In our neurometrics approach, AI plays a central role in identifying treatment effects using
support vector machines and in discovering the effects of interventions using unsupervised
clustering approaches. When analyzing animal behavior, we use AI-based approaches to
classify behavior, including DeepLabCut and MoSeq, in addition to the criteria accepted in
published studies for performance in a task, such as choosing a particular arm in a maze or
scoring food intake by the observer in a video recording. To ensure rigorous analysis, we
adhere to a strict methodology in which neither the experimenter nor the analyst is aware
of which groups receive the vehicle or various drugs and their doses. This blinding process
ensures unbiased assessments. We also ensure the robustness of our results through 

exhaustive resampling and shuffling to control for any spurious dependencies within the dataset.