Related Experiment Video
Updated: Jun 11, 2026
04:35
A Rapid Approach to High-Resolution Fluorescence Imaging in Semi-Thick Brain Slices
Published on: July 26, 2011
Automatic identification of fluorescently labeled brain cells for rapid functional imaging
Ilya Valmianski1, Andy Y Shih, Jonathan D Driscoll
1Dept. of Physics 0374, Univ. of California, 9500 Gilman Dr., La Jolla, CA 92093-0374, USA.
Journal of Neurophysiology
|July 9, 2010
Summary
This study introduces a supervised learning algorithm for rapid, automated identification of fluorescently labeled cells in microscopy images. The method optimizes scan paths, significantly speeding up data acquisition in neuroscience research.
Area of Science:
- Neuroscience
- Computational Biology
- Microscopy
Background:
- Automated cell identification is crucial for efficient scanning microscopy.
- Current methods face limitations in speed and adaptability.
Purpose of the Study:
- To develop a supervised learning algorithm for rapid, on-line identification of fluorescently labeled cells.
- To construct an optimized scan path for efficient microscopy data acquisition.
Main Methods:
- Supervised learning algorithm applied to full-field images of fluorescently labeled cortical cells.
- Algorithm trained for automatic tagging of somata and scan path optimization.
- Utilized two-photon laser scanning microscopy with calcium indicators in rat parietal cortex.
Main Results:
- The algorithm automatically identifies approximately 50 cells within 1 minute.
- Achieved sampling rates of approximately 100 Hz with a signal-to-noise ratio of around 10.
- A single classifier demonstrated robustness across different subjects and imaging conditions without retraining.
Conclusions:
- The developed algorithm significantly accelerates the rate-limiting step of cell identification in scanning microscopy.
- The method offers a robust and adaptable solution for automated cell analysis in neuroscience.
- Enables efficient, high-throughput imaging of neural activity in vivo.
