Machine learning reveals immediate disruption in mosquito flight when exposed to Olyset nets

Area of Science:

• Entomology
• Malariology
• Machine Learning Applications

Background:

• Insecticide-treated nets (ITNs) are vital for malaria control.
• Mosquito behavioural responses to ITNs, especially insecticide resistance, require further understanding.

Purpose of the Study:

• To analyze Anopheles gambiae flight behaviour around Olyset nets (OL) using machine learning.
• To compare responses between insecticide-resistant (IR) and susceptible (IS) strains to treated vs. untreated nets.

Main Methods:

• Utilized machine learning models to classify mosquito flight trajectories.
• Employed SHAP analysis to identify key flight behaviour predictors.
• Conducted experiments comparing mosquito behaviour around Olyset nets versus untreated nets.

Main Results:

• Both IR and IS mosquitoes exhibited immediate, convoluted flight patterns around OL nets.
• Flight disruption was characterized by changes in angle and velocity, indicating irritancy.
• Insecticide resistance did not prevent behavioural disruption, though IR mosquitoes survived longer.

Conclusions:

• Permethrin in ITNs acts as an irritant, not a repellent, upon contact.
• Machine learning trajectory analysis effectively reveals mosquito behavioural responses to insecticides.
• Findings inform ITN design and the development of new malaria control interventions.