K-complex detection using multi-layer perceptrons and recurrent networks.

B H Jansen; P R Desai

doi:10.1016/0020-7101(94)90123-6

Home/Archive/PubMed

PubMed · 7705906

K-complex detection using multi-layer perceptrons and recurrent networks.

B H Jansen·P R Desai

View original source ↗Export metadata ↓

Abstract

The feasibility of using a multi-layer perceptron and Elman's recurrent network for the detection of specific waveforms (K-complexes) in electroencephalograms (EEGs), regardless of their location in the signal segment, is explored. Experiments with simulated and actual EEG data were performed. In case of the perceptron, the input consisted of the magnitude and/or phase values obtained from 10-s signal intervals, whereas the recurrent net operated on the digitized data samples directly. It was found that both nets performed well on the simulated data, but not on the actual EEG data. The reasons for the failure of both nets are discussed.

Explore related subjects

Electroencephalography Humans Models, Neurological Neural Networks, Computer Signal Processing, Computer-Assisted Journal Article

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

B H Jansen, P R Desai. K-complex detection using multi-layer perceptrons and recurrent networks.. https://doi.org/10.1016/0020-7101(94)90123-6

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed1998

Nonstationarity in epileptic EEG and implications for neural dynamics.

In this paper, we use a recently developed method to analyze the nonstationarity in time series from intracranial depth and subdural recordings of patients with temporal lobe epilepsy. We show that the nonstationarity in the signal can be accounted for by the variation of a single parameter. We then show that the various dominant nonlinear waveforms observed in different electrodes can be explained by a simple stochastic model in which the mesoscopic collection of neurons, whose potential the electrodes measure, can be on one of two states. The nonstationarity observed in our analysis is a consequence of a time-dependent transition probability between these two states. In general, this transition probability increases as a seizure is approached. The model that we propose incorporates this bistability. We find good agreement between real data and simulated data generated by our model. We understand that this mesoscopic bistability may be associated with the existence of excitation waves traversing the brain in these patients.

Electroencephalography↗

PubMed1998

Responses of the nervous system to low frequency stimulation and EEG rhythms: clinical implications.

The present paper reviews literature data on the role of the non-specific central nervous system response mechanisms on the therapeutic effects of relatively weak external stimulations used in clinical practice. The factors affecting the stimulation efficiency and increased sensitiveness of living things to extra-low-frequency periodic stimulations (in the range of from less than 1 Hz to tens of Hz) are discussed. Among the factors determining such effects, the non-specific response mechanisms of the nervous system, the resonance phenomena in different organism systems, and the interaction of external stimulation with endogenous rhythmic processes are analyzed. Most attention is given to endogenous rhythms of the electrical brain activity reflected in the EEG rhythms. A high resolution EEG processing approach that is used to reveal the intrinsic oscillators in the individual EEG spectrum is described. Synchronization of sensory stimulation parameters with the frequencies of intrinsic EEG oscillators is supposed to be an appropriate way to enhance the therapeutic effects of various sensory stimulation treatments. Specific methods for utilizing resonance therapy via sensory stimulation with intrinsic EEG frequencies, and for automatic modulation of stimulation parameters by endogenous organism rhythms are delineated; some preliminary results are described.

Electroencephalography↗

PubMed1998

Prior spontaneous nocturnal waking duration and EEG during quiet sleep in infants: an automatic analysis approach.

To ascertain the role of spontaneous nocturnal waking duration on the EEG dynamics during quiet sleep (QS) periods, we analysed the nocturnal polygraphic recordings of 12 infants aged 9 47 weeks old. Their sleep was characterised by two sleep episodes, separated by spontaneous waking and containing at least two QS-paradoxical sleep (PS) cycles each. Automatic analysis of the EEG activity recorded by the centro-occipital lead and reflecting the degree of synchronisation allowed us to compute a parameter whose values ranged from 0 (maximum of EEG synchronisation) to 10 (maximum of EEG de-synchronisation). Three indicators of the time course of the parameter value were computed during the first QS period of the sleep episode subsequent to nocturnal waking: (i) the parameter range (difference between the EEG parameter value at the QS onset and that at the trough-maximum of EEG synchronisation); (ii) the trough latency (time interval between QS onset and trough); and (iii) rate of synchronisation (range/trough latency). These three indicators were the dependent variables in a multiple regression model, where the independent variables were age and the logarithm of the duration of prior waking. The parameter range was correlated with the duration of prior waking. Both the trough latency and the rate of synchronisation were correlated with age only, respectively, positively and negatively. The marked decline in the rate of synchronisation throughout the first year of life could account for the failure to find a significant correlation between prior waking and the above indicator of the EEG dynamics. The relationship between the duration of prior waking and the parameter range in following sleep in infants supports the hypothesis of the early emergence of the homeostatic regulation of sleep.

Electroencephalography↗