Date Published: July 12, 2017
Publisher: Public Library of Science
Author(s): Yuichiro Hayashi, Satoshi Yawata, Kazuo Funabiki, Takatoshi Hikida, Thomas Abraham.
A combination of genetically-encoded calcium indicators and micro-optics has enabled monitoring of large-scale dynamics of neuronal activity from behaving animals. In these studies, wide-field microscopy is often used to visualize neural activity. However, this method lacks optical sectioning capability, and therefore its axial resolution is generally poor. At present, it is unclear whether wide-field microscopy can visualize activity of densely packed small neurons at cellular resolution. To examine the applicability of wide-field microscopy for small-sized neurons, we recorded calcium activity of dentate granule cells having a small soma diameter of approximately 10 micrometers. Using a combination of high numerical aperture (0.8) objective lens and independent component analysis-based image segmentation technique, activity of putative single granule cell activity was separated from wide-field calcium imaging data. The result encourages wider application of wide-field microscopy in in vivo neurophysiology.
To record neural activity in living animals, extracellular electrophysiological recording has been widely used [1–3]. This technique offers high temporal resolution and low invasion of brain tissue, but recording the same neurons for a long period of time (days to weeks) and to determine their precise location and cell types are difficult. In contrast, calcium imaging with genetically-encoded calcium indicators provides a long-term, cell-type-specific method of recording neural activity [4,5]. Because two-photon microscopy provides deep tissue penetration and optical sectioning capability, it is suitable for in vivo calcium imaging of deep brain structures. However, because each pixel in an image is sampled serially in two-photon microscopy, an inevitable tradeoff exists between the size of imaging area and frame rate. Two-photon microscope using wide-field excitation or multi-beam scanning technique offers wide field of view at high frame rate [6,7]. However, their applicability to in vivo calcium imaging is yet to be demonstrated. On the other hand, wide-field microscopy have also been used for visualizing neural dynamics in various brain structures [8–12]. Unlike conventional two-photon microscopy, all pixels in an image are sampled simultaneously in wide-field microscopy. Therefore, this method is advantageous over two-photon microscopy for high speed, high resolution imaging. However, due to the lack of optical sectioning capability, axial resolution of wide-field microscopy is worse than that of confocal or two-photon microscopy. At present, it is unclear whether wide-field microscopy is applicable to densely packed small cells. To examine the applicability of wide-field fluorescence microscopy for smaller sized neurons, we recorded calcium activity of dentate granule cells (GCs) having a small soma diameter of approximately 10 micrometers. By employing high-numerical aperture (NA) (0.8) objective lens and independent component analysis (ICA)-based cell sorting technique, activity of individual hippocampal GCs in head-restrained mouse were separated.
Wide-field microscopy have been used to image calcium activity in various brain regions from awake behaving animals [10–12]. Noteworthy, owing to its simple mechanism, the method can be miniaturized for imaging from free-moving rodents, which have expanded the scope of applications . A major disadvantage of wide-field microscopy is lacking of optical sectioning capability, and hence poor axial resolution. To separate individual activity of small, densely packed cells, we first used high-NA (0.8) objective lens. This lens yielded high lateral and axial resolution, which is enough to separate individual dentate GCs (Fig 1D and S2 Fig). Second, we used PCA-ICA based cell sorting software (Cellsort) (Fig 3B) to separate overlapping fluorescence from adjacent cells. Spatial filters derived from the software appear to represent individual GCs with minimal overlap (Fig 3B), suggesting that the software successfully extracted individual cell activity. Activity traces of the dentate GCs during rest and running on the treadmill were highly sparse (Fig 3C), consistent with previous electrophysiological studies [26,28–31] and calcium imaging studies [26,27]. The two calcium imaging studies using two-photon microscopy observed extremely low calcium transient rate (0.1~0.5 events/min)[26,27]. Our present result (0.17 events/min) was in the range of previously reported values, also indicating that our method successfully extracted calcium activity at cellular resolution. We did not analyzed place-related activity of the GCs, because we did not collect sufficient number of calcium transient for reliable estimation of place field in a recording session (5 min), owing to their extremely low calcium transient rate. Much longer recording time is needed to determine place field properties of the dentate GCs.