A Data-Driven Threshold Standard for Olfactory Behavior: A Case Study with Trichogramma japonicum

Objectively defining an effective choice is crucial for the reliability of results in insect olfactory behavior research. Relying on subjective time thresholds often leads to inconsistent findings, primarily due to the high likelihood of mistaking an accidental intrusion for an effective choice. Our analysis indicated that 29% of preference responses were not consistent among 0 s, 15 s, 30 s, and 60 s, when these thresholds were applied to evaluate the behavioral responses of Trichogramma japonicum Ashmead (Hym.: Trichogrammatidae) to 20 odorants in a four-arm olfactometer. This inconsistency arises because subjective thresholds fail to distinguish between accidental, transient entries and sustained, purposeful choices, leading to false positives or false negatives. To address this issue, we compiled a database of the initial stay duration of T. japonicum in the odor zone, and applied the 5th percentile of its empirical distribution-a widely accepted statistical standard for excluding extreme outliers (i.e., accidental brief entries) in medical and biological research-as the threshold to differentiate between accidental intrusion and effective choice. We then systematically assessed the behavioral responses of T. japonicum to 20 odorants and validated the authenticity of olfactory preferences through oviposition choice experiments. A continuous stay of ≥10 s was determined as an effective choice. Based on this criterion, T. japonicum exhibited significant preference for 0.1 μg/μL α-pinene, 0.01 μg/μL D-limonene and 0.001 μg/μL citral, whereas avoiding 10 μg/μL β-ionone. The results of the oviposition choice experiments showed high consistency with the olfactory behavioral preferences, confirming the reliability of this standard in predicting its actual host selection decisions. Collectively, this study proposes a data-driven protocol for establishing an objective and stable behavioral criterion in insect olfactory behavior research.