all of the features as above is, of course, impossible to be performed.

However, it would be already important to have studies fulfilling at least some of the abovequoted features; these should be multicenter to recruit as many patients as possible.

This is an open access article distributed under the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. List of Partners vendors. A sleep-deprived EEG, or an electroencephalogram , is a type of EEG that requires the patient to acquire less sleep than usual before undergoing the test. Like standard EEGs, this non-invasive test is used to record the electrical activity of the brain and can pick up on abnormal brain waves through electrodes attached to the scalp.

A standard EEG can detect seizures and diagnose epilepsy, but a sleep-deprived EEG may better detect more subtle seizures, like absence seizures or focal seizures. Learn about sleep-deprived EEGs, their purpose in diagnosing seizures, potential risks, and costs, and what to expect before, during, and after the testing is completed. The relationship between sleep and epilepsy has been studied for years.

The latter are abnormal electrical patterns that are characteristic of epilepsy and occur between clinical seizures. A board-certified neurologist may recommend a sleep-deprived EEG after a person with suspected seizures has had a standard EEG test that failed to show any unusual electrical activity.

Sleep deprivation can improve the accuracy of the diagnosis of epilepsy and increase the probability of detecting the characteristic electrical patterns known as epileptiform discharges.

Standard EEGs may detect many findings, including evidence of:. A sleep-deprived EEG further assesses changes in brain activity that can indicate various brain disorders, like epilepsy or other seizure disorders. A sleep-deprived EEG can be used to diagnose and differentiate various types of epilepsies. Sometimes seizure activity can manifest with psychiatric symptoms. Therefore, in some psychiatric presentations, a sleep-deprived EEG may be ordered by your healthcare provider to identify abnormalities that are typically seen with seizures.

The amount of sleep the person obtains the night before, the duration of the EEG, and the time of day the examination is administered are not specific to the test.

These may contribute to some differences in the results at times encountered when comparing studies done at different institutions. A sleep-deprived EEG is safe, painless, and poses no significant risk. Most people experience little or no discomfort during an EEG. Remember, the electrodes do not transmit electrical charges, they only pick up electrical activity from the brain itself. Like in alternative activation procedures involving photic stimulation fast, flashing lights or patterns or hyperventilation very quick breathing , sleep deprivation can trigger a seizure during the exam.

If you are undergoing a sleep-deprived EEG, you will be carefully monitored throughout the procedure. In case you have a seizure, which is a possibility among those who are predisposed to this condition and thus undergoing the testing, you would be treated with a fast-acting anti-seizure medication immediately. If the seizure is prolonged, as would occur in a condition called status epilepticus, oxygen and the appropriate safety equipment are kept nearby the monitoring room and a protocol would be followed to quickly end the disturbance.

A standard EEG procedure can be about one hour to an hour and a half, with time spent applying the electrode wires and a 20 to 40 minute period for recording brain activity, while the sleep-deprived EEG procedure usually takes a few hours. The recording will continue while that patient is falling asleep or dozing. Once the test is over and the patient wakes up, they can immediately go home.

Prior to the test, the healthcare provider will ask you to sleep less or avoid sleep completely the night before the test. It is likely that your healthcare provider may have you go in for the sleep-deprived EEG early in the morning in order to ensure you are drowsy during the test and don't accidentally fall asleep during the day.

A sleep-deprived EEG is typically an outpatient procedure, meaning that it occurs without the need for hospitalization. In some cases, a sleep-deprived EEG may occur as part of longer video EEG monitoring on an epilepsy monitoring unit in a hospital. Because you are likely to be drowsy during and after the sleep-deprived EEG, it is in your best interest to arrange for someone else to drive you to and from the testing. Because you will be seated or lying down during the exam, you should wear something comfortable.

A top that buttons or zips up is advised, so you don't have to pull anything over your head. Jewelry is permitted, but keep in mind that large or dangling earrings could get in the way depending on where the electrodes are placed. Stam ]. A PLI value for each pair of electrodes was computed for each epoch and the average PLI over four epochs was used for further analysis. The PLI was calculated separately for the following frequency bands: delta 0.

The PLI is a measure that is less sensitive for confounding than other functional connectivity measures and quantifies the phase coupling between two time series as a value between 0 and 1 Stam et al. The synchronization between time series is based on the consistency of the nonzero phase lag of one time series with the other. To compute the instantaneous phase difference for each time sample we used the analytical signal concept and the Hilbert transform.

During the calculation the influence of volume conduction is diminished by disregarding phase differences of zero. A more detailed description of the calculation and specifications of PLI can be found elsewhere Stam et al. Global network properties were quantified via weighted clustering coefficient and path length. We repeated the construction of functional networks and analysis for different frequency ranges.

The clustering coefficient is a measure of degree to which nodes in a graph tend to cluster together. We used the weighted clustering coefficient C as described in Grindrod , Higham et al.

These conditions are fulfilled when using PLI as weight definition. The weighted C of node i is then defined as. In the case in which w ij equals either 0 or 1 the definition was equivalent to the definition for unweighted graphs Watts and Strogatz, The mean weighted clustering coefficient was defined as.

For the weighted path length, the path between two nodes i and j is found by minimizing the sum of weights assigned to the edges on their path. The average path length L for node i to all other nodes is defined as. We considered high values of the PLI as close functional distance and low values of the synchronization index as large functional distance i. In our dataset no disconnected nodes were present. The mean weighted path was defined as. The MST creates a unique network based on the weighted connections between nodes.

Furthermore, it connects all the nodes in the network without forming cycles and thereby reduces the connection costs. Thus, if adding an edge results in formation of a cycle within the network, this edge will be skipped. Given the number of electrodes included in our analysis, trees contained 17 nodes and 16 edges.

From these trees, several measures can be quantified Boersma et al. For our study, we focused on the two most straightforward MST measures: leaf number and diameter that give information on topological features of trees Figure 2. The leaf number presents an upper bound to the diameter of the spanning tree that is the largest distance between any possible pair of nodes of the tree. Both leaf number and diameter were normalized between 0 and 1.

In summary, by excluding less important connections in the network, the MST network is built from the most efficient connections and enables a direct comparison between two networks since the number of nodes and connections are similar Figure 1. Figure 1. Two schematic illustrations of networks. A a standard network and B a MST network wherein all nodes are connected only once resulting in a loopless network.

In panel B the black lines represent the most efficient connections in the MST network; grey lines represent the excluded functional connections. First, we explored the effect of sleep deprivation in each group separately by comparing relative power spectra, network and MST measures from routine EEG and SD-EEG recordings for each frequency band with a paired t -test.

To investigate whether a different network alteration was observed after sleep deprivation in patients compared with controls, we used a repeated measures analysis of variance ANOVA with type of EEG EEG vs.

SD-EEG as within factor and group patients vs. Differences between groups in age and gender were explored using independent t -test and chi-square test respectively. To correct for multiple comparison, we used false discovery rate correction per frequency band as each frequency band is associated with distinct network and functions Basar et al.

All analyses were performed in SPSS. A p -value below 0. In total, 21 children with a definitive diagnosis of focal epilepsy were included in this study 5 girls, mean age Clinical details are summarized in Table 1. The control group consisted of seventeen children 7 girls, mean age Paired t -tests were performed for relative power spectra per frequency band for patients and controls separately. Table 2. Paired t -tests were performed for each network and MST characteristic per frequency band for patients and controls separately Tables 3A , B respectively.

In the other frequency bands, no significant differences in network measures were found after sleep deprivation. Table 3A. Table 3B. Table 4.

An EEG after sleep deprivation is often performed in patients suspected of epilepsy when the standard EEG recording is inconclusive. Little is known, however, about the mechanism behind the increased sensitivity of SD-EEG recordings in epilepsy.

In this study we investigated whether a network analytical approach could clarify this phenomenon. Our repeated measures analysis for leaf number and diameter suggested an interaction between patients and controls after sleep deprivation: patients showed a shift toward a more path-like MST network whereas controls showed a shift toward a more star-like MST network Figure 2. This shift difference was more pronounced when including only the patients in whom the SD-EEG recording was of additional clinical value Table 4.

Together with the increased presence of epileptiform abnormalities after sleep deprivation in patients, the network shift toward a more path-like MST network could possibly reflect an inadequate compensatory mechanism of the epileptic brain, although this requires to be confirmed using data from future studies with larger sample sizes. Figure 2. Three network topologies based on MST network. On the left a path-like topology with few leafs and long diameter; on the right a star-like topology with many leaves and a moderate diameter.

In the middle an intermediate form combining the qualities of a line-like and star-like topology. Notify that all networks have the same number of nodes and connections. Leafs colored in green. Figure 3. Illustration of interaction effects from MST measures diameter left graph and leaf number right graph per frequency band as revealed with a repeated measures ANOVA mean values and standard error of the mean bars.

There was a significant interaction for diameter in the alpha band; the diameter increased in patients whereas an opposite effect was found for controls. For leaf number, a significant interaction was found in the alpha band; the leaf number decreased in patients whereas an opposite effect was found for controls. Together, these results in the alpha band suggest a shift toward a path-like topology for patients after sleep deprivation and a shift toward a star-like topology for controls.

Previous research has shown that functional networks change during seizure generation and propagation into a more regular network organization i. Furthermore, long-term continuous evaluation of functional networks derived from intracranial recordings, revealed large fluctuations in clustering coefficient and path length during the day Kuhnert et al.

These fluctuations over time, largely attributed to daily rhythms, showed an increased regularization of functional networks during night-time in patients with focal epilepsy Kuhnert et al. Possibly, this shift toward a more regular network during sleep in patients with epilepsy explains why the epileptic brain is more susceptible to both epileptiform discharges and seizures during sleep.

However, this remains speculative as we cannot infer a causal relation between network alteration and an increased presence of interepileptic discharges after sleep deprivation based on our results.

We did not investigate network alterations during sleep, but a similar mechanism could explain the increased presence of epileptiform discharges after sleep deprivation. This study suggests that the network organization shift toward a more path-like topology in patients with epilepsy i.

Considering the path-like network as a network wherein nodes are less centrally connected and share basic characteristics with a regular network, it might be possible that the mechanisms underlying sleep deprivation-induced network alterations mimic the changes of a functional network during the ictal state or during sleep.