Collective behavior quantification on human odor effects against female Aedes aegypti mosquitoes—Open source development

Abstract

Classifying and quantifying mosquito activity includes a plethora of categories, ranging from measuring flight speeds, repellency, feeding rates, and specific behaviors such as home entry, swooping and resting, among others. Entomologists have been progressing more toward using machine vision for efficiency for this endeavor. Digital methods have been used to study the behavior of insects in labs, for instance via three-dimensional tracking with specialized cameras to observe the reaction of mosquitoes towards human odor, heat and CO2, although virtually none was reported for several important fields, such as repellency studies which have a significant need for a proper response quantification. However, tracking mosquitoes individually is a challenge and only limited number of specimens can be studied. Although tracking large numbers of individual insects is hailed as one of the characteristics of an ideal automated image-based tracking system especially in 3D, it also is a costly method, often requiring specialized hardware and limited access to the algorithms used for mapping the specimens. The method proposed contributes towards (a) unlimited open source use, (b) a low-cost setup, (c) complete guide for any entomologist to adapt in terms of hardware and software, (d) simple to use, and (e) a lightweight data output for collective behavior analysis of mosquitoes. The setup is demonstrated by testing a simple response of mosquitoes in the presence of human odor versus control, one session with continuous human presence as a stimuli and the other with periodic presence. A group of female Aedes aegypti (Linnaeus) mosquitoes are released into a white-background chamber with a transparent acrylic panel on one side. The video feed of the mosquitoes are processed using filtered contours in a threshold-adjustable video. The mosquitoes in the chamber are mapped on the raster where the coordinates of each mosquito are recorded with the corresponding timestamp. The average distance of the blobs within the frames against time forms a spectra where behavioral patterns can be observed directly, whether any collective effect is observed. With this method, 3D tracking will not be required and a more straightforward data output can be obtained.

Effects of low-powered RF sweep between 0.01-20 GHz on female Aedes Aegypti mosquitoes: A collective behaviour analysis

“…However, in this study, a more specific design is required as shown in Fig 1, which is a modified version from a previous closely related research [20]. …”

Keeping track of mosquitoes: a review of tools to track, record and analyse mosquito flight

“…Overall, the technical development has led to a variety of tracking systems to choose from, a reduction in costs and an increase in the portability of recording set-ups. Advantages and disadvantages of quantification techniques have recently been summarized in a non-exhaustive list by Poh et al [ 153 ]. It is remarkable that the table lists a variety of techniques that are described as either ‘complex to set up’ or ‘cover a small area of observation’ indicating that there is still much room for improvement.…”

A minimal 3D model of mosquito flight behavior around the human baited bed net

“…This paper 26 describes one such model: a fine-grained agent-based approach for modelling how indoor 27 insecticide treatments deployed as residues on bed nets affect the behavior and survival 28 of mosquitoes. 29 To capture the complex interactions between mosquitoes and a human host lying 30 beneath an insecticide-treated bed net, individual mosquito flight paths and local 31 interactions within the indoor environment are required. Yet most modelling approaches 32 to disease transmission by insect vectors set the scale at a much higher level, 33 emphasizing interactions at village scale (and above) [11], [12].…”

“…Recent technological developments in imaging and computing have enabled accurate 73 tracking of mosquito flight dynamics. This is achieved by analyzing video recordings of 74 mosquitoes and applying image processing and analysis methods to reconstruct and log 75 the flight paths and range from relatively simple approaches using a single 76 camera [30], [31] to more complex multi-camera methods [9], and 3D reconstruction 77 approaches [32], [33]. These multi-camera approaches allow the tracking of mosquito 78 flight around a bed net and therefore a more complex representation of bed net 79 occupancy at a 'life-size scale.…”