Two-photon calcium imaging of neuronal activity

Abstract

In vivo two-photon calcium imaging (2PCI) is a technique used for recording neuronal activity in the intact brain. It is based on the principle that, when neurons fire action potentials, intracellular calcium levels rise, which can be detected using fluorescent molecules that bind to calcium. This Primer is designed for scientists who are considering embarking on experiments with 2PCI. We provide the reader with a background on the basic concepts behind calcium imaging and on the reasons why 2PCI is an increasingly powerful and versatile technique in neuroscience. The Primer explains the different steps involved in experiments with 2PCI, provides examples of what ideal preparations should look like and explains how data are analysed. We also discuss some of the current limitations of the technique, and the types of solutions to circumvent them. Finally, we conclude by anticipating what the future of 2PCI might look like, emphasizing some of the analysis pipelines that are being developed and international efforts for data sharing.

Fig. 1 |. The principles of fluorescence calcium imaging.

a | Mechanism of calcium indicator function. A calcium sensor requires a calcium-binding element and a fluorescent molecule. Genetically encoded calcium indicators (GECIs) are combinations of two proteins, a calcium-binding protein and a fluorescent protein. Action potential firing leads to an influx of Ca²⁺ ions into the cell. Binding of Ca²⁺ to the calcium-sensing module causes a conformational change in the fluorescent molecule, which causes a change in its brightness. b | Relationship between neural spiking and calcium traces. Owing to limitations in signal-to-noise ratio and in temporal resolution, calcium imaging cannot reliably detect individual spikes within trains of action potentials. c | Three steps in calcium imaging: introducing the calcium indicator via direct injection into the brain (step 1); implantation of a cranial window or optical fibre (not shown) to gain optical access into brain tissue and in vivo imaging of changes in fluorescence of the indicator with a two-photon (2P) microscope (step 2); and analysis of the fluorescence images and the extracted calcium traces (step 3).

Fig. 2 |. Types of calcium indicators.

a | Synthetic chemical indicators, which include Indo-1, Fura-2 and Fluo-4, possess a fluorophore domain and a calcium-binding domain (chelating site). When these indicators bind to calcium, their fluorescence intensity increases or decreases. b | Fluorescence resonance energy transfer (FRET)-based genetically encoded calcium indicators (GECIs), including TN-XXL, combine a calcium-sensing domain (for example, troponin) with a pair of complementary fluorescent proteins. In the case of TN-XXL, one donor protein (cyan fluorescent protein (CFP)) has an emission spectrum that highly overlaps with the absorption spectrum of the acceptor protein (citrine, or yellow fluorescent protein). A conformation change in the sensing domain induced by calcium brings the fluorescent proteins into close proximity, such that the FRET acceptor will increase its fluorescence by absorbing a fraction of the energy that would otherwise be emitted as photons by the FRET donor. c | GECIs based on circularly permuted green fluorescent protein (GFP), including GCaMP6s, are single-wavelength indicators in which circularly permuted GFP is linked to the calcium-sensing protein (typically calmodulin) and the M13 peptide from myosin light chain kinase. Upon calcium binding, calmodulin undergoes a conformational change and tightly binds to M13, preventing water molecules from accessing circularly permuted GFP, which in its anionic form fluoresces brightly.

Fig. 3 |. Expressing GECIs in the brain and common window preparations for in vivo imaging.

a | Virus injection directly into the brain. Depending on the virus strain, some cell types may not be transfected (for example, layer 4 with adeno-associated virus 1 (AAV1), AAV5 or AAV9). Triangles and circles represent excitatory and inhibitory neurons, respectively, and squares represent layer 4 spiny stellate neurons. Not shown are systemic virus injections. b | Virus injection strategies: retrograde virus injection strategy for labelling somata of subpopulations of neurons that project to a particular brain region and standard anterograde virus injection strategy for labelling axons. Not shown are trans-synaptic anterograde viruses, or viruses that are injected systemically, such as in the tail vein. c | Transgenic mice can be used for two-photon calcium imaging (2PCI) of specific cortical layers (SCNN1A-Cre), cell classes (SST-Cre) or subcortical brain regions (not shown). SCNN1A-Cre line and SST-Cre line restrict expression to layer 4 cortical neurons and somatostatin inhibitory interneurons, respectively. d | Standard cranial window for imaging neocortex. The skull is removed but the dura and underlying cortex are left intact. e | Partial window for simultaneous electrophysiology. f | Prism for imaging medial structures or across all cortical layers. A prism is inserted into a fissure or through a cut into the cortex. g | Top hat window for imaging hippocampus or striatum. The skull is removed, the cortex aspirated and a glass plug inserted into the brain to gain optical access to deeper brain structures. h | Micro-endoscope with gradient refractive index (GRIN) lens for imaging deep structures or for imaging in freely moving mice with a miniaturized, portable two-photon (2P) microscope. Note that GRIN lenses can be used alone, as an optical relay for a standard 2P microscope, which is ideal for 2PCI in subcortical brain structures in head-fixed mice. i | Cranial windows for in vivo 2PCI can also be implanted in other vertebrate species besides mice/rats, including zebra finches (shown) or non-human primates. BLA, basolateral amygdala; GECI, genetically encoded calcium indicator; VPM, ventral posteromedial nucleus.

Fig. 4 |. Components of a 2P microscope, adaptive optics and scanning approaches for high-speed and volumetric 2PCI.

a | Two-photon (2P) microscope and its components, including the Ti:sapphire laser, a beam expander (used to ensure slight overfilling of the back aperture of the objective lens), a half-wave plate combined with a polarizing cube or a Pockels cell to control and measure the laser power, a shutter to control exposure, scanning mirrors, a scan lens, a tube lens, a dichroic mirror, the objective lens (with high numerical aperture and long working distance), a collection lens, the photomultiplier tubes (PMTs) and a CPU to integrate and control the image acquisition process. b | Adaptive optics for wavefront optimization and improved depth penetration. Top: modified emission path in 2P microscope that includes a wavefront sensor (Shack-Hartmann lens array) to estimate the degree of aberrations, and a deformable mirror to shape the excitation beam to compensate for those aberrations. Bottom: uncompensated wavefront distorted by sample inhomogeneities (left) and optimized wavefront after adaptive optics (right). c | Various scanning patterns in two-photon calcium imaging (2PCI). Dotted lines indicate scanning path. Spacing of dots indicates relative speed of scanning. d | Techniques for imaging volumes. In spatio-temporal multiplexing (left), the excitation beam is split into n beamlets that are slightly delayed in time with respect to one another, which can be used to image n times faster, or in n planes. Fluorescence resulting from each beamlet can be separated in time by the PMT, provided that n × fluorescence decay time of the genetically encoded calcium indicator (GECI) does not exceed the repetition rate of the laser (80 MHz). For piezo-moving objectives (right), two or more planes at different depths can be scanned in quick succession with the use of a piezo motor that moves the objective quickly to each of those depths. AOD, acousto-optical deflector; galvo, galvanometer-mounted mirror.

Fig. 5 |. Examples of good and bad cranial windows and fields of view.

a | Example of good-quality cranial window. Blood vessels can be clearly seen and there is no vascular proliferation, blood or bone growth. b | Example of a window with frank haemorrhage due to puncture of the dura during drilling or lifting the bone flap. c | Examples of windows with bone growth starting at the periphery (left) and excessive bone growth that obscures the middle of the window (right). Asterisks indicate areas of bone growth. d | Example fields of view with GCaMP6 expression in hippocampal pyramidal neurons, layer 4 neurons of barrel cortex and somatostatin neurons of barrel cortex. Scale bars are 100 μm. Part d images courtesy of N. Kourdougli and A. Suresh.

Fig. 6 |. Illustration of analysis pipeline concepts.

a | Analysis pipeline and example relationship to archiving frameworks. b | Example of source localization (left) and activity extraction (right). The blue traces represent the raw calcium signal before demixing. The red, orange and purple traces represent the revised traces after demixing to remove contamination from adjacent neurons. Bottom row represents component contours overlaid on the correlation image (left) and the neuropil contamination signal estimated with CNMF (right). Scale bar in ROI1 (10 μm) is the same for all regions of interest (ROIs). c | Example of high-quality (top) and low-quality (bottom) temporal (left) and spatial (right) sources extracted with CNMF. Scale bar is 10 μm. Traces in panels b and c obtained from data used in REF. d | Example of simultaneous GCaMP6s traces from two-photon calcium imaging (2PCI) (orange) and patch-clamp electrophysiology (blue). Number of ground truth spikes shown at top. GCaMP6s is able to detect isolated single spikes (arrows). Traces obtained from data used in REF. e | Field of view of viral expression of GCaMP8s in primary somatosensory cortex of adult mouse. Scale bar is 100 μm. f | Example of raw fluorescence signal from GCaMP8s (blue trace) from two cortical neurons and corresponding deconvolved signals. For deconvolution we used CaImAn’s OASIS (green trace) or CASCADE (purple trace). ΔF/F, change in fluorescence over the baseline fluorescence; NWB, Neurodata Without Borders.