Malaria is an infectious disease caused by Plasmodium parasites, transmitted through mosquito bites. Symptoms include fever, headache, and vomiting, and in severe cases, seizures and coma. The World Health Organization reports that there were 228 million cases and 405,000 deaths in 2018, with Africa representing 93% of total cases and 94% of total deaths. Rapid diagnosis and subsequent treatment are the most effective means to mitigate the progression into serious symptoms. However, many fatal cases have been attributed to poor access to healthcare resources for malaria screenings. In these low-resource settings, the use of light microscopy on a thin blood smear with Giemsa stain is used to examine the severity of infection, requiring tedious and manual counting by a trained technician. To address the malaria endemic in Africa and its coexisting socioeconomic constraints, we propose an automated, mobile phone-based screening process that takes advantage of already existing resources. Through the use of convolutional neural networks (CNNs), we utilize a SSD multibox object detection architecture that rapidly processes thin blood smears acquired via light microscopy to isolate images of individual red blood cells with 90.4% average precision. Then we implement a FSRCNN model that upscales 32 × 32 low-resolution images to 128 × 128 high-resolution images with a PSNR of 30.2, compared to a baseline PSNR of 24.2 through traditional bicubic interpolation. Lastly, we utilize a modified VGG16 CNN that classifies red blood cells as either infected or uninfected with an accuracy of 96.5% in a balanced class dataset. These sequential models create a streamlined screening platform, giving the healthcare provider the number of malaria-infected red blood cells in a given sample. Our deep learning platform is efficient enough to operate exclusively on low-tier smartphone hardware, eliminating the need for high-speed internet connection.

No abstract

Obstacle crossing requires visuospatial working memory to guide the trailing leg trajectory when vision in unavailable. Visuospatial working memory, as assessed with neuropsychological tests, declines with age, however, this remains to be investigated functionally in obstacle crossing. There is also evidence that visuospatial encoding during a secondary task interferes with balance control during stepping and walking in older people. Here, we studied the interaction effects of age by delay (study 1) and age by secondary visuospatial task (study 2) conditions on obstacle clearance in a visuospatial working memory -guided obstacle crossing task. Healthy young adults aged 19 to 36 years (n = 20 in study 1 and n = 17 in study 2) and healthy older adults aged 66 to 83 years (n = 29 in study 1 and n = 21 in study 2) were instructed to step over an obstacle with their leading leg and straddle it for a delay period before completing the crossing with their trailing leg. In study 1, two obstacle height conditions (12 cm, 18 cm) and two delay durations (20 s, 60 s) were presented in random order. In study 2, participants were required to attend to either no secondary task (control), a visuospatial secondary (star movement) task, or a nonspatial secondary (arithmetic) task, while straddling the obstacle for a delay duration of 20 s, at obstacle heights of 12 cm and 18 cm, randomly presented. Trailing leg kinematics (mean and variability of maximum toe clearance over the obstacle) were determined via motion capture. There were no statistically significant age by delay or age by secondary task interactions. In study 1, toe clearance variability was significantly greater in young adults and increased with increasing delay duration in both groups. In study 2, compared with the control condition, toe clearance variability was significantly greater in the non-spatial secondary task condition but not in the visuospatial condition. Contrary to our hypotheses, these findings suggest that young and older adults alike can store an obstacle representation via visuospatial working memory for durations of at least 60 s and use this information to safely scale their trailing leg over an obstacle. However, the increase in trailing leg toe clearance variability with delay duration suggests that obstacle representation starts to deteriorate even within the first 20 s regardless of age. The finding that undertaking a concurrent arithmetic task impaired visuospatial working memory-guided obstacle clearance suggests a potential increased risk of tripping during obstacle crossing while dual-tasking in both young and older people.

No abstract