A novel method for estrous cycle staging using supervised object detection.

Area of Science:

• Reproductive Biology
• Computational Biology
• Neuroscience

Background:

• The estrous cycle is critical for female mammal reproduction and health, analogous to the human menstrual cycle.
• Traditional monitoring via vaginal cytology has limitations, including time-consuming training and researcher variability.
• Accurate estrous cycle tracking is vital for interpreting results in preclinical research, especially in neuropsychiatric studies.

Purpose of the Study:

• To assess the feasibility of machine learning for estrous cycle staging.
• To introduce and validate an object detection-based model for automated estrous cycle classification.
• To provide a reliable and efficient alternative to manual vaginal cytology.

Main Methods:

• Developed and employed an object detection-based machine learning model named Object Detection Estrous Staging (ODES).
• Utilized a dataset of 730 stained vaginal cytology images for training, validation, and testing.
• Derived a novel classification rule set by analyzing training images to categorize estrous cycle stages.

Main Results:

• ODES achieved an average accuracy of 80% in classifying estrous cycle stages.
• Model performance was comparable to human accuracy (66%) and superior to previous image classification models (41-79%).
• ODES processed 100 test images in 2.67 minutes, demonstrating high efficiency.

Conclusions:

• ODES provides a fast, reliable, and accessible method for estrous cycle monitoring.
• The model's efficiency makes it suitable for large-scale studies, particularly in neuropsychiatric research involving female rodents.
• This automated approach can improve the integration of sex-based variables into neurological and psychiatric research.