ABSTRACT
Freezing of gait (FOG) is a debilitating problem that markedly impairs the mobility and independence of 38-65% of people with Parkinson's disease. During a FOG episode, patients report that their feet are suddenly and inexplicably "glued" to the floor. The lack of a widely applicable, objective FOG detection method obstructs research and treatment. To address this problem, we organized a 3-month machine-learning contest, inviting experts from around the world to develop wearable sensor-based FOG detection algorithms. 1,379 teams from 83 countries submitted 24,862 solutions. The winning solutions demonstrated high accuracy, high specificity, and good precision in FOG detection, with strong correlations to gold-standard references. When applied to continuous 24/7 data, the solutions revealed previously unobserved patterns in daily living FOG occurrences. This successful endeavor underscores the potential of machine learning contests to rapidly engage AI experts in addressing critical medical challenges and provides a promising means for objective FOG quantification.
Subject(s)
Algorithms , Gait , Machine Learning , Parkinson Disease , Humans , Gait/physiology , Parkinson Disease/diagnosis , Parkinson Disease/physiopathology , Wearable Electronic Devices , Gait Disorders, Neurologic/diagnosis , Gait Disorders, Neurologic/physiopathology , Male , FemaleABSTRACT
BACKGROUND: Although enamel matrix derivative (EMD) has demonstrated the ability to promote angiogenesis and osteogenesis both in vitro and in vivo, the specific elements within the EMD compound responsible for these effects remain unknown. METHODS: Nine different protein pools from a commercially produced EMD were collected based on molecular weight. Six of these pools, along with the complete EMD unfractionated compound and positive and negative controls, were tested for their ability to induce bone formation in a calvarial induction assay. Immunocytochemistry of phosphorylated SMAD1/5/8 (phospho-SMAD), osterix, and vascular endothelial growth factor A (VEGF-A) was carried out at selected time points. Finally, proteomic analysis was completed to determine the specific protein-peptide content of the various osteoinductive pools. RESULTS: One of the lower-molecular-weight pools tested, pool 7, showed bone induction responses significantly greater than those of the other pools and the complete EMD compound and was concentration dependent. Dynamic bone formation rate analysis demonstrated that pool 7 was optimally active at the 5- to 10-µg concentration. It was demonstrated that EMD and pool 7 induced phospho-SMAD, osterix, and VEGF-A, which is indicative of increased bone morphogenetic protein (BMP) signaling. Proteomic composition analysis demonstrated that pool 7 had the highest concentration of the biologically active amelogenin-leucine-rich amelogenin peptide and ameloblastin 17-kDa peptides. CONCLUSIONS: These studies demonstrate that the low-molecular-weight protein pools (7 to 17 kDa) within EMD have greater osteoinductive potential than the commercially available complete EMD compound and that the mechanism of action, in part, is through increased BMP signaling and increased osterix and VEGF-A. With this information, selected components of EMD can now be formulated for optimal osteo- and angio-genesis.