Blog

‹ Back

Blog 2 - 600+ Hours of dLight Recordings: Key Lessons in Fiber Photometry
Chapter 1 – Surgeries

Takeaway: Surgeries take years of practice. I'm still learning tips and tricks. Here are things I've learned over the years that make my surgeries better.

 

 

Here’s my qualifications for writing this post (copied from my first blog post). “One of my PIs introduced me as ‘the person who’s done more dLight fiber photometry than anyone else.’” Whether or not that’s true, at least 170 recording sites across 100 mice later, I’ve certainly logged my hours at the surgery bunch.

 

 

I’m going to present my full workflow but focus on things I’m doing to ensure success with my photometry experiments. I won't comment on alternative approaches very much unless I want to call attention to approaches that are causing people to unknowingly compromise their experiments. The reason for this is simple: everyone has their own way of doing things, and I really want it to be clear that I am not claiming my way is “right.” There is no “right” or “wrong,” these are just the methods that work best for me.

 

1. Planning the surgeries.
Is your experiment a free-moving or head-fixed setup? Are the animals group-housed or single? Will you be recording from multiple sites in one animal? Are you using interesting/complicated genetic/viral targeting strategies?

 

While these details are extremely important to a neuroscientist planning a neuroscience experiment, the engineer in me sees things a little differently. Successful photometry surgery boils down to this: precision. You need to secure the animal and place your fiber with the precision of a brain surgeon—and maybe a little patience.


For fiber photometry of dLight this is injecting virus and placing your optical fiber. I’ve only ever done one ephys surgery (it was also the first/only surgery I have done on a rat at the MINT workshop), but the principle of the procedure is the same. I had to secure the skull so I could implant my optoelectrode precisely where I wanted. Let’s break this procedure down into some engineering principles.

 

Here’s the breakdown for dLight photometry success in terms of engineering characteristics:
I. Precise injection at the target location
II. Fiber placement exactly at the injection site
III. Secure implant to ensure stability for weeks of recording
These steps might sound simple, but as every surgeon knows, getting them right takes more than just a steady hand.

 

2. Doing the surgeries
a) Get mouse on nosecone: place the mouse on the nosecone and make sure the nosecone is pulled tight to minimize isoflurane leakage. Then tape the whiskers aside. I saw Dr. Weijian Zong tape them back when I learned mini2p surgeries in his lab at the Kavli Institute, but he only kept the tape on when he was shaving the skull. We should keep whiskers out of the way throughout the procedure because mice rely on their whiskers almost as much as their eyes to navigate the world. Just like we put eye goop to protect the eyes, we should protect the whiskers by moving them out of the way.


b) Remove hair and clean the scalp: I’ve user scissors, a tiny shaver, and Nair. I like Nair the most. It keeps all the hair together in clumps which prevents it from getting everywhere like with cutting or shaving. After that, I wipe the scalp down with alcohol-provodone/iodine-alcohol-provodone/iodine-alcohol-antiseptic wipe. I know this is overkill, but it gets the scalp super clean since the first couple of passes remove residual Nair and hair clumps.


c) Secure animal in earbars and level the skull: The reason I wait until this point to secure the animal with earbars is so I have more space around the animal’s head while I prep the animal. Securing the animal is one of the most important steps. From here on out the surgery is very similar to machining with a CNC machine, so I’m going to take that perspective. Just like you need to secure the material you want to machine super well to ensure your part is machined correctly with correct tolerances, you must secure the skull as best as you possibly can.


I cannot stress enough how important it is to move the animal to the earbar position while the earbars are loose, and only secure the earbars in place once they are “seated” correctly in the skull. The earbars are going to end up in the position determined by the earbar clamp, and if you angle the earbars to the animal’s head to seat them and then clamp down the earbars, you’re just going to force the animal’s head to move into the earbar's position and possibly damage their skull. After this, I make sure the animal is under before cutting open the scalp. Then I level the skull Bregma-Lambda using my drill burr.


d) Drilling the craniotomy: A stereotax is great for precision, but the moment you start doing things ‘by eye,’ that precision goes out the window. When I was first taught to perform craniotomies for performing cranial windows (shoutout to some people from Liston lab - Dr. Thu Huynh, Dr. Puja Parekh, and David Rosenthal), I learned to touch my leveling instrument (insulin syringe, injection needle, fiber implant, etc.) to a few points on the skull, mark the points with a marker, and then try drilling out a perfectly circular craniotomy by connecting an imaginary circle that connects the points together.


While I got mixed results using this method (especially because I was first learning), I can guarantee that these surgeries were not precise at all. The hole I ended up drilling was based on educated guesswork and nothing more. However, this was inconsequential for imaging experiments through a window since you have room to play with for finding your imaging plane. Photometry implants and surgeries are less forgiving. Unfortunately, I continued this bad habit with the marker on the skull for my first couple of years of photometry surgeries in grad school, and I only grew out of it recently.


I’ve started using a stereotaxic drill holder and using the exact sized drill burr I need (including clearance space) for the size of my implant. For example, I use a .5mm drill bit size for a 400 micron-core optical fiber implant. I’m using a 1.4mm trephine for 1mm GRIN lens implants (for imaging experiments, not photometry). Because the drill is held in the stereotaxic arm, the hole only ends up being a bit bigger than .5mm, not 2-3mm wide like it can sometimes be when I drill by hand right after drinking coffee. Note that you might need to adjust your drill angle for regions like amygdala. Skull curvature varies by animal and getting it right can make all the difference.


e) Injection and implantation: next, I center the injection needle of my Hamilton syringe^I) in the hole I just drilled by using the same stereotaxic precision I had when I drilled (see diagram below). It's like using an edge finder on a CNC where you need to make sure to account for the thickness of the burr when making your measurements.  I usually inject two small volumes (50-150nL) at two depths -- .25 mm above the tip of my fiber and .25 mm below the start of the taper.

After the injection I simply center one of the tapered fibers of my (dual-)implant^II) above the site and slowly lower the implant as it goes into the brain. Because there is only a ~100 micron difference between the size of my craniotomy and the size of my fiber core, I feel confident that I’m implanting my implant right in the injection site, whereas I previously had access to too much brain when drilling by hand, and could easily place my implant slightly off-target from my injection site, but significantly enough to ruin the quality of my recordings.

 

f) Cement implant(s): the last step is securing the implant, and I have been doing this by layering metabond, venus diamond flow (this is a new adhesive Dr. Zong introduced me to), and acrylic that I add a dark-colored chalk^III) to. The metabond is like a base layer for the venus to grab onto, and the acrylic is for distinguishing between animals in a cage (I’ll typically use different colors/make different patterns on the headcaps). I usually remove the earbars and implant holder after I cure the venus with UV light so I have complete access to the skull for applying acrylic.

 

That's the whole procedure. I put the animal back in its home cage on a heat pad and monitor them. I did my best to be thorough, and I will edit this post if there's anything I realize I'm missing when I do surgeries in the coming weeks. If you have any surgery hacks you’d like to share, please submit a comment on the Home page.

 

 

 

Quick asides:

 

I) A quick aside on Hamilton syringes. Hamilton syringes vs glass pipettes: the injection world’s version of a political divide. I’ve dabbled in both, and I feel more comfortable with Hamilton syringes.
 

 

I was initially trained in the Liston lab with WPI-brand Hamilton’s. When I first tried glass pipettes in Dr. Brian Wiltgen’s lab during one of my grad school rotations, I had a hard time seeing the pipette tip. I quickly solved this problem by adding fast blue dye to my virus, but I quickly discovered one of the other big problems with glass pipettes. They clog easily. I tried dozens of pipette pulling recipes using both mechanical and gravity-based pullers, and I find that they clog more frequently than Hamilton syringes.

 

The worst part about pipette clogs is that you must throw out all your virus and start over with a new pipette. This can be extremely stressful while you are on the first of six surgeries you plan to do that day. If I can’t get my Hamilton syringe unclogged, I simply swap out my needle and take the L of whatever volume of virus I lost in the hub of the clogged syringe. Instead of sitting there and praying that my pipette magically unclogs itself so I don’t have to run back to the -80 in the middle of my surgery, I solve the problem and am back to injecting within minutes.


There is a cool technique that one of my Liston labmates, Dr. Tim Spellman, introduced me to that I am still troubleshooting but doesn’t require any injection at all. Instead, you create a film out of your sensor that you deposit onto your implant such that the sensor will start to express directly around the site of the implant as the film dissolves and allows the sensor to diffuse into the brain. If you are using this technique and have tips for me, please send them my way.




 

II) A quick aside on implant type. I have written enough IACUC amendments to know that “reduction” should be a major consideration of all labs that do animal work. The simplest way I can think of “reducing” animal numbers is maximizing the amount of data we can collect from a single animal. I have lots of ideas of how we can tackle this when it comes to the data collected during the experiment itself, and I will share these in a later post.


Right now, I want to focus on the fact that multi-site fiber photometry is almost as easy as single-site, and I implore anyone that is interested in studying multiple brain regions in their studies to consider recording from multiple regions in the same animal (note, this encompasses everyone doing fiber photometry because these don’t need to be two distinct brain regions – they can be the same region in both hemispheres). I’ve been doing a lot of PFC-NAc dual-site recordings with dLight so I can monitor mesocortical and mesolimbic dopamine simultaneously. I’ve also done dual-NAc recordings with two different dLight variants so I could run direct comparisons between variants within animal.


There are just a couple of ways that I have used multi-site recordings to reduce my animal numbers while extracting some meaningful data (I will update this blog once the upcoming dLight paper is published so I can explain by annotating my figure). Here are several other ways you can use multi-site recordings to your advantage (none that I have personally tried, but there are plenty of papers out there that support using these methods for these aims):
 

• Record upstream calcium activity of _____ nucleus simultaneously with downstream neuromodulator activity in _____ region that _____ nucleus projects to. (Calcium in VTA and dLight in NAc.)

• Use Cre-strategies to express _____ sensor in one cell-type in one hemisphere and a complimentary cell-type in the opposite hemisphere to investigate differences in ligand signaling dynamics between these cell-types. (dLight in DRD1-cells in hemisphere one and DRD2-cells in hemisphere two.)

• This last suggestion is super exploratory but is something I’ve wanted to try for a while – see if you can enhance behaviorally time-locked neuromodulator signaling dynamics to a particular behavioral event by artificially inducing plasticity using optogenetics. (Enhance dLight reward response in one hemisphere by stimulating one hemisphere at the time of reward collection such that the dopamine response is conditioned to be larger to future reward collections. This is a super exploratory experiment to look at the “emergence” of conditioned behaviors. Can I modulate conditioning by altering the dopamine release that is reinforcing the conditioning?)
   

Moving on to the other piece of the implant type: the tip. There is a paper that came out several years ago about using tapered fibers for getting depth-resolved photometry measurements out of a single fiber. Instead of recording from a 200micron tall volume of tissue directly below the full width of your implant, you record from a 200micron radial width all along the length of your tapered tip (see diagram below). For a 400 micron core fiber, you are increasing your effective recording volume from 0.05mm^2 to .126mm^2 for a 1mm long taper, a 250% increase in recording volume. 

 

Unfortunately, you are agnostic to the specific depth along the taper that your signals are coming from unless you have some sort of way of raster scanning your excitation light. However, that is a non-issue for most papers that use fiber photometry techniques.


Here’s why: every single fiber photometry paper that claims: “____ region is responsible for facilitating ____ behavior” fails to explain that what they really mean is: “the signals I am recording from the 0.05mm^2 of tissue I am recording from in ____ region suggests that this tiny volume of brain might contribute to this behavior.” Tapered fibers are simply making the original claim more accurate since the larger recording volume more closely represents the activity of ____ region than the smaller volume you get with a blunt fiber.






III. A quick aside on black chalk. Seeing as fiber photometry measurements are measuring light, these measurements are extremely sensitive to all sources of “noisy light.” We recently had a setback with our photometry measurements where we realized that a small LED from our behavior testing apparatus was leaking through the animal’s skull and presenting an artifact in our recordings. In the figure below I have a raw fiber photometry trace from a brain region where there isn't much fluorescence signal. The dashed lines indicate times when the LED is flashed on for a couple of seconds, and you can clearly see how the light from the LED traveling through the animal's skull modulated the photometry signal (shout out to Tanner Stevenson on bringing my attention to this problem).

 

We ran a bunch of troubleshooting experiments, and while the small slit in the sleeve used to couple fiber implants and patch cords does allow for a small amount of light leakage, this pales in comparison to the amount of light that comes through the skull itself. I use the black chalk in my final layer of adhesive to minimize the amount of light that can pass through the top of the skull. It also serves as a nice way to tell which animal is which because I can use different colors of chalk or draw different patterns in the acrylic.

 

 

‹ Back

Blog 2 - 600+ Hours of dLight Recordings: Key Lessons in Fiber Photometry
Chapter 1 – Surgeries

Takeaway: Surgeries take years of practice. I'm still learning tips and tricks. Here are things I've learned over the years that make my surgeries better.

 

 

Here’s my qualifications for writing this post (copied from my first blog post). “One of my PIs introduced me as ‘the person who’s done more dLight fiber photometry than anyone else.’” Whether or not that’s true, at least 170 recording sites across 100 mice later, I’ve certainly logged my hours at the surgery bunch.

 

 

I’m going to present my full workflow but focus on things I’m doing to ensure success with my photometry experiments. I won't comment on alternative approaches very much unless I want to call attention to approaches that are causing people to unknowingly compromise their experiments. The reason for this is simple: everyone has their own way of doing things, and I really want it to be clear that I am not claiming my way is “right.” There is no “right” or “wrong,” these are just the methods that work best for me.

 

1. Planning the surgeries.
Is your experiment a free-moving or head-fixed setup? Are the animals group-housed or single? Will you be recording from multiple sites in one animal? Are you using interesting/complicated genetic/viral targeting strategies?

 

While these details are extremely important to a neuroscientist planning a neuroscience experiment, the engineer in me sees things a little differently. Successful photometry surgery boils down to this: precision. You need to secure the animal and place your fiber with the precision of a brain surgeon—and maybe a little patience.


For fiber photometry of dLight this is injecting virus and placing your optical fiber. I’ve only ever done one ephys surgery (it was also the first/only surgery I have done on a rat at the MINT workshop), but the principle of the procedure is the same. I had to secure the skull so I could implant my optoelectrode precisely where I wanted. Let’s break this procedure down into some engineering principles.

 

Here’s the breakdown for dLight photometry success in terms of engineering characteristics:
I. Precise injection at the target location
II. Fiber placement exactly at the injection site
III. Secure implant to ensure stability for weeks of recording
These steps might sound simple, but as every surgeon knows, getting them right takes more than just a steady hand.

 

2. Doing the surgeries
a) Get mouse on nosecone: place the mouse on the nosecone and make sure the nosecone is pulled tight to minimize isoflurane leakage. Then tape the whiskers aside. I saw Dr. Weijian Zong tape them back when I learned mini2p surgeries in his lab at the Kavli Institute, but he only kept the tape on when he was shaving the skull. We should keep whiskers out of the way throughout the procedure because mice rely on their whiskers almost as much as their eyes to navigate the world. Just like we put eye goop to protect the eyes, we should protect the whiskers by moving them out of the way.


b) Remove hair and clean the scalp: I’ve user scissors, a tiny shaver, and Nair. I like Nair the most. It keeps all the hair together in clumps which prevents it from getting everywhere like with cutting or shaving. After that, I wipe the scalp down with alcohol-provodone/iodine-alcohol-provodone/iodine-alcohol-antiseptic wipe. I know this is overkill, but it gets the scalp super clean since the first couple of passes remove residual Nair and hair clumps.


c) Secure animal in earbars and level the skull: The reason I wait until this point to secure the animal with earbars is so I have more space around the animal’s head while I prep the animal. Securing the animal is one of the most important steps. From here on out the surgery is very similar to machining with a CNC machine, so I’m going to take that perspective. Just like you need to secure the material you want to machine super well to ensure your part is machined correctly with correct tolerances, you must secure the skull as best as you possibly can.


I cannot stress enough how important it is to move the animal to the earbar position while the earbars are loose, and only secure the earbars in place once they are “seated” correctly in the skull. The earbars are going to end up in the position determined by the earbar clamp, and if you angle the earbars to the animal’s head to seat them and then clamp down the earbars, you’re just going to force the animal’s head to move into the earbar's position and possibly damage their skull. After this, I make sure the animal is under before cutting open the scalp. Then I level the skull Bregma-Lambda using my drill burr.


d) Drilling the craniotomy: A stereotax is great for precision, but the moment you start doing things ‘by eye,’ that precision goes out the window. When I was first taught to perform craniotomies for performing cranial windows (shoutout to some people from Liston lab - Dr. Thu Huynh, Dr. Puja Parekh, and David Rosenthal), I learned to touch my leveling instrument (insulin syringe, injection needle, fiber implant, etc.) to a few points on the skull, mark the points with a marker, and then try drilling out a perfectly circular craniotomy by connecting an imaginary circle that connects the points together.


While I got mixed results using this method (especially because I was first learning), I can guarantee that these surgeries were not precise at all. The hole I ended up drilling was based on educated guesswork and nothing more. However, this was inconsequential for imaging experiments through a window since you have room to play with for finding your imaging plane. Photometry implants and surgeries are less forgiving. Unfortunately, I continued this bad habit with the marker on the skull for my first couple of years of photometry surgeries in grad school, and I only grew out of it recently.


I’ve started using a stereotaxic drill holder and using the exact sized drill burr I need (including clearance space) for the size of my implant. For example, I use a .5mm drill bit size for a 400 micron-core optical fiber implant. I’m using a 1.4mm trephine for 1mm GRIN lens implants (for imaging experiments, not photometry). Because the drill is held in the stereotaxic arm, the hole only ends up being a bit bigger than .5mm, not 2-3mm wide like it can sometimes be when I drill by hand right after drinking coffee. Note that you might need to adjust your drill angle for regions like amygdala. Skull curvature varies by animal and getting it right can make all the difference.


e) Injection and implantation: next, I center the injection needle of my Hamilton syringe^I) in the hole I just drilled by using the same stereotaxic precision I had when I drilled (see diagram below). It's like using an edge finder on a CNC where you need to make sure to account for the thickness of the burr when making your measurements.  I usually inject two small volumes (50-150nL) at two depths -- .25 mm above the tip of my fiber and .25 mm below the start of the taper.

After the injection I simply center one of the tapered fibers of my (dual-)implant^II) above the site and slowly lower the implant as it goes into the brain. Because there is only a ~100 micron difference between the size of my craniotomy and the size of my fiber core, I feel confident that I’m implanting my implant right in the injection site, whereas I previously had access to too much brain when drilling by hand, and could easily place my implant slightly off-target from my injection site, but significantly enough to ruin the quality of my recordings.

 

f) Cement implant(s): the last step is securing the implant, and I have been doing this by layering metabond, venus diamond flow (this is a new adhesive Dr. Zong introduced me to), and acrylic that I add a dark-colored chalk^III) to. The metabond is like a base layer for the venus to grab onto, and the acrylic is for distinguishing between animals in a cage (I’ll typically use different colors/make different patterns on the headcaps). I usually remove the earbars and implant holder after I cure the venus with UV light so I have complete access to the skull for applying acrylic.

 

That's the whole procedure. I put the animal back in its home cage on a heat pad and monitor them. I did my best to be thorough, and I will edit this post if there's anything I realize I'm missing when I do surgeries in the coming weeks. If you have any surgery hacks you’d like to share, please submit a comment on the Home page.

 

 

 

Quick asides:

 

I) A quick aside on Hamilton syringes. Hamilton syringes vs glass pipettes: the injection world’s version of a political divide. I’ve dabbled in both, and I feel more comfortable with Hamilton syringes.
 

 

I was initially trained in the Liston lab with WPI-brand Hamilton’s. When I first tried glass pipettes in Dr. Brian Wiltgen’s lab during one of my grad school rotations, I had a hard time seeing the pipette tip. I quickly solved this problem by adding fast blue dye to my virus, but I quickly discovered one of the other big problems with glass pipettes. They clog easily. I tried dozens of pipette pulling recipes using both mechanical and gravity-based pullers, and I find that they clog more frequently than Hamilton syringes.

 

The worst part about pipette clogs is that you must throw out all your virus and start over with a new pipette. This can be extremely stressful while you are on the first of six surgeries you plan to do that day. If I can’t get my Hamilton syringe unclogged, I simply swap out my needle and take the L of whatever volume of virus I lost in the hub of the clogged syringe. Instead of sitting there and praying that my pipette magically unclogs itself so I don’t have to run back to the -80 in the middle of my surgery, I solve the problem and am back to injecting within minutes.


There is a cool technique that one of my Liston labmates, Dr. Tim Spellman, introduced me to that I am still troubleshooting but doesn’t require any injection at all. Instead, you create a film out of your sensor that you deposit onto your implant such that the sensor will start to express directly around the site of the implant as the film dissolves and allows the sensor to diffuse into the brain. If you are using this technique and have tips for me, please send them my way.




 

II) A quick aside on implant type. I have written enough IACUC amendments to know that “reduction” should be a major consideration of all labs that do animal work. The simplest way I can think of “reducing” animal numbers is maximizing the amount of data we can collect from a single animal. I have lots of ideas of how we can tackle this when it comes to the data collected during the experiment itself, and I will share these in a later post.


Right now, I want to focus on the fact that multi-site fiber photometry is almost as easy as single-site, and I implore anyone that is interested in studying multiple brain regions in their studies to consider recording from multiple regions in the same animal (note, this encompasses everyone doing fiber photometry because these don’t need to be two distinct brain regions – they can be the same region in both hemispheres). I’ve been doing a lot of PFC-NAc dual-site recordings with dLight so I can monitor mesocortical and mesolimbic dopamine simultaneously. I’ve also done dual-NAc recordings with two different dLight variants so I could run direct comparisons between variants within animal.


There are just a couple of ways that I have used multi-site recordings to reduce my animal numbers while extracting some meaningful data (I will update this blog once the upcoming dLight paper is published so I can explain by annotating my figure). Here are several other ways you can use multi-site recordings to your advantage (none that I have personally tried, but there are plenty of papers out there that support using these methods for these aims):
 

• Record upstream calcium activity of _____ nucleus simultaneously with downstream neuromodulator activity in _____ region that _____ nucleus projects to. (Calcium in VTA and dLight in NAc.)

• Use Cre-strategies to express _____ sensor in one cell-type in one hemisphere and a complimentary cell-type in the opposite hemisphere to investigate differences in ligand signaling dynamics between these cell-types. (dLight in DRD1-cells in hemisphere one and DRD2-cells in hemisphere two.)

• This last suggestion is super exploratory but is something I’ve wanted to try for a while – see if you can enhance behaviorally time-locked neuromodulator signaling dynamics to a particular behavioral event by artificially inducing plasticity using optogenetics. (Enhance dLight reward response in one hemisphere by stimulating one hemisphere at the time of reward collection such that the dopamine response is conditioned to be larger to future reward collections. This is a super exploratory experiment to look at the “emergence” of conditioned behaviors. Can I modulate conditioning by altering the dopamine release that is reinforcing the conditioning?)
   

Moving on to the other piece of the implant type: the tip. There is a paper that came out several years ago about using tapered fibers for getting depth-resolved photometry measurements out of a single fiber. Instead of recording from a 200micron tall volume of tissue directly below the full width of your implant, you record from a 200micron radial width all along the length of your tapered tip (see diagram below). For a 400 micron core fiber, you are increasing your effective recording volume from 0.05mm^2 to .126mm^2 for a 1mm long taper, a 250% increase in recording volume. 

 

Unfortunately, you are agnostic to the specific depth along the taper that your signals are coming from unless you have some sort of way of raster scanning your excitation light. However, that is a non-issue for most papers that use fiber photometry techniques.


Here’s why: every single fiber photometry paper that claims: “____ region is responsible for facilitating ____ behavior” fails to explain that what they really mean is: “the signals I am recording from the 0.05mm^2 of tissue I am recording from in ____ region suggests that this tiny volume of brain might contribute to this behavior.” Tapered fibers are simply making the original claim more accurate since the larger recording volume more closely represents the activity of ____ region than the smaller volume you get with a blunt fiber.






III. A quick aside on black chalk. Seeing as fiber photometry measurements are measuring light, these measurements are extremely sensitive to all sources of “noisy light.” We recently had a setback with our photometry measurements where we realized that a small LED from our behavior testing apparatus was leaking through the animal’s skull and presenting an artifact in our recordings. In the figure below I have a raw fiber photometry trace from a brain region where there isn't much fluorescence signal. The dashed lines indicate times when the LED is flashed on for a couple of seconds, and you can clearly see how the light from the LED traveling through the animal's skull modulated the photometry signal (shout out to Tanner Stevenson on bringing my attention to this problem).

 

We ran a bunch of troubleshooting experiments, and while the small slit in the sleeve used to couple fiber implants and patch cords does allow for a small amount of light leakage, this pales in comparison to the amount of light that comes through the skull itself. I use the black chalk in my final layer of adhesive to minimize the amount of light that can pass through the top of the skull. It also serves as a nice way to tell which animal is which because I can use different colors of chalk or draw different patterns in the acrylic.