Researchers have developed MemoCare, an interactive mobile system designed for automated cognitive screening. This system, built with React Native, integrates multiple modalities including spoken responses, spatial orientation, touchscreen interactions, and visuoconstruction. The system utilizes Google Speech-to-Text for transcription and a custom natural language processing approach for scoring, alongside modules for spatial resolution and touch event analysis. A convolutional neural network consensus model, specifically ShuffleNetV2 x1.5, was employed for the drawing task, achieving notable accuracy rates in validation tests. Clinician feedback on the end-to-end workflow was positive, with a median rating of 4 out of 5 across several criteria. AI
IMPACT This system could streamline cognitive assessments, making them more accessible and objective through multimodal data analysis.
RANK_REASON Research paper detailing a new system and its validation. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →