TY - JOUR

T1 - Use of Naïve Bayes Classifier to Assess the Effects of Antipsychotic Agents on Brain Electrical Activity Parameters in Rats

AU - Sysoev, Yu. I.

AU - Shits, D. D.

AU - Puchik, M. M.

AU - Prikhodko, V. A.

AU - Idiyatullin, R. D.

AU - Kotelnikova, A. A.

AU - Okovityi, S. V.

PY - 2022/7/1

Y1 - 2022/7/1

N2 - Research and development of novel methods to determine the effects of antipsychotic agents is an important challenge for experimental biomedicine. Although behavioural tests, the ones most commonly used for pharmacological screening, are quite efficient for the evaluation of drug effects on animal anxiety and locomotion, they hardly allow to detect antipsychotic activity. Pharmacoelectroencephalography (pharmaco-EEG), which is based on the principle of different psychoactive agents producing distinct changes in brain electrical activity, could represent a viable alternative approach to that task. The rapid evolution of machine learning techniques has opened new possibilities for using pharmaco-EEG data for the purposes of classification and prediction. This work describes an experimental approach to the assessment of specific activity and pharmacological profiling of antipsychotic agents using naive Bayes classifier, a simple probabilistic classifier widely employed in biomedical research. The experiments were conducted in white outbred male rats with chronically implanted electrocorticographic electrodes. To serve as the training set, a library was assembled containing electrocorticograms (ECoG) following the administration of antipsychotic agents: chlorpromazine, haloperidol, droperidol, tiapride, and sulpiride. For each sample, ECoG parameters before and after drug administration were calculated, and a total of 132 amplitude and spectral signal parameters were taken into analysis. Principal component analysis was used to reduce dimensionality. Using naive bayes classifier, we were able to detect and qualify distinct effects of antipsychotic agents on brain electrical activity parameters in rats, allowing them to be differentiated from phenazepam, a benzodiazepine tranquilizer with sedative properties. Moreover, this approach proved effective to distinguish among the antipsychotics as well as between them and other agents with similar receptor binding affinity profiles, e.g., the tricyclic antidepressant amitriptyline. Thus, the method we propose can be used to discern between antipsychotic and sedative effects of drugs as well as to compare the effects across different antipsychotic agents. PU - PLEIADES PUBLISHING INC PI - NEW YORK PA - PLEIADES HOUSE, 7 W 54 ST, NEW YORK, NY, UNITED STATES

AB - Research and development of novel methods to determine the effects of antipsychotic agents is an important challenge for experimental biomedicine. Although behavioural tests, the ones most commonly used for pharmacological screening, are quite efficient for the evaluation of drug effects on animal anxiety and locomotion, they hardly allow to detect antipsychotic activity. Pharmacoelectroencephalography (pharmaco-EEG), which is based on the principle of different psychoactive agents producing distinct changes in brain electrical activity, could represent a viable alternative approach to that task. The rapid evolution of machine learning techniques has opened new possibilities for using pharmaco-EEG data for the purposes of classification and prediction. This work describes an experimental approach to the assessment of specific activity and pharmacological profiling of antipsychotic agents using naive Bayes classifier, a simple probabilistic classifier widely employed in biomedical research. The experiments were conducted in white outbred male rats with chronically implanted electrocorticographic electrodes. To serve as the training set, a library was assembled containing electrocorticograms (ECoG) following the administration of antipsychotic agents: chlorpromazine, haloperidol, droperidol, tiapride, and sulpiride. For each sample, ECoG parameters before and after drug administration were calculated, and a total of 132 amplitude and spectral signal parameters were taken into analysis. Principal component analysis was used to reduce dimensionality. Using naive bayes classifier, we were able to detect and qualify distinct effects of antipsychotic agents on brain electrical activity parameters in rats, allowing them to be differentiated from phenazepam, a benzodiazepine tranquilizer with sedative properties. Moreover, this approach proved effective to distinguish among the antipsychotics as well as between them and other agents with similar receptor binding affinity profiles, e.g., the tricyclic antidepressant amitriptyline. Thus, the method we propose can be used to discern between antipsychotic and sedative effects of drugs as well as to compare the effects across different antipsychotic agents. PU - PLEIADES PUBLISHING INC PI - NEW YORK PA - PLEIADES HOUSE, 7 W 54 ST, NEW YORK, NY, UNITED STATES

UR - https://www.mendeley.com/catalogue/5a4db489-5882-36a6-b3d2-cd1650cce616/

U2 - 10.1134/s0022093022040160

DO - 10.1134/s0022093022040160

M3 - Article

VL - 58

SP - 1130

EP - 1141

JO - Journal of Evolutionary Biochemistry and Physiology

JF - Journal of Evolutionary Biochemistry and Physiology

SN - 0022-0930

IS - 4

ER -