Researchers at the Flatiron Institute have developed software to automate the monitoring of neuron activity via a mix of standard computational methods and machine learning techniques.
The researchers demonstrated that the calcium imaging analysis (CaImAn) software can localize active neurons with near-human accuracy, and greater efficiency.
CaImAn facilitates real-time data analysis, and the challenges it faced include analyzing up to 1 terabyte of data generated hourly by calcium imaging; compounding the challenge was noisy data caused by overlapping fluorescent signals from different neurons.
Said Flatiron's Andrea Giovannucci, "The algorithm subtracts the background voices and focuses on a few," showing individual cerebellar granule cells exhibit distinct activity patterns.
From News-Medical
View Full Article
No entries found