Neuroscientists use large data tools to analyze brain cell activity

Beijing Time July 29 news, according to the Daily Science web site, the Big Data age, a single computer does not always find the user most needed solutions. On the contrary, the computing task should be distributed in the cluster of computers, so that cluster computers could analyze large datasets together. This is the technology that Google, Facebook has dug up your browsing history and presented the ads you need. But big data doesn't stop there. The latest technology used to monitor brain activity has brought a lot of information. Understanding these data may give humans a new understanding of how their brains work. To this end, neuroscientists use distributed computing Tools Thunder (Lightning), a lot of research.

The latest technology to monitor brain activity brings with it an unprecedented amount of information.

Thunder is a tool library developed by the Janelia Research Park of the Hughes Medical Institute (Howard Hughes Jude Cato). Thunder can improve the speed of data analysis, which would have been running on a workstation for days or weeks. The relevant research results are published in the July 27 issue of the journal Nature methodology. Scientists at the University of California, Berkeley, Jeremy Freeman, Misha Ahrens and others completed the study.

More importantly, they also use thunder to analyze a microscopic imaging technique. Ahrens and his colleagues used the technique to observe the brain cell activity of zebrafish when responding to visual stimuli. The findings are also published in the journal Nature methodology.

Freeman, Ahrens said the factors that make it difficult for a single computer to go beyond the bottleneck are not just the sheer size of the data, but also the complexity of the data. "When you record information about brain activity, you don't know the best way to sort out the data you need." Each dataset is different. You may have ideas about some of the data, but whether it is meaningful or not remains to be explored. ”

Neuroscientists rarely burst into new ideas when considering data for the first time. On the contrary, the initial study may imply a more promising approach, which requires only a small amount of tweaking and a new computational analysis to make the original data more useful.

That's why the use of turtle-speed computing tools to analyze neurological data can be frustrating. Freeman said: "For some analysis, you can upload data, start running data, and come back the next day." But if you need to adjust the analysis and run it again, you have to wait another night. ”

The imaging results can be analyzed in minutes by using Thunder,freeman and Ahrens, and policy adjustments can be made to the analysis results without the need for long waiting time.

(Responsible editor: Mengyishan)