Intrinsic Imaging (as cheap as possible)
I was recently tasked with building the lab’s first rig for intrinsic optical imaging. The budget according to my PI? “As cheap as possible.” Thus began my weeklong obsession crafting a slick but cheap scope for intrinsic imaging. Along, the way, I came across great resources such as this Nature Protocols paper from Juavinett, et al. and these helpful tips from LabRigger. However, I couldn’t help but wish for more clarity on the details. Also, let’s be real, searching for parts to home-build any type of lab equipment is always a massive pain in the ass. To save anyone else this headache, below you will find a ton of detail, tips, and parts for a rig that I have been very pleased with thus far.
In this post, I’ll discuss the detail of the macroscope itself. I use this setup to target whisker barrels in somatosensory cortex (check out this post if you’re interested in a cheap whisker stimulator), but this scope should generalize to a lot of applications. The parts cost ~$2,500. I also briefly discuss some modifications that could be made.
The Scope
To be honest just click here if you are done reading and want to get straight to the parts list. I don’t want to be like those online dinner recipes that tell you my life story with all the actual ingredients buried in a table at the end. But read on if you’re curious about my thought process in constructing this scope, in addition to helpful tips and alternatives.
If you are familiar with intrinsic imaging at all, you’ve probably heard of a “tandem macro lens” by now. Read more about this here. Most groups achieve this configuration by putting two DLSR camera lenses front-to-front (see Juavinett, et al.). Yes—camera lenses, not objectives. There are a ton of great reasons to do this that are described elsewhere, so I won’t get into that. While the theory behind tandem lenses is prevalent, the actual construction of a scope that uses massive DLSR camera lenses as “objectives” is really difficult to find. My goal here is to provide you everything you need to know to actually build a macroscope.
First, I will highlight two main takeaways that you probably won’t find in the literature:
1) I house the macro lens on a Thorlabs microscope body. I really like this packaging as opposed to the more prevalent boomstand stereoscope, like in Juavinett, et al. As far as I can tell it is a significantly cheaper option as well. In addition, I couldn’t find any off-the-shelf parts to mount the tandem lens to the stereoscope. Does everyone get these custom-machined? If you have the answer, feel free to share and I’ll update this section!
The primary tradeoff with the Thorlabs body is a loss of flexibility in movement, particularly in the XY plane. For my purposes, this tradeoff doesn’t matter much.
2) Potentially the most useful bit of info for many of y’all: I use a ~$600 CMOS camera from Flir and their free SpinView imaging software as opposed to more commonly used Qicams, Teledyne Dalsa M60s or M30s, or Retiga cameras that easily cost several thousands of dollars (even up to $10-12K). I’m really not sure why these specific cameras are so prevalent in the literature. While these cameras are great quality, they are not necessary if your application is to use intrinsic imaging for simply targeting a brain region for ephys or calcium imaging. However, I do suggest being more selective in your choice if you are looking for very high-quality imaging. For example, maybe you need to be able to see the time course of the optical signal, instead of just locating a cortical area. An investment in a pricey camera is maybe warranted for that purpose.
I spent a lot of time looking into these cameras, trying to decide what to do. I gathered the camera specs for these expensive cameras, compared them to more modern CMOS cameras, and reached out to multiple neuro labs that do intrinsic imaging. If you want to compare specs, you can check out this PDF. To put it plainly, I will unequivocally state that this $600 camera absolutely works for intrinsic imaging in barrel cortex. During an experiment, I get 30 fps (individual frames are trigger, 16 bit pixel depth, 2x2 binning, 960x600 pixel image). Here is a shot of the FOV in mm (note that the images quality in real life is very good, weird compression here is causing the awkward artifacts).
A few things to note here and some alternative approaches.
When putting this together, I specifically had in mind the ability to tilt the objective. To achieve this I ordered the rail carriage and objective arm, and had our shop here on campus machine the rotatable aluminum L-bracket. The tilt modality is amazing and critical for proper imaging of barrel cortex. If you don’t need the tilt modality, or just don’t want to go through the trouble of machining custom parts, I suggest trying this alternative objective arm from Thorlabs. It can attach directly to the optical rail without the extra rail carriage component. All of the other components (adapters, etc) pictured for holding the camera and objectives can remain the same.
One downside to both my approach (and the alternative objective arm described in #1) is a lack of z-motion focus. To move the objective up and down you will need to loosen the dovetail adapter attached to the optical rail, move the whole arm assembly to the position you need, then retighten. Keep in mind that you can still focus on the brain using the manual focus on each camera lens (be careful though, you MUST buy a manual focus lens!). If all of this sounds like a pain and you prefer the feel of a more conventional microscope z-focus, try these manual or motorized translation arms from Thorlabs. Again, as long as the objective arm has SM2 internal threading, everything thing else pictured will be compatible. These translation arms are a minimum of $1,600, which is the primary reason I chose to forego easy z-movement when it’s not necessary for my purposes.
You’ll notice the parts list leaves out a light source. I got luckily, and we had a very fancy Lumencor Spectra X multi-wavelength LED laying around. However, this expensive light source is totally unnecessary. A super cheap and easy choice are these RGB LED ring lights from adafruit. Stick one right on the bottom of the camera lens closest to the brain for bright, steady, and even illumination.
To attach the tandem lens to the objective arm, I used a Vello F-mount-to-C-mount adapter and a Thorlabs C-mount-to-SM2 adapter. However, this clunky combination can be replaced by this single Thorlabs F-mount to SM2 adapter. See the detailed PDF where I also make note of this. So…why did I use two components instead of one? Easy….this special adapter currently has an 8 week lead time!
Some groups need to provide visual stimulation during intrinsic imaging, which can lead to unwanted wavelengths obscuring the optical signal. As in Juavinett, et al. I would suggest a bandpass emission filter. In their Nature Protocols paper, it appears they machined a custom part to place a filter between the two camera lenses. Another option would be to mount a filter cube on top of the imaging arm, similar to what shown here on the Thorlabs website. (Note that you don’t need the entire epi-illumination setup unless you are using a white LED for illumination).
That wraps up everything I can think of right now on the scope. Below, you can see results from my first time trying the system out in mouse barrel cortex while stimulating two different whiskers (10 trials each). The barrels are definitely there! Looking forward to improving the rig, protocol, and analysis for cleaner images. Maybe I’ll be back here eventually to post those. (UPDATE: check out my next blog post for an update and improved images. We’re easily resolving 4 adjacent whiskers.)
Final Thoughts
I genuinely hope this post saves some people a ton of time (and some money). Despite tearing my hair out triple checking all these parts were compatible and finding the most efficient way to build a macroscope, I really enjoyed this little project. Definitely email if you found this helpful, have a few questions, and especially if you come up with a neat modification!
-Daniel
