Computer-Based Prediction of Psychotropic Drug Classes Based on a Discriminant Analysis of Drug Effects on Rat Sleep

Abstract

The goal of the present study was to classify psychotropic drugs on the basis of EEG-defined rat sleep-waking behaviour. Using an automated sleep classification system it was found that some of the drug-induced changes in sleep-waking behaviour were specific for the pharmacotherapeutic treatment class to which the drug belonged. In several preliminary experiments we further found that drugs may have effects on rat EEG independent of their effects on rat sleep-waking behaviour and that these pharmaco-EEG effects may be different for the various sleep and waking stages. By analysing sleep class-independent EEG-spectral parameters a single drug effect score can, moreover, be obtained giving information on drug pharmacodynamics. The drug-induced changes in sleep-waking behaviour were used to classify a large number of drugs into several therapy classes by means of a discriminant analysis procedure. Antidepressants, antipsychotics and stimulants were discriminated successfully from each other and from placebo by this system, whereas nootropics classified as placebo. Anxiolytics, hypnotics and anticonvulsants classified poorly. Their classification is hampered by the lack of specific compounds. Assigned drug class and assignment probability were dose dependent. In the discussion of the present study it is suggested that animal pharmaco-sleep and pharmaco-EEG studies are not mutually exclusive approaches, but that they may complement each other.

Pharmaco-EEG Studies in Animals: A History-Based Introduction to Contemporary Translational Applications

“…One may try to entrain animals into a stable state of sustained vigilance, for instance by forced locomotor activity [42] or by training the animals to attend to sensory stimuli and study the effects of drugs on the EEG of such a stable state of vigilance. Alternatively, ethnographical analysis of the behaviour of the animal can be used to segment the EEG into discontinuous segments belonging to different behavioural states and then study the effects of a drug on the more homogeneous EEG for the pooled segments of each of these different states [33,34,35]. This has the added advantage that apart from the effect of a drug on the various behaviour-defined types of EEG, one also obtains the effects of a drug on the ethogram itself [43].…”

“…Animal p-EEG was relatively cheap when compared to clinical EEG recordings, and the throughput of compounds was much higher, especially if small rodents were used rather than cats, dogs and monkeys. Thus, animal p-EEG groups emerged in the industry in the early 1980s, exemplified by Ciba-Geigy [27], Duphar [28,29,30], Janssen Pharmaceutica [31,32], Organon [33,34,35], Sandoz [36] and Synthélabo [37]. When clinical p-EEG failed to accurately predict the therapeutic value of a number of compounds (e.g.…”

Guidelines for the Recording and Evaluation of Pharmaco-Sleep Studies in Man: The International Pharmaco-EEG Society (IPEG)

“…Although sleep is disrupted in several psychiatric disorders (and their animal models), pharmaco-sleep biomarkers do not need to be disease specific in order to be valuable in drug discovery. Drug-induced changes in sleep-wake patterns and associated EEG spectra in healthy animals and human volunteers have been successfully used to characterise novel CNS-active drugs as potential therapeutic agents for different CNS disorders in a probabilistic manner (71-95%) [311,312]. …”

Modulation of Neurotransmission by a Specified Oregano Extract Alters Brain Electrical Potentials Indicative of Antidepressant-Like and Neuroprotective Activities

“…We used the discriminant analysis to compare the pattern of frequency changes generated by OE with pharmacological effects of known compounds [20][21][22]. Spherical projection of the result of this calculation led to an obvious separation of the action of OE.…”

“…In general, electropharmacograms show prominent differences for drugs prescribed for different indications and are similar for drugs with similar clinical indications [10]. In addition, the discriminant analysis provides the statistical fundament for classification of single drug's effects into disease specific clusters [20][21][22]. All compounds within each cluster exhibit a similar pattern of changes providing a disease specific "fingerprint" of electrical activity.…”