Integrating Global Citizen Science Platforms to Enable Next-Generation Surveillance of Invasive and Vector Mosquitoes

Machine learning in healthcare citizen science: A scoping review

OBJECTIVES

This scoping review aims to clarify the definition and trajectory of citizen-led scientific research (so-called citizen science) within the healthcare domain, examine the degree of integration of machine learning (ML) and the participation levels of citizen scientists in health-related projects.

MATERIALS AND METHODS

In January and September 2024 we conducted a comprehensive search in PubMed, Scopus, Web of Science, and EBSCOhost platform for peer-reviewed publications that combine citizen science and machine learning (ML) in healthcare. Articles were excluded if citizens were merely passive data providers or if only professional scientists were involved.

RESULTS

Out of an initial 1,395 screened, 56 articles spanning from 2013 to 2024 met the inclusion criteria. The majority of research projects were conducted in the U.S. (n = 20, 35.7 %), followed by Germany (n = 6, 10.7 %), with Spain, Canada, and the UK each contributing three studies (5.4 %). Data collection was the primary form of citizen scientist involvement (n = 29, 51.8 %), which included capturing images, sharing data online, and mailing samples. Data annotation was the next most common activity (n = 15, 26.8 %), followed by participation in ML model challenges (n = 8, 14.3 %) and decision-making contributions (n = 3, 5.4 %). Mosquitoes (n = 10, 34.5 %) and air pollution samples (n = 7, 24.2 %) were the main data objects collected by citizens for ML analysis. Classification tasks were the most prevalent ML method (n = 30, 52.6 %), with Convolutional Neural Networks being the most frequently used algorithm (n = 13, 20 %).

DISCUSSION AND CONCLUSIONS

Citizen science in healthcare is currently an American and European construct with growing expansion in Asia. Citizens are contributing data, and labeling data for ML methods, but only infrequently analyzing or leading studies. Projects that use "crowd-sourced" data and "citizen science" should be differentiated depending on the degree of involvement of citizens.

iNaturalist community observations provide valuable data on human-mosquito encounters

Mosquitoes (Diptera: Culicidae) and the pathogens they transmit represent a threat to human and animal health. Low-cost and effective surveillance methods are necessary to enable sustainable monitoring of mosquito distributions, diversity, and human interactions. This study examined the use of iNaturalist, an online, community-populated biodiversity recording database, for passive mosquito surveillance in the United Kingdom (UK) and Ireland, countries under threat from the introduction of invasive mosquitoes and emerging mosquito-borne diseases. The Mozzie Monitors UK & Ireland iNaturalist project was established to collate mosquito observations in these countries. Data were compared with existing long-term mosquito UK datasets to assess representativeness of seasonal and distribution trends in citizen scientist-recorded observations. The project collected 738 observations with the majority recorded 2020-2022. Records were primarily associated with urban areas, with the most common species Culex pipiens and Culiseta annulata significantly more likely to be observed in urban areas than other species. Analysis of images uploaded to the iNaturalist project also provided insights into human-biting behavior. Our analyses indicate that iNaturalist provides species composition, seasonal occurrence, and distribution figures consistent with existing datasets and is therefore a useful surveillance tool for recording information on human interactions with mosquitoes and monitoring species of concern.

Deep learning in disease vector image identification

Vector-borne diseases (VBDs) represent a critical global public health concern, with approximately 80% of the world's population at risk of one or more VBD. Manual disease vector identification is time-consuming and expert-dependent, hindering disease control efforts. Deep learning (DL), widely used in image, text, and audio tasks, offers automation potential for disease vector identification. This paper explores the substantial potential of combining DL with disease vector identification. Our aim is to comprehensively summarize the current status of DL in disease vector identification, covering data collection, data preprocessing, model construction, evaluation methods, and applications in identification spanning from species classification to object detection and breeding site identification. We also discuss the challenges and possible prospects for DL in disease vector identification for further research. © 2024 Society of Chemical Industry.

Assessing and correcting neighborhood socioeconomic spatial sampling biases in citizen science mosquito data collection

Climatic, ecological, and socioeconomic factors are facilitating the spread of mosquito-borne diseases, heightening the importance of vector surveillance and control. Citizen science is proving to be an effective tool to track mosquito populations, but methods are needed to detect and account for small scale sampling biases in citizen science surveillance. In this article we combine two types of traditional mosquito surveillance records with data from the Mosquito Alert citizen science system to explore the ways in which the socioeconomic characteristics of urban neighborhoods result in sampling biases in citizen scientists' mosquito reports, while also shaping the spatial distribution of mosquito populations themselves. We use Barcelona, Spain, as an example, and focus on Aedes albopictus, an invasive vector species of concern worldwide. Our results suggest citizen scientists' sampling effort is focused more in Barcelona's lower and middle income census tracts than in its higher income ones, whereas Ae. albopictus populations are concentrated in the city's upper-middle income tracts. High resolution estimates of the spatial distribution of Ae. albopictus risk can be improved by controlling for citizen scientists' sampling effort, making it possible to provide better insights for efficiently targeting control efforts. Our methodology can be replicated in other cities faced with vector mosquitoes to improve public health responses to mosquito-borne diseases, which impose massive burdens on communities worldwide.

New Insights into Rabbit Viral Diseases

Viruses are responsible for many devastating rabbit diseases that impact their health and welfare and put their conservation and economic revenue at risk [...].

Robust mosquito species identification from diverse body and wing images using deep learning

Mosquito-borne diseases are a major global health threat. Traditional morphological or molecular methods for identifying mosquito species often require specialized expertise or expensive laboratory equipment. The use of convolutional neural networks (CNNs) to identify mosquito species based on images may offer a promising alternative, but their practical implementation often remains limited. This study explores the applicability of CNNs in classifying mosquito species. It compares the efficacy of body and wing depictions across three image collection methods: a smartphone, macro-lens attached to a smartphone and a professional stereomicroscope. The study included 796 specimens of four morphologically similar Aedes species, Aedes aegypti, Ae. albopictus, Ae. koreicus and Ae. japonicus japonicus. The findings of this study indicate that CNN models demonstrate superior performance in wing-based classification 87.6% (95% CI: 84.2-91.0) compared to body-based classification 78.9% (95% CI: 77.7-80.0). Nevertheless, there are notable limitations of CNNs as they perform reliably across multiple devices only when trained specifically on those devices, resulting in an average decline of mean accuracy by 14%, even with extensive image augmentation. Additionally, we also estimate the required training data volume for effective classification, noting a reduced requirement for wing-based classification compared to body-based methods. Our study underscores the viability of both body and wing classification methods for mosquito species identification while emphasizing the need to address practical constraints in developing accessible classification systems.

Livestock keeping, mosquitoes and community viewpoints: a mixed methods assessment of relationships between livestock management, malaria vector biting risk and community perspectives in rural Tanzania

BACKGROUND

Livestock keeping is one of the potential factors related to malaria transmission. To date, the impact of livestock keeping on malaria transmission remains inconclusive, as some studies suggest a zooprophylactic effect while others indicate a zoopotentiation effect. This study assessed the impact of livestock management on malaria transmission risks in rural Tanzania. Additionally, the study explored the knowledge and perceptions of residents about the relationships between livestock keeping and malaria transmission risks in a selected village.

METHODS

In a longitudinal entomological study in Minepa village, South Eastern Tanzania, 40 households were randomly selected (20 with livestock, 20 without). Weekly mosquito collection was performed from January to April 2023. Indoor and outdoor collections used CDC-Light traps, Prokopack aspirators, human-baited double-net traps, and resting buckets. A subsample of mosquitoes was analysed using PCR and ELISA for mosquito species identification and blood meal detection. Livestock's impact on mosquito density was assessed using negative binomial GLMMs. Additionally, in-depth interviews explored community knowledge and perceptions of the relationship between livestock keeping and malaria transmission risks.

RESULTS

A total of 48,677 female Anopheles mosquitoes were collected. Out of these, 89% were Anopheles gambiae sensu lato (s.l.) while other species were Anopheles funestus s.l., Anopheles pharoensis, Anopheles coustani, and Anopheles squamosus. The findings revealed a statistically significant increase in the overall number of An. gambiae s.l. outdoors (RR = 1.181, 95%CI 1.050-1.862, p = 0.043). Also, there was an increase of the mean number of An. funestus s.l. mosquitoes collected in households with livestock indoors (RR = 2.866, 95%CI: 1.471-5.582, p = 0.002) and outdoors (RR = 1.579,95%CI 1.080-2.865, p = 0.023). The human blood index of Anopheles arabiensis mosquitoes from houses with livestock was less than those without livestock (OR = 0.149, 95%CI 0.110-0.178, p < 0.001). The majority of participants in the in-depth interviews reported a perceived high density of mosquitoes in houses with livestock compared to houses without livestock.

CONCLUSION

Despite the potential for zooprophylaxis, this study indicates a higher malaria transmission risk in livestock-keeping communities. It is crucial to prioritize and implement targeted interventions to control vector populations within these communities. Furthermore, it is important to enhance community education and awareness regarding covariates such as livestock that influence malaria transmission.

An optical system to detect, surveil, and kill flying insect vectors of human and crop pathogens

Sustainable and effective means to control flying insect vectors are critically needed, especially with widespread insecticide resistance and global climate change. Understanding and controlling vectors requires accurate information about their movement and activity, which is often lacking. The Photonic Fence (PF) is an optical system that uses machine vision, infrared light, and lasers to identify, track, and interdict vectors in flight. The PF examines an insect's outline, flight speed, and other flight parameters and if these match those of a targeted vector species, then a low-power, retina-safe laser kills it. We report on proof-of-concept tests of a large, field-sized PF (30 mL × 3 mH) conducted with Aedes aegypti, a mosquito that transmits dangerous arboviruses, and Diaphorina citri, a psyllid which transmits the fatal huanglongbing disease of citrus. In tests with the laser engaged, < 1% and 3% of A. aegypti and D. citri, respectfully, were recovered versus a 38% and 19% recovery when the lacer was silenced. The PF tracked, but did not intercept the orchid bee, Euglossa dilemma. The system effectively intercepted flying vectors, but not bees, at a distance of 30 m, heralding the use of photonic energy, rather than chemicals, to control flying vectors.

Classifying stages in the gonotrophic cycle of mosquitoes from images using computer vision techniques

The ability to distinguish between the abdominal conditions of adult female mosquitoes has important utility for the surveillance and control of mosquito-borne diseases. However, doing so requires entomological training and time-consuming manual effort. Here, we design computer vision techniques to determine stages in the gonotrophic cycle of female mosquitoes from images. Our dataset was collected from 139 adult female mosquitoes across three medically important species-Aedes aegypti, Anopheles stephensi, and Culex quinquefasciatus-and all four gonotrophic stages of the cycle (unfed, fully fed, semi-gravid, and gravid). From these mosquitoes and stages, a total of 1959 images were captured on a plain background via multiple smartphones. Subsequently, we trained four distinct AI model architectures (ResNet50, MobileNetV2, EfficientNet-B0, and ConvNeXtTiny), validated them using unseen data, and compared their overall classification accuracies. Additionally, we analyzed t-SNE plots to visualize the formation of decision boundaries in a lower-dimensional space. Notably, ResNet50 and EfficientNet-B0 demonstrated outstanding performance with an overall accuracy of 97.44% and 93.59%, respectively. EfficientNet-B0 demonstrated the best overall performance considering computational efficiency, model size, training speed, and t-SNE decision boundaries. We also assessed the explainability of this EfficientNet-B0 model, by implementing Grad-CAMs-a technique that highlights pixels in an image that were prioritized for classification. We observed that the highest weight was for those pixels representing the mosquito abdomen, demonstrating that our AI model has indeed learned correctly. Our work has significant practical impact. First, image datasets for gonotrophic stages of mosquitoes are not yet available. Second, our algorithms can be integrated with existing citizen science platforms that enable the public to record and upload biological observations. With such integration, our algorithms will enable the public to contribute to mosquito surveillance and gonotrophic stage identification. Finally, we are aware of work today that uses computer vision techniques for automated mosquito species identification, and our algorithms in this paper can augment these efforts by enabling the automated detection of gonotrophic stages of mosquitoes as well.

Utilizing Molecular Epidemiology and Citizen Science for the Surveillance of Lagoviruses in Australia

Australia has multiple lagoviruses with differing pathogenicity. The circulation of these viruses was traditionally determined through opportunistic sampling events. In the lead up to the nationwide release of RHDVa-K5 (GI.1aP-GI.1a) in 2017, an existing citizen science program, RabbitScan, was augmented to allow members of the public to submit samples collected from dead leporids for lagovirus testing. This study describes the information obtained from the increased number of leporid samples received between 2015 and 2022 and focuses on the recent epidemiological interactions and evolutionary trajectory of circulating lagoviruses in Australia between October 2020 and December 2022. A total of 2771 samples were tested from January 2015 to December 2022, of which 1643 were lagovirus-positive. Notable changes in the distribution of lagovirus variants were observed, predominantly in Western Australia, where RHDV2-4c (GI.4cP-GI.2) was detected again in 2021 after initially being reported to be present in 2018. Interestingly, we found evidence that the deliberately released RHDVa-K5 was able to establish and circulate in wild rabbit populations in WA. Overall, the incorporation of citizen science approaches proved to be a cost-efficient method to increase the sampling area and enable an in-depth analysis of lagovirus distribution, genetic diversity, and interactions. The maintenance of such programs is essential to enable continued investigations of the critical parameters affecting the biocontrol of feral rabbit populations in Australia, as well as to enable the detection of any potential future incursions.

Global mosquito observations dashboard (GMOD): creating a user-friendly web interface fueled by citizen science to monitor invasive and vector mosquitoes

BACKGROUND

Mosquitoes and the diseases they transmit pose a significant public health threat worldwide, causing more fatalities than any other animal. To effectively combat this issue, there is a need for increased public awareness and mosquito control. However, traditional surveillance programs are time-consuming, expensive, and lack scalability. Fortunately, the widespread availability of mobile devices with high-resolution cameras presents a unique opportunity for mosquito surveillance. In response to this, the Global Mosquito Observations Dashboard (GMOD) was developed as a free, public platform to improve the detection and monitoring of invasive and vector mosquitoes through citizen science participation worldwide.

METHODS

GMOD is an interactive web interface that collects and displays mosquito observation and habitat data supplied by four datastreams with data generated by citizen scientists worldwide. By providing information on the locations and times of observations, the platform enables the visualization of mosquito population trends and ranges. It also serves as an educational resource, encouraging collaboration and data sharing. The data acquired and displayed on GMOD is freely available in multiple formats and can be accessed from any device with an internet connection.

RESULTS

Since its launch less than a year ago, GMOD has already proven its value. It has successfully integrated and processed large volumes of real-time data (~ 300,000 observations), offering valuable and actionable insights into mosquito species prevalence, abundance, and potential distributions, as well as engaging citizens in community-based surveillance programs.

CONCLUSIONS

GMOD is a cloud-based platform that provides open access to mosquito vector data obtained from citizen science programs. Its user-friendly interface and data filters make it valuable for researchers, mosquito control personnel, and other stakeholders. With its expanding data resources and the potential for machine learning integration, GMOD is poised to support public health initiatives aimed at reducing the spread of mosquito-borne diseases in a cost-effective manner, particularly in regions where traditional surveillance methods are limited. GMOD is continually evolving, with ongoing development of powerful artificial intelligence algorithms to identify mosquito species and other features from submitted data. The future of citizen science holds great promise, and GMOD stands as an exciting initiative in this field.

Decision-support tools to build climate resilience against emerging infectious diseases in Europe and beyond

Climate change is one of several drivers of recurrent outbreaks and geographical range expansion of infectious diseases in Europe. We propose a framework for the co-production of policy-relevant indicators and decision-support tools that track past, present, and future climate-induced disease risks across hazard, exposure, and vulnerability domains at the animal, human, and environmental interface. This entails the co-development of early warning and response systems and tools to assess the costs and benefits of climate change adaptation and mitigation measures across sectors, to increase health system resilience at regional and local levels and reveal novel policy entry points and opportunities. Our approach involves multi-level engagement, innovative methodologies, and novel data streams. We take advantage of intelligence generated locally and empirically to quantify effects in areas experiencing rapid urban transformation and heterogeneous climate-induced disease threats. Our goal is to reduce the knowledge-to-action gap by developing an integrated One Health-Climate Risk framework.

Nondestructive Methods of Pathogen Detection: Importance of Mosquito Integrity in Studies of Disease Transmission and Control

Mosquitoes are vectors of many pathogens, including viruses, protozoans, and helminths, spreading these pathogens to humans as well as to wild and domestic animals. As the identification of species and the biological characterization of mosquito vectors are cornerstones for understanding patterns of disease transmission, and the design of control strategies, we conducted a literature review on the current use of noninvasive and nondestructive techniques for pathogen detection in mosquitoes, highlighting the importance of their taxonomic status and systematics, and some gaps in the knowledge of their vectorial capacity. Here, we summarized the alternative techniques for pathogen detection in mosquitoes based on both laboratory and field studies. Parasite infection and dissemination by mosquitoes can also be obtained via analyses of saliva- and excreta-based techniques or of the whole mosquito body, using a near-infrared spectrometry (NIRS) approach. Further research should be encouraged to seek strategies for detecting target pathogens while preserving mosquito morphology, especially in biodiversity hotspot regions, thus enabling the discovery of cryptic or new species, and the determination of more accurate taxonomic, parasitological, and epidemiological patterns.

A Habitat Model for Disease Vector Aedes aegypti in the Tampa Bay Area, FloridA

Within the contiguous USA, Florida is unique in having tropical and subtropical climates, a great abundance and diversity of mosquito vectors, and high rates of human travel. These factors contribute to the state being the national ground zero for exotic mosquito-borne diseases, as evidenced by local transmission of viruses spread by Aedes aegypti, including outbreaks of dengue in 2022 and Zika in 2016. Because of limited treatment options, integrated vector management is a key part of mitigating these arboviruses. Practical knowledge of when and where mosquito populations of interest exist is critical for surveillance and control efforts, and habitat predictions at various geographic scales typically rely on ecological niche modeling. However, most of these models, usually created in partnership with academic institutions, demand resources that otherwise may be too time-demanding or difficult for mosquito control programs to replicate and use effectively. Such resources may include intensive computational requirements, high spatiotemporal resolutions of data not regularly available, and/or expert knowledge of statistical analysis. Therefore, our study aims to partner with mosquito control agencies in generating operationally useful mosquito abundance models. Given the increasing threat of mosquito-borne disease transmission in Florida, our analytic approach targets recent Ae. aegypti abundance in the Tampa Bay area. We investigate explanatory variables that: 1) are publicly available, 2) require little to no preprocessing for use, and 3) are known factors associated with Ae. aegypti ecology. Out of our 4 final models, none required more than 5 out of the 36 predictors assessed (13.9%). Similar to previous literature, the strongest predictors were consistently 3- and 4-wk temperature and precipitation lags, followed closely by 1 of 2 environmental predictors: land use/land cover or normalized difference vegetation index. Surprisingly, 3 of our 4 final models included one or more socioeconomic or demographic predictors. In general, larger sample sizes of trap collections and/or citizen science observations should result in greater confidence in model predictions and validation. However, given disparities in trap collections across jurisdictions, individual county models rather than a multicounty conglomerate model would likely yield stronger model fits. Ultimately, we hope that the results of our assessment will enable more accurate and precise mosquito surveillance and control of Ae. aegypti in Florida and beyond.

Using community science data to assess the association between urbanization and the presence of invasive Aedes species in Hungary

BACKGROUND

Urbanization can be a significant contributor to the spread of invasive mosquito vector species, and the diseases they carry, as urbanized habitats provide access to a great density of food resources (humans and domestic animals) and offer abundant breeding sites for these vectors. Although anthropogenic landscapes are often associated with the presence of invasive mosquito species, we still have little understanding about the relationships between some of these and the built environment.

METHODS

This study explores the association between urbanization level and the occurrence of invasive Aedes species, specifically Aedes albopictus, Aedes japonicus, and Aedes koreicus, in Hungary, using data from a community (or citizen) science program undertaken between 2019 and 2022.

RESULTS

The association between each of these species and urbanized landscapes within an extensive geographic area was found to differ. Using the same standardized approach, Ae. albopictus showed a statistically significant and positive relationship with urbanization, whereas Ae. japonicus and Ae. koreicus did not.

CONCLUSIONS

The findings highlight the importance of community science to mosquito research, as the data gathered using this approach can be used to make qualitative comparisons between species to explore their ecological requirements.

Artificial intelligence versus natural selection: Using computer vision techniques to classify bees and bee mimics

Many groups of stingless insects have independently evolved mimicry of bees to fool would-be predators. To investigate this mimicry, we trained artificial intelligence (AI) algorithms—specifically, computer vision—to classify citizen scientist images of bees, bumble bees, and diverse bee mimics. For detecting bees and bumble bees, our models achieved accuracies of 91.71% and 88.86%, respectively. As a proxy for a natural predator, our models were poorest in detecting bee mimics that exhibit both aggressive and defensive mimicry. Using the explainable AI method of class activation maps, we validated that our models learn from appropriate components within the image, which in turn provided anatomical insights. Our t-SNE plot yielded perfect within-group clustering, as well as between-group clustering that grossly replicated the phylogeny. Ultimately, the transdisciplinary approaches herein can enhance global citizen science efforts as well as investigations of mimicry and morphology of bees and other insects.

•

AI models for classifying bees and bumble bees achieved 92% and 89% accuracy

•

AI models were fooled most by bee mimics exhibiting both aggressive and defensive mimicry

•

Class activation maps explained the anatomical reasoning of AI model classifications

•

t-SNE plot exhibited perfect phylogenetic clustering within and between groups