Precision Telemedicine through Crowdsourced Machine Learning: Testing Variability of Crowd Workers for Video-Based Autism Feature Recognition

Mobilized telemedicine is becoming a key, and even necessary, facet of both precision health and precision medicine. In this study, we evaluate the capability and potential of a crowd of virtual workers-defined as vetted members of popular crowdsourcing platforms-to aid in the task of diagnosing autism. We evaluate workers when crowdsourcing the task of providing categorical ordinal behavioral ratings to unstructured public YouTube videos of children with autism and neurotypical controls. To evaluate emerging patterns that are consistent across independent crowds, we target workers from distinct geographic loci on two crowdsourcing platforms: an international group of workers on Amazon Mechanical Turk (MTurk) (N = 15) and Microworkers from Bangladesh (N = 56), Kenya (N = 23), and the Philippines (N = 25). We feed worker responses as input to a validated diagnostic machine learning classifier trained on clinician-filled electronic health records. We find that regardless of crowd platform or targeted country, workers vary in the average confidence of the correct diagnosis predicted by the classifier. The best worker responses produce a mean probability of the correct class above 80% and over one standard deviation above 50%, accuracy and variability on par with experts according to prior studies. There is a weak correlation between mean time spent on task and mean performance (r = 0.358, p = 0.005). These results demonstrate that while the crowd can produce accurate diagnoses, there are intrinsic differences in crowdworker ability to rate behavioral features. We propose a novel strategy for recruitment of crowdsourced workers to ensure high quality diagnostic evaluations of autism, and potentially many other pediatric behavioral health conditions. Our approach represents a viable step in the direction of crowd-based approaches for more scalable and affordable precision medicine.

Drought reshuffles plant phenology and reduces the foraging benefit of green-wave surfing for a migratory ungulate

To increase resource gain, many herbivores pace their migration with the flush of nutritious plant green-up that progresses across the landscape (termed "green-wave surfing"). Despite concerns about the effects of climate change on migratory species and the critical role of plant phenology in mediating the ability of ungulates to surf, little is known about how drought shapes the green wave and influences the foraging benefits of migration. With a 19 year dataset on drought and plant phenology across 99 unique migratory routes of mule deer (Odocoileus hemionus) in western Wyoming, United States, we show that drought shortened the duration of spring green-up by approximately twofold (2.5 weeks) and resulted in less sequential green-up along migratory routes. We investigated the possibility that some routes were buffered from the effects of drought (i.e., routes that maintained long green-up duration irrespective of drought intensity). We found no evidence of drought-buffered routes. Instead, routes with the longest green-up in non-drought years also were the most affected by drought. Despite phenological changes along the migratory route, mule deer closely followed drought-altered green waves during migration. Migrating deer did not experience a trophic mismatch with the green wave during drought. Instead, the shorter window of green-up caused by drought reduced the opportunity to accumulate forage resources during rapid spring migrations. Our work highlights the synchronization of phenological events as an important mechanism by which climate change can negatively affect migratory species by reducing the temporal availability of key food resources. For migratory herbivores, climate change poses a new and growing threat by altering resource phenology and diminishing the foraging benefit of migration.