Figure 1.

Overview of data collection and implementation for machine learning. Neuro ICU data including patient demographics, clinical findings, waveforms, genetics, lab results, and imaging are collected in the electronic medical records. Data collected from patients can be uploaded onto a universal data cloud for research purposes. Likewise, large datasets from other neuro ICUs can be downloaded for analysis. Machine learning (ML) will then be trained using patient data collected from the neuro ICU and/or cloud. ML can then be used by neuro ICU teams to assist in clinical decision-making for their patients.