Which first rig should I buy? Part I.

After these challenging quarantine times a lot of postdocs, phd students and undergrads are facing the fact that they have to go back to the lab (which in most of the cases is  safer than anything else..) continue an existing project or even start a new one. In this case the very same question might pop up, namely 'which rig should I buy?'

This simple question can be transformed easily into a super complex problem. You must check many parameters step by step and quite frankly it is quite far from low hanging fruits, especially because these parameters are not independent from one another!

Here I will try to highlight some popular imaging techniques which are all superb in different aspects.

My main parameters are the type of the animal model, the spatial and temporal resolution, the type of the detector and biosensor, the universality of the rig and maybe the most important of all the size of the budget.

Let’s start with the first, the animal model, one of the most important parameters. It can be divided into two parts, the size of the animal and the type of the investigation (in vivo, ex vivo, in toto). If we were considering examining mammalians (mice, rats, cats, monkeys etc.) in vivo chronical experiments I would be looking for the very popular two-photon rigs. These scopes have outstanding parameters regarding sample penetration depth, SNR or optogenetic applications. Unfortunately, they demand special mode locked infrared pulse laser sources, which are quite expensive stuff. Moreover, if the rig has special features like multiple lightpaths and a detector system, it may quickly increase the final pricetag(!).

However, if we have already made our decision, we should consider other options as well. For example, do we really need a three axis (galvo-galvo-resonant) or a dual scanhead (galvo-galvo & resonant-galvo) for our experiments? A Resonant-galvo scanner is ideal for in vivo chronical awake experiments because of the simplicity and functional volume imaging possibility. It is also useful for motion correction algorithms but at the same time it generates gigabytes of unnecessary data every minute. I would say this is a classical trade-off. Nowadays, this is the best option for in vivo imaging because of the slow calcium sensors.

So, for in vivo rodent neuronal population imaging the solution is crystal clear, but what if we ought to measure dendrites or axons in acute slices? Well, in this case I would choose a conventional galvo-galvo scanhead. With the galvo-galvo scanhead we will not face the sample motion problem, we can select the desired ROIs very precisely plus the trajectories are very flexible and fast. Another less known advantage is the size of the final data. The analysis will be much faster if we can reduce the amount of the unnecessary data using line or point scanning mode instead of raster scan. Galvo-galvo scanners  also offer a standalone option for precise optogenetic and photostimulation applications like opsin activation or using caged compounds. All the applications are possible with CW lasers and infrared pulsed lasers as well. This is one of the cheapest solutions for optogenetical questions, but always keep in mind this is just a semi-simultaneous method as opposed to other techniques like holographic activation. About holo…in my opinion it is a quite fascinating technique for optogenetic activation but it also has some notable disadvantages as well. Next time I will write about them in detail.

Now let’s go back to the main question. If you are not interested in functional imaging or your sample is thin (for example retina tissue) a confocal system might be the good choice. It is also an excellent option if the experiment is about anatomy. Surprisingly, a lot of people cannot tell which rig can deliver better image quality the confocal or the two-photon scope? They just think the ‘two’ means better than one. The truth is that  the confocal can give better spatial resolution because of the short excitation wavelength. But if the project demands extremely deep imaging then the only reasonable choice is the cutting edge three-photon microscopy. This technique can provide excellent tissue penetration thanks to the longer excitation wavelength. Another quite interesting feature of it  is its capability of imaging through the skull. In that case you will not go that much deep but the whole experiment could be truly non-invasive. Unfortunately, this method is still quite new, the laser source has low repetition rate, so it cannot be used with a resonant-galvo scanner which would be essential for in vivo. Moreover, the price tag of the laser is insane.

However, if the grant budget can afford high-end stuff, acousto imaging is also a considerable choice. Since it has not got any conventional scanners (the focal point is modulated by ultrasound waves) it is completely silent during acquisition. There is no physical movement during functional scanning, so there are no restrictions regarding the orientation of the image plane. The user can select any point in given volume, which can open new horizons in any project. All of these are also true for photostimulation. Last but not least, real time motion correction is also possible with this kind of technique, which could be a gamechanger in vivo imaging. Of course nothing is perfect so it also has some considerable disadvantages. You must be aware of the question of dispersion, the relatively small FOV and the slow scanning speed if you want to collect information from every single pixel.

If you would like to examine smaller animals like zebrafish, the lightsheet microscopy might be the best choice. Instead of point by scanning, the lightsheet illuminates a whole image plane at the same time and uses fast CMOS camera for acquisition. If the sample is small and transparent, this is the best technique for functional and developmental imaging on the market.

There are other popular useful techniques such as the mesoscope, which is designed for ultralarge field of view. The spatial and temporal resolution is moderate but it can record 300k neurons simultaneously over 5x5 mm FOV.

All in all, I can say that making the right choice is a very complex question for a scientist even if she/he is an expert in this field. Seeking assistance from an application specialist is always a very good start.