Back to search

Aristotle University Of Thessaloniki (AUTH)

Η χρησιμοποίηση μεθόδων τεχνητής νοημοσύνης για την αναγνώριση των επιδράσεων νευροτρόπων ουσιών στα προκλητά δυναμικά

Abstract

dc:description

Spikes and sharp waves with different polarity, morphology, size and field distributions were recorded from the interictal electroencephalograms obtained from epileptic patients. Normally such signal, (epileptic discharges), are ‘hidden’ within the whole signal that the electroencephalograph records. In our study we developed a method that uses the linear correlation function in order to reveal the epileptiform activity even in channels and montages where such activity is not evident. Moreover, we investigated the influence of specific antiepileptic agents and compared them with one another in these interictal recordings. For this purpose we used generalized linear models, analysis of variance (ANOVA) and the Tukey multiple comparison test. Finally, we propose a methodology which is based on power equations in order to describe in a mathematical way (and with a mathematical model the ‘behavior’ of the epileptiform discharges in patients under specific anticonvulsive monotherapy. The antiepileptics we study are Valproate, Carbamazepine and Oxcarbazepine. Briefly, our conclusions are: 1) All three independent variables (channel, frequency and antiepileptic) do influence our dependent variable, the amplitude. 2) The mean amplitude (the amplitude of the mean waveform) for every channel is influenced: a) By the frequency (observed in all channels) and b) By the drug administered (observed in specific channels and with different level of statistical significance). 3) Valproate is linked with lower values of the mean amplitude over the largest part of the frequency spectrum. This observation is most evident in the range 7 -10 Hz. It probably reflects a more evident suppressive action of Valproate in comparison with the other two antiepileptics. 4) Carbamazepine and Oxcarbazepine are clearly connected with higher mean amplitude (for the different frequencies in the recordings) in comparison with Valproate. This may reflect a weaker anticonvulsive action. Carbamazepine and Oxcarbazepine also present with a common ‘behavior’ regarding the fluctuation of the mean amplitude for the different frequencies of the EEG spectrum. The last observation may be due to a common action mechanism for the two drugs. 5) Oxcarbazepine is associated with the highest value of the mean amplitude over the lowest frequencies of the frequency spectrum (1 - 4 Hz). 6) Carbamazepine and Oxcarbazepine show a sharp increase of the mean amplitude followed by rapid decrease in the frequencies 10 - 13 Hz and 16 - 19 Hz, while Valproate is constantly associated with low values of the mean amplitude. 7) For all the above antiepileptics, the most intensive differences in the mean amplitude are seen in the frequencies 1 – 13 Hz, less intensive differences in the frequencies 14 – 50 Hz, while at higher frequencies they all follow a smoother common course. 8) The fewest (quantitatively) and smallest (qualitatively) differences of the mean amplitude for all three drugs are found in the channels 1, 5, 11 and 15. The highest values of the mean amplitude for the Valproate are seen in the channels 3, 6, 8, 15 and 16, for the Carbamazepine in the channels 3, 4, 6, 7, 8, 10 and 16, and for the Oxcarbazepine at last in the channels 1, 5, 6, 11, 15, 16 and 18. These positions are reflecting the topography of the potentially weaker anticonvulsive/suppressive action for each antiepileptic agent separately, and reveal a differentiation which in the cases of Carbamazepine and Oxcarbazepine cannot be explained on the basis of their action mechanism but could be attributed to the interruption of different (for each drug) cortex circuits. 9) Finally, we believe that we proved that is possible through the use of mathematical models and equations to predict the ‘behavior’ of antiepileptics as far as interictal epileptiform discharges are concerned.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2011

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Πιτσαλίδης, Χρήστος

Subjects

dc:subject × 20

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/30281

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Πιτσαλίδης, Χρήστος. Η χρησιμοποίηση μεθόδων τεχνητής νοημοσύνης για την αναγνώριση των επιδράσεων νευροτρόπων ουσιών στα προκλητά δυναμικά. Aristotle University Of Thessaloniki (AUTH), 2011. http://hdl.handle.net/10442/hedi/30281