Speeding up the imaging of neural activity. Credit: Marina Corral Spence/Springer Nature

Calcium- or voltage-sensitive probes have made it possible to watch neurons in action. But observing larger neural networks together at the same time is challenging, as many relevant imaging techniques are afflicted by a trade-off between temporal resolution and field of view. In transparent organisms, wide-field techniques such as light-sheet or light-field microscopy can resolve this conundrum. But the situation is more challenging in the light-scattering rodent brain, where imaging neuronal activity with conventional approaches such as two-photon scanning is slow.

Several strategies are available to increase the number of imaged neurons or the temporal resolution, with each strategy having its own advantages and drawbacks. The most intuitive option is to simply increase the scan speed, but this comes at the cost of shorter pixel dwell times, thereby requiring brighter probes. Alternatively, if the neurons of interest are sparsely distributed, they may be selectively imaged by random-access microscopy, which avoids scanning 'empty' areas. Tissue movement can compromise the signals, but this problem can be overcome with a random-access scanning approach (Nat. Methods 13, 1001–1004, 2016; Neuron 92, 723–738, 2016).

Another possible means for increasing the imaging speed is multiplexing: i.e., using multiple imaging beams, which may be temporally interleaved or spatially separated. In sparsely labeled brain regions, multiple layers can be scanned together (by splitting the beam with a spatial light modulator) while the emitted light from the different layers is collected together, and later the signals can be separated computationally (Neuron 89, 269–284, 2016; Neuron 89, 285–299, 2016).

Finally, reducing the number of voxels per volume speeds the imaging process up as well. Instead of scanning a diffraction-limited spot across the sample, the voxel size can be adapted to approach the size of a neuronal cell body, which substantially increases the imaging speed (Nat. Methods 13, 1021–1028, 2016). Alternatively, the axial extent of a voxel may be increased, as in extended-depth-of-field imaging.

While the described general strategies have been around for a while, the recent developments in microscope hardware and the combination of these strategies with advanced computational approaches have certainly moved the field forward. It will be interesting to watch how the boundaries in volumetric imaging speeds are further stretched when the different strategies are combined.