Biomedical Engineering Reference
In-Depth Information
Finally the results of the above mentioned works were gathered in an open system
for sleep staging, based explicitly on the R&K criteria [Malinowska et al., 2009].
The system started with detection and parametrization of relevant waveforms and
then combining these results (together with information from EMG and EOG sig-
nals) into a final decision assigning epochs to sleep stages. Automatic sleep staging
was performed in a hierarchical way presented in
Figure 4.14.
In the first step, each
20-s epoch is tested for muscle artifacts in EEG or EMG derivation, which indicate
the movement time (MT). If at least one of the analyzed derivations (C3, C4, or
EMG) exceeded a corresponding threshold in more than 50% of the 20 s epoch, the
epoch is scored as MT. In the second step, the algorithm detects slow wave sleep
stages 3 and 4 by applying fixed 20% and 50% thresholds to the amount of epochs
time occupied by slow waves. In the following step, the algorithm detects sleep spin-
dles and K-complexes, which are related to stage 2. If at least one sleep spindle or
K-complex occurs in the 20 s epoch, and less than 75% of the epoch is occupied by
alpha activity, the epoch is scored as stage 2. If alpha activity occupies above 75%,
EOG and EMG signals of this epoch are examined to distinguish stage REM from
wake. Performance of the system was evaluated on 20 polysomnographic record-
ings scored by experienced encephalographers. The system gave 73% concordance
with visual staging—close to the inter-expert concordance. Similar results were ob-
tained for other systems reported in the literature [Penzel et al., 2007], however these
expert systems were tuned explicitly for maximizing the concordance with visual
scoring and were usually based on a black-box approach, so their parameters were
difficult to relate to the properties of EEG signals observed in visual analysis. Also,
most of these systems are closed-source commercial solutions. On the contrary, a
bottom-up approach presented in [Malinowska et al., 2009] together with an open
implementation (including complete source code) freely available from the internet
The parameters of the system can be easily adapted to the other criteria concerning
sleep stage classification.
4.1.6.4
Analysis of EEG in epilepsy
Epilepsy is one of the most common neurological disorders second only to stroke;
it affects about 0.8% of the world's population. The most important treatment is
pharmacological, however in 25% of patients seizures are drug resistant. Epilepsy is
manifested by a sudden and recurrent brain malfunction which has its origin in ex-
cessive and hypersynchronous activity of neurons. The seizures occur at random and
impair the normal function of the brain. During the seizure the electroencephalogram
changes dramatically: its amplitude increases by an order of magnitude and charac-
teristic patterns varying in time and frequency appear. The mechanisms of the seizure
generation are still not known. It is presumed that seizure occurrence is connected
with the decrease of the inhibition in brain. There are several models of seizure gen-
eration and some of them are compatible with the observed electroencephalographic
activity [Lopes da Silva et al., 2003a].
Electroencephalography is the most useful and cost-effective modality for the
study of epilepsy. For the purpose of the localization of epileptic foci imaging meth-
Search WWH ::
Custom Search