A comparison of doubly hierarchical discriminant analyses for multiple class longitudinal data from EEG experiments

Area of Science:

• Neuroscience
• Pharmacology
• Data Science

Background:

• Longitudinal data analysis is crucial in preclinical pharmaco-electroencephalogram (EEG) studies.
• Characterizing psychotropic drug effects requires sophisticated classification methods.
• Existing methods may not fully capture the complexity of multiple drug classes and temporal data.

Purpose of the Study:

• To propose a general and simple procedure for establishing classification rules for multiple-class longitudinal data.
• To apply this procedure to preclinical pharmaco-EEG studies for characterizing psychotropic drug effects.
• To develop a flexible hierarchical supervised learning tool tailored for EEG data.

Main Methods:

• Development of a general classification procedure for longitudinal data.
• Application of spectral EEG analysis to characterize drug effects.
• Utilizing a flexible hierarchical supervised learning approach.
• Testing several variations of the proposed procedure on EEG datasets.

Main Results:

• The proposed procedure successfully established classification rules for longitudinal EEG data.
• Comparable results were obtained across different variations of the procedure.
• A significant association was found between sleeping stages and psychotropic drug classes.
• The method demonstrated flexibility in handling multiple drug classes and longitudinal data.

Conclusions:

• The developed procedure offers a general and simple approach for classifying multiple-class longitudinal data.
• This method is effective for characterizing psychotropic drug effects using spectral EEG analysis in preclinical studies.
• The hierarchical supervised learning tool accounts for the specific nature of drug classes and the longitudinal aspect of EEG data.