Building Pediatric and Clinical Data Pipelines for MNE-Python
Daniel McCloy (University of Washington)
To enhance MNE-Python for clinical neuroscience uses by improving spectral and spectro-temporal data handling, and by providing standardized preprocessing pipelines for data.
MNE-Python is widely used in neuroscience research, but less so in clinical neuroscience. This proposal aims to make MNE-Python easier to use for clinical and diagnostic purposes by improving support for spectral and spectro-temporal data, as well as by building automated data cleaning and analysis pipelines. MNE-Python’s support for frequency (spectral) and time-frequency (spectro-temporal) data was added piecemeal, and for reasons of expediency, existing data containers and visualization methods were re-used and expanded to accommodate these data types. This work will upgrade our spectral and spectro-temporal classes to “first-class citizens” with purpose-built analysis and visualization methods. This will support a range of clinical neuroscience applications, where spectral and spectro-temporal representations of brain signals are used for classification or diagnosis (e.g., analysis of sleep states or epileptic seizures). Expanding the automated data cleaning pipelines for pediatric and adult data will also make MNE-Python better suited for clinical use. To complement these development goals, the team will continue outreach efforts started during the previous grant period by providing stipends for new developer mentoring,code-of-conduct enforcement training, and managing collaboration server costs.