This study will evaluate how the state of being completely deprived of sleep has an effect
on recordings of magnetoencephalography (MEG) and electroencephalography (EEG), in relation
to how alert someone is and how sleepy someone perceives himself or herself to be. EEG
measures electronic potential differences on the scalp. On the other hand, MEG is a
non-invasive technique for recording the activity of neurons in the brain, through recording
of magnetic fields caused by synchronized neural currents. It has the ability to detect
seizures. Because magnetic signals of the brain vary, this technique must balance two key
problems: weakness of the signal and strength of the noise. The EEG is sensitive to
extra-cellular volume currents, whereas the MEG primarily registers intra-cellular currents.
Because electrical fields are quite dependent on the conductive properties of the tissues,
and magnetic fields are significantly less distorted by tissue, the MEG has better spatial
resolution. There is a great deal of evidence that EEG and MEG provide complementary data
about underlying currents of ions.
The complex relationship of sleep and epilepsy is well known. Sleep has been used for many
years as a powerful EEG activator. Many researchers have supported the hypothesis that there
is a specific activating effect of sleep deprivation on epileptic discharges. Sleep
deprivation is defined as a sleepless state of longer than 24 hours. The increased use of
MEG in diagnosis could improve the procedure for evaluating patients before surgery for
epilepsy, by making invasive studies less necessary.
Patients 18 years of age or older, with a diagnosis of epilepsy and with a documented last
routine EEG (at least 2 weeks earlier) and routine EEG on the day of a baseline MEG-EEG
without interictal epileptiform discharges (IEDs) may be eligible for this study.
Participants will be rated according to the Epworth, Stanford, and Karolinska Sleepiness
Scales, to determine their subjective sleepiness. They will be randomly assigned to stay
awake all night or sleep in the hospital overnight. That is, a sleep deprivation and
non-sleep deprivation synchronized MEG-EEG recording will be performed in random order. Then
the sequence of sleep deprivation and non-sleep deprivation will be reversed within 14 to 21
days. During the recordings, the patient will either sit or lie with his or her head in a
helmet covering the entire head, with openings for the eyes and ears. Brain magnetic fields
will be recorded with a 275-channel OMEGA system. Throughout the session, visual and two-way
audio communication will be maintained with the patient. Recording sessions will last 90 to
180 minutes, with the patient allowed to take breaks after at least 10 minutes in a scanner.
Attempts will be made to encourage patients to stay awake and sleep for about the same
amount of time during each recording, to acquire comparable amounts of sleep and awake
We would like to evaluate the activating effects of complete sleep deprivation (SD) on
synchronized MEG-EEG recordings, and on each of the components singly, in relation to the
degree of alertness during recording (awake vs. sleep) and the subjective degree of
sleepiness as assessed by standardized scales. We postulate that acute SD will increase the
diagnostic yield of synchronized MEG-EEG and activate both modalities (MEG, EEG) to the same
degree. The increased diagnostic utility of MEG could improve the epilepsy surgery
evaluation procedure for many patients by rendering invasive studies less necessary. The
medical and economic utilization of such expensive resources as MEG could thus be
Participants of this study will be epilepsy patients whose last routine interictal EEG
(performed at least two weeks earlier), subsequent pre-screening EEG and screening MEG-EEG
show no interictal epileptiform discharges (IEDs), and are therefore considered
We will use a 275-channel Whole-head MEG System (CTF Systems 2001 Inc.). Patients will have
a screening, non-SD and SD MEG-EEG after their degree of sleepiness is assessed using the
Epworth, Stanford and Karnolinska sleepiness scales. Starting one day after the MEG-EEG, SD
and non-SDMEG-EEG will be performed in random order within 14-21 days of each other. This
will ensure an equal amount of sampling effect in SD and non-SD data sets. The MEG-EEG
session will last 90 to 180 minutes. Patients may take a break after at least 10 minutes in
a scanner. We will attempt to record a comparable amount of awake and sleep data. At least
thirty minutes of artifact-free baseline, non-SD and SD MEG-EEG will be analyzed. For the
purpose of blinding, each modality will be read independently by two readers, each of whom
will be blinded to the relationship of the MEG-EEG data to sleep deprivation, results
obtained by the other modality and subjective degree of sleepiness. Only interpretations of
each modality agreed upon by both readers will be accepted. When there is no agreement, a
third independent reader will resolve the disagreement.
The primary outcome measure will be the proportion of seizure foci detected and delineated
after SD on synchronized MEG-EEG recordings. Comparisons will also be made for each
recording modality and between them, according to the state of alertness during recording
and subjective feeling of sleepiness before each recording.
Patients 18 years or older
Patients of both sexes and of any race or ethnicity who have a diagnosis of epilepsy
routine EEG on the day of a baseline MEG-EEG without IEDs.
Patients younger than 18 years
Patients with progressive neurological disorders
Patients who are unable to lie still or cooperate adequately with MRI or MEG/EEG
Patients who are unable to sustain the required degree of sleep deprivation.
Patients who have had brain surgery.
Patients unable to sign and understand the study consent
Patients with IEDs on a routine EEG
Patients with IEDs on the baseline MEG-EEG