Neural map in nanoscale: analysis of serial electron microscope images

  • Hit 126
  • Writer 최고관리자
  • 2017-05-04



▶Subject: Neural map in nanoscale: analysis of serial electron microscope images

▶Speaker: Jinseop Kim, Ph. D. (Korea Brain Research Institute)
▶Date: 4:00 PM/May. 4(Thur.)/2017
▶Place: Life Science Bldg. #104
Mapping the connectivity between individual neurons for each synapse is critical to understanding the functions of neural networks. Since the thickness of neuronal arbors can go below a few hundred nanometers and the scale of synaptic structures is in the order of ten nanometers, electron microscopes (EMs) have been used to study the ultrastructure of neurons and their connectivity. While terabytes of data became available thanks to the advancement in high-throughput serial EM imaging technologies, the methods to analyze those images reconstructing volumetric neural structure from serial images remain primitive. In this talk I will introduce recent discoveries found by analyzing a large stack of EM images. The analysis was performed with help from artificial intelligence powered by deep learning and also from the EyeWire, an online community of ‘citizen neuroscientists’. We will see that anatomical difference in nanoscale discriminates the types of neurons and that sub-cellular connection specificity determines the functions of neurons.

▶Inquiry: Prof. Joung-Hun Kim(279-2347)
* This seminar will be given in Korean.
Please refrain from taking photos during seminars. *