1. Researchers from New York City’s Columbia University and New York University developed a machine learning model that analyzes EEG recordings to identify “covert consciousness” in unresponsive patients following severe brain injury.
2. A group of algorithms determined with 93.4 percent accuracy whether a patient is a suitable candidate for the oft-misdiagnosed corneal refractive surgery.
3. South San Francisco-based precision medicine company MyoKardia’s proprietary AI analysis of smartwatch data was proven to be a reliable screening system for hypertrophic cardiomyopathy.
4. A noninvasive brain-computer interface created by researchers from Carnegie Mellon University in Pittsburgh and the University of Minnesota in Minneapolis allows individuals to control a robotic arm with only their neural signals.
5. The Sonopill, a miniature magnetic robot, was successfully guided through a model colon by an automated arm, potentially replacing painful endoscopic examinations.
6. Alphabet’s Verily developed a deep learning technology that detects diabetic retinopathy with equal or better accuracy than that of trained graders and retina specialists.
7. Researchers from the University of Washington in Seattle trained an algorithm embedded in commercial voice assistants such as the Amazon Echo to recognize audible sounds of agonal breathing, a symptom of cardiac arrest, and alert emergency services.
8. A machine and deep learning algorithm developed at the IBM Research lab in Haifa, Israel, analyzes health records and mammograms to predict the development of breast cancer up to 12 months before its onset.
9. Automated analysis of the wording of Facebook posts was proven to correctly indicate 21 health conditions, including diabetes, anxiety, depression and psychosis.
10. Scientists at the University of Virginia in Charlottesville produced machine learning AI that expedites the process of imaging, diagnosing and treating pediatric gut diseases by analyzing duodenal biopsy images.
11. AI-powered language analysis not only detected speech patterns linked to psychosis, but also identified a new pattern associated with the prodromal phase of psychosis, therefore predicting the later emergence of psychosis with more than 90 percent accuracy.
12. A group of 139 algorithms initially developed for the International Skin Imaging Collaboration challenge in 2018 diagnosed benign and malignant skin lesions more accurately than dermatologists, dermatology residents and general practitioners.
13. Engineers from Rensselaer Polytechnic Institute in Troy, N.Y., and radiologists from Massachusetts General Hospital and Harvard Medical School, both in Boston, developed a deep learning model that produced low-dose CT images with greater speed and accuracy than previous attempts to use less radiation in CT imaging.
14. When combined with genetic data, AI software that analyzes front-facing portraits of patients for facial dysmorphism was able to correctly identify disease-causing genes.
15. Assistance from the HeadXNet deep learning model improved radiologists’ ability to detect brain aneurysms from CT angiography imaging.
16. A natural language processing algorithm developed at Northern Illinois University in DeKalb detects patterns and specific features in infants’ cries to determine their meanings.
17. Smartphones equipped with a machine learning program were able to distinguish between and diagnose common respiratory disorders including asthma and pneumonia by analyzing the sound of a child’s cough.
18. Scientists at Yokohama City University Hospital in Japan trained facial recognition technology to detect high-risk arm movements in ICU patients and predict when they were about to remove their breathing tube or engage in other accidental unsafe behavior.
More articles about AI:
Physician viewpoint: What physicians need to know about the data that powers medical AI
HP Enterprise is using AI, blockchain to advance automation while enhancing data privacy
Most people feel unprepared for the AI-powered future, poll finds: 4 notes