Table 3.

Methodological considerations for detection and analysis of freezing behavior during fear learning procedures.

Experimental Feature	Methodological Recommendations	Additional Considerations
Video Recording	• Avoid adverse effects of video quality by recording trials at a minimum of 25 frames per second • Eliminate experimenter-induced variability (scent, movement, gender, etc.) by not having personnel in the behavioral testing room during trials • Mitigate data loss by using cloud-based backup for video storage	Additional features that can optimize video and data quality:• High contrast between the animal and the test background • Optimal lighting conditions (minimized shadows and glare, consistent lux value) • High image resolution • Maintain view of the entire animal in the frame (even while it is rearing) • Ensure the only in-frame movement is that of the animal
Manual Scoring of Freezing Behavior	• Use a detailed scoring protocol to standardize scoring across experiments and lab personnel turnover • Average scores from two independent, highly trained raters • Ensure raters are blinded to animal treatment condition • Continuous scoring may be more comprehensive than time sampling approaches
Automated Scoring of Freezing Behavior	• Optimize threshold of detection for local lighting conditions then keep software settings consistent for all experimental animals • Background subtraction may enhance freezing accuracy	Automated scoring procedures and settings can be validated by correlating freezing data with corresponding data generated by manual scoring, see Recommendations for Freezing Detection and Data Analysis