Using mobile phones as acoustic sensors for high-throughput mosquito surveillance.

Haripriya Mukundarajan,Manu Prakash,Cooper Newby,Erica Araceli Castillo,Felix Jan Hein Hol

doi:10.7554/elife.27854

Abstract

The direct monitoring of mosquito populations in field settings is a crucial input for shaping appropriate and timely control measures for mosquito-borne diseases. Here, we demonstrate that commercially available mobile phones are a powerful tool for acoustically mapping mosquito species distributions worldwide. We show that even low-cost mobile phones with very basic functionality are capable of sensitively acquiring acoustic data on species-specific mosquito wingbeat sounds, while simultaneously recording the time and location of the human-mosquito encounter. We survey a wide range of medically important mosquito species, to quantitatively demonstrate how acoustic recordings supported by spatio-temporal metadata enable rapid, non-invasive species identification. As proof-of-concept, we carry out field demonstrations where minimally-trained users map local mosquitoes using their personal phones. Thus, we establish a new paradigm for mosquito surveillance that takes advantage of the existing global mobile network infrastructure, to enable continuous and large-scale data acquisition in resource-constrained areas.

Highlights

Frequent, widespread, and high resolution surveillance of mosquitoes is essential to understanding their complex ecology and behaviour
We found that different lab-reared strains of Anopheles gambiae, including permethrin susceptible and resistant variants sourced from Kenya and an additional bendiocarb resistant strain sourced from Benin, produced highly similar frequency distributions with large overlaps of 0.84 to 0.89 as measured by their mutual Bhattacharya Coefficeint (BC) (Figure 3 - figure supplement 1C)
We demonstrate a new technique for mosquito surveillance, using mobile phones as networked acoustic sensors to collect wingbeat frequency data for automated identification

Summary

Introduction

Widespread, and high resolution surveillance of mosquitoes is essential to understanding their complex ecology and behaviour. Efforts to map mosquito populations primarily rely on interpolative mathematical models based on factors such as clinical disease burdens or climatological data, with field inputs from entomological surveys being a comparatively sparse contribution (Hay et al, 2010; Mordecai et al, 2017). This scarcity of field data stems from the absence of high-throughput, lowcost surveillance techniques to map the distribution of mosquitoes. For disease control strategies to be truly effective, it is critical for them to be strongly informed by current mosquito population distributions

Methods

Results

Conclusion