Ch 3 Flashcards
Why was the system originally set-up in that way?
The original system set-up was used for gcamp monitoring by exciting at 470 nm and 405 nm was the isosbestic signal which is a channel that records the fluorescence devoid of the sensor.
What do the bandpass filters do?
These will allow only the desired wavelength to pass and remove all unwanted wavelengths from the illumination path.
What are LED drivers?
These are the voltage supply of the LEDs by controlling the current that passes the the LED.
When triggered this will allow current flow to a certain rate in order to protect the LED.
What does the f and NA mean for the aspheric lens? What is an aspheric lens?
Aspheric lens are lens that are not a sphere in shape. They are used as they can improve abberations. They allow light to be focussed using the aspheric side with minimal aberration and can be used for collimation by applying light to the other side.
F is the focal length and this is the distance from the lens to where the light converges.
NA is the angle that the light is spread from the lens. The higher NA allows high ligh-gathering.
What does a 425 nm long pass dichroic mirror do?
This will reflect light shorter than 425 nm, such as 405 nm.
Whereas, 470 nm will pass through the mirror.
Therefore, the two beams will occur next to each other.
What is a dichroic mirror?
Glass that can split a beam of light into two beams with differing wavelengths. It does this by reflecing some wavelengths, while letting the selevtive wavelength pass.
Basically its a filter but on a mirror.
Can you use a camera instead of a photodetector for fibre photometry?
For this particular system no.
But with the correct optical devices this has been completed. It allows imaging of a specific field of view.
Why was a 525 nm emission filter used?
This allows 525 nm wavelength of light to pass with a band width of 39 nm.
This was used for the previous set-up where gcamp that has a peak emission of 510 nm. However, this is off-peak for Methoxy-x04.
In any case, this allows only the emitted light to pass and excludes excitation light from the signal.
Why did you choose a 2.5 mm setting for ferrules, and the appropriate parameters for the patch cable and fibre?
We choose 200 um core as this was less invasive and showed good fluorescence in vitro.
We choose 50o as this increased the light cone and was compatible with the low AF cable.
We choose the low AF cable as this would help improve our SNR.
2.5 was compatible with low AF setting at this time.
Where was this Methoxy-x04 spectra acquired?
An online source that allows their data to be shared.
Why were the mirrors swapped for reconfigured system?
We only want 405 nm excitation and we want to collect our signal now at 440 nm. In the first set-up the 498 nm dichroic mirror will reflect light below this, therefore removing our desired signal.
Therefore, this is moved to reflect the LED light to the path as this is desired. Then the 425 nm DC will allow light above this to pass which is desired.
Why were LEDs used?
As they are lower in cost and have a wider availability fo wavelengths.
Why was a 498 nm dichroic mirror used?
This allowed reflection of the excitation light to the launch system and transmission of emitted light through the mirror to the detection pathway.
What are broadband mirrors and why were they used?
These are mirrors that should have near total reflectance of all wavelengths of light and allow modification of the direction of light through the system.
What is a drop-in filter holder?
This was a holder that could attach to the mount rods with screws which can easily be removed and reattached without having to disassemble the whole system.
Why were lens tubes used?
To sheild background light from the emission pathway. Also, we covered this in cardboard to reduce background light.
How was the system realigned?
First, the light output at the end of the patch cable was measured using a light sensor while the launch system was focuseed into the patch cable and the mirror was focused onto the photodetector.
Once this was at its maximum, the fibre was put into high concentration DAPI and these mirrors was modified until the maximum read out was measured on the PD.
After, system calibrations were completed to see if read outs were comparable.
How was the system cleaned?
Weekly the system was sprayed with dust remover and patch cable ferrules were cleaned with lint free tissue and propanol.
Annually the system was cleaned if necessary with propanol, but as this was covered this was not often.
What does photobleaching the cable do?
By leaving light on this causes internal reflectance within the patch cable. As the patch cable contains cladding that can generate autofluorescence, this works to quench these signals before recording.
Does the low AF cable help?
We found that the low AF cable reduced the overall noise/AF of the system and when it was bleached, the AF remained low for the duration of the recording and was relatively stable. This helped retrieve stronger signals.
Why was enzyme remover used?
Some fibres had been used for in vivo measures and therefore, may have some tissue residue. Therefore, enzyme remover helped remove this and clear the fibre.
Why did all recordings have to be completed in the dark?
We did some test recordings with lights on and we found that the photodetector was very sensitive and this background light was picked up by the sensor and therefore reduced our SNR.
For long-term acute and chronic experiments we used a red light bulb as this interfered with the signal less.
What powers were used for system calibration measurements?
This varied across experiments. First we used 0-5 V. Then, we used 0.1 to 1mW/mm2 to reflect in vivo measures.
Why did you reduce the number of repetitions for the system sensitivity measure?
Methoxy-x04 appears to be quite photo unstable. So in attempt to minimise the photobleaching of the dye throughout the experiment, we reduced the time that the fluorophore was being excited.
Why did you use 6-9-month old mice?
For this part of the project, we really wanted to confirm that this approach was a feasible way to monitor plaque pathology. To do so, we wanted the best chance of monitoring a FP plaque signal. Therefore, using mice that had a strong plaque pathology was desirable.
Why were these experiments performed under terminal anaesthetic?
We wanted to acquire signals at increasing depths. As flat fibres have only one illumination and collection depth, it would not be possible in a waking condition, without using a microdrive. Therefore, terminal anaesthetic was required.
How did you align between bregma and lambda?
By using a glass pipette and stereotax, I measured the depth distance from bregma and lambda. I wanted this <50 um as this meant it was less likely to miss my target.
Why were some recordings performed within the surgery room?
At this point thsi was due to availability of space within the lab. We also wanted depth profile experiments to be only under urethane as this had been shown to be less effective on neuronal activity. Therefore, we moved rigs so the surgery room would be available for others.
What does fluovac do?
Absorbs the anaesthetic.
How was the heatmat temperature controlled?
there was a probe underneath the plastic cover that was weight activated and controlled the termperature.
If you were concerned about the bleaching of M04, why did you maintain the 10 repetitions for DP recordings?
As this was a one-off measurement and not occuring every 5-s, it wasnt as big of a concern.
How did you create a linear regression model?
We measured the light output at the end of the fibre using the light sensor at 1-5 V. By fitting a linear line, we could use the coefficients to determine the voltage of LED required for a desired output.
Why did you record at 5000Hz?
Because we are delivering light pulses at a millisecond range and we want to sample around this.
5000 Hz allows sampling of everything.
Why did you choose 440 and 550 nm?
440 nm is the peak of M04 emission spectra so we expect this to boost the fluorescence recorded.
550 nm should be mostly devoid of a M04 signal and should be high in AF as most AF can occur in this wavelength. So this was a trial control channel.
Why did you choose these implant sites?
These three implant sites pass brain areas that are highly dense in plaque pathology and sparse in plaque pathology. This should provide a signal that has a good comparison between high and low fluorescence to allow confirmation that a signal is being detected.
Why did you inject 24-h before?
As in many previous papers and the landmark study, this was shown to provide the best signals for 2PM as it had the least background fluorescence. Therefore, this suggests this is most reflective of plaque staining.
What did you expect/see with the DP experiments before the APK recording?
Despite Methoxy-x04 not being injected yet, plaques can provide strong autofluorescence and therefore we expected once we began to hit our target, we would see a rise in fluorescence in the 440 and more specifically the 550 nm channel. We done this to be able to have a better idea that we are in the correct area as this is a one-chance experimental protocol.
Did you see this with the DP?
Yes, I tended to see a small increase in fluorescence as I reached my target and if the coordinates and fluorescence matched, this is when I would start the recording.
How do you think the autofluorescence of plaques will interfere with the recording?
As the recording channel is so specific to 440 nm with a small bandwidth of 10 nm I believe this signal should be a strong M04 channel.
Why did you perform a 30-minute baseline recording before injection?
So that we had a measure of the fluorescence within the brain without M04 present as a baseline. This would allow us to normalise against this baseline to see small increases in fluorescence.
Will the mouse being under anaesthetic affect the pharmacokinetic properties of M04?
It remaisn unclear as we couldnt acquire a positive signal in acute to compare.
However, the neurinal activity is meant to be maintained, but I would assume it would be slowed due to a enhanced parasynmpathetic nervous system.
Why did you sample at 30-s?
As the change in fluorescence would be fairly slow, a rapid sampling interval that is required for calcium imaging is not needed here. This was used to identify if a slower sampling interval could be used.
Why are some recordings completed with voltage inputs and other irradiances?
At first, I used voltages. But as this could vary from fibre to fibre depending on their functionality, the light output would not be consistent across mice.
Therefore, I implemented an approach to calculate the voltage requried for a chosen range of irradiances so that we could be consistent across recordings.
Why did you choose that brain region for APK?
Subiculcum is the most dense region of plaques in the 5xFAD brain so should allow ius to detect the strongest signal available.
Why did you fix the tissue with paraformaldehyde?
Because this maintains the tissue and cell architecture in a life-like state which allows us to stain the tissue. Additionally, they allow good preservation of the tissue and make it sturdy enough for sectioning and processing.
How long do you fix the tissue and why?
We fix it for 24-h after to allow the process to occur. But no longer as this can make the tissue fragile.
Why do you use sucrose in preparing brain tissue?
For cyroprotection where we want to protect the tissue from freezing processes throughout histology to maintain the tissue structure.
Why did you use 100 um sections?
As my fibre was 200 um in width, we wanted to choose a section thickness that would allow identification of the fibre track.
Why did you use sodium azide?
This works to prevent bacterial infection of the sections and allows them to be preserved for longer.
Why did you stain with TS if you had already injected with M04?
At first, we wanted to confirm that the M04 stain had a similar plaque marking as TS.
Additionally, it could be used as a backup channel if something went wrong with the injection as we were still confirming if the system could detect M04.
What does TS stain?
Like M04, they bind to any type of amyloid as their target is beta-pleated sheets.
What does neurotrace red stain and why did you use it?
It is a nissl stain that is used to mark nissl in the cell soma of neurons within the brain. This is important substance for protein synthesis and channels in neurons.
We first used this to help identify the structure of the brain when imaging. However, we later found that imaging just the blue channel was sufficient for identification of brain regions and further analysis.
What are free-floating sections and why were these used?
This is sections that are floating within a well that contains the specific staining agent.
The reason we used this approach rather than fixing them onto the slides is because this is better at maintaining the brain structure and preventing tissue loss as in fixed conditions its common for some of the section to be washed away. Also, it allows better antibody penetration on both sides and reduces background fluorescence due to better washings.
Why did you stain adjacent sections?
We done this because we wanted to identify the full fibre track. Occasionally this could be at a slight angle and as we are quantifying plaques on the contralateral hemisphere, this sometimes was at an angle. So this approach ensured we would have enough brain tissue either side of the implant so no plaques that should be quantified were missed.
Why do you use PBS in triton X?
We used this because it increases the permability of the section and therefore allows better staining.
Why is gelatin used?
To penetrace the section to hydrate it and maintain its structure as well as allowing it to stick to the section appropriately.
Why did you use epifluorescent microscopy rather than others?
As this allows quick imaging of the whole brain section.
Our goal was to identify the fibre track, get images of the whole section and plaque pathology.
Therefore, this is a more suitable approach as we do not require depth analysis of the full section. While this would be nice, like light sheet microscopy this has high acquisition times, and can be expensive for numerous samples.
What is epifluorescence?
An imaging approach where both the excitation and emission light flow through the same objective.
Why x4? Did you try higher objectives?
Becuase this was suitable to get the plaque pathology across the whole brain section.
Occasionally, I would look at x10 to confirm presence of the fibre track damage.
Why is there differences in the exposure times?
At this point, a lab technician took over my histology and he believed imaging at this exposure provided better images. Perhaps as we were only imaging the blue channel, a lower brightness was fine.
Why did data need to be offsetted and how exactly was this completed?
Within the system there was a slight grounding offset where the baseline signals were ~-0.05V rather than at 0.
To abolish this, the system and DAQ channels were grounded and then we would take an average value of the baseline signal of each channel and add this on to each channels to boost signals to have a baseline at 0.
This would allow us to see a true and more understandable value of the fluorescence measured.
Why did you detect signals that were greather than 0.05V?
This was because throughout all recordings all irradiances that we supplied would have a voltage that was greater than this value.
This meant that by detecting the values in the data matrix that had this voltage or greater, we could identify the times in the mastrix when the LED was on.
How was event times and data extracted?
Using ther detected events from the sync channels, we would take the first value as the event onset.
Then, at each of these data points, the photodetector channel was extracted from data points 3 to 9 to accounr for some offsets in LED power on either side of the pulse.
Then, this data was averaged.
How did you acquire a fluorescence signal for a single concentration?
As the concentration was increased every 2-mins, the median fluorescence was calculated in 2-min time bins.
Why did you calculate median and not mean?
We found that the fluorescence signal can provide some outliers. Specifically in this experiment when completed in a plastic well. Therefore, we found calculating the median would be a better reflection of the fluorescence that is not skewed by these measurements.
How did you determine that a change in fluorescence at a specific concentration was seen?
When we seen that the fluorescence was increasing from baseline.
Why did you choose these three phases for linear modelling and how did you pick them?
Upon analysis we found that there seemed to be a period where the fluorescence did not decrease or increase and we thought this is pribably due to a balanced state between the bleaching of the system/M04 and the increase of M04 concentration.
We picked them by manually determining the time-points that show a change in fluorescence.
Why did you use mean for everything bar system sensitivity?
Because the stirrer sometimes resulted in large false fluorscence recordings.
For DP exps, why did you use the highest power? Was 1mW/mm2 much different from 5V?
Because this provided the best signals. Weaker powers showed the profile but not as intense.
Yes, commonly the 1mm/mm2 was within the upper 4V range.
Why did you normalise against depths 0-500?
Because this was the brain surface and could provide us with the starting fluorescence. Additionally, the plaque pathology should be weak at these upper cortical layers so should present a baseline signal of brain fluorescence with minimal plaque pathology.
Why did you use 3 different normalisation approaches for FF data?
For DP fluorescence analysis it was important to see the small changes in fluorescence from a baseline so we used baseline normalisation.
For HC analysis we wanted to see small changes that will be as sensitive as possible as histology and have a normalisation approach that allows comparison of 0-1 values for correlation.
For APK, we z-scored from the baseline because this allows detection of small increases from the baseline.
Why did you apply a smoothed profile?
To illustrate a better profile of the plaque fluorescence without some noise.
How did you analyse:
- DP F data
- DP corr data
- APK data?
1) baseline normalisation where I found the mean fluorescence signal from 0-500 um. Then for each depth I subtracted this and divided by it.
2) minimum-based normalisation where I subtracted the monomum sgnal and divided it. Then, I correlated with histology DP.
3) Z-scored by subtracting the mean fluorescence from 30-mins and then dividing by the standard error. Then, 5-min time bins were determined.
How did you manually label the fibre track?
By looking at the aligned histology images, we could see some damage where the fibre had been inserted. We would manually mark dots along this track, ensuring we mark appropriately at the brain surface and the deepest region of the track as this is vital for appropriate depth resolution and quantification sampling.
What is the difference between the manually labelled fibre track and the estimated track coordinates?
The manually labelled only has several data points that have been applied to show the track. Therefore, there may only be ~5 data points.
However, for analysis we want the coordinates of the track at the same sampling depths as the photmetry analysis. Therefore, using the appropriate calculations, we can get the AP, ML and DV coordinates at 41 measures from the brain surfacer.
How exactly is the fibre track detected?
First, I manually mark on the sections where I see the fibre track.
Then, using the alignment information, this provided coordinates for each data point I marked on the images.
Then, by taking the deepest and most superficial marked data point, the estimated depth was determined. This was divided by the actual depth of in vivo experiments to achieve a scaling factor.
This scaling factor was multiplied by the recording resolution of 100 um to establish the histology depth measurement resolution.
Then, from the depth of the most superficial point equal sampling of this value was done until the deepest region.
Using these points and the coordinates, a two step regression model using the three dimensions of the coordinates were used.
This would provide AV/ML coordinates for the appropriate sampling depths.
What is the two-step regression model?
Create a linear trend between the AP coordinates and sampling depths and multiply these coefficients by the sampling depths.
Same for ML
Why did you use pearsons correlation coefficient?
Because we wanted to see if there was a linear relationship between the photometry and histology signals to prove that this approach acquires signals that are the same as a well-established approach.
Why would noise be quantified as a plaque in 5xFAD- mice?
As this is an automated approach, using the chosen parameters, sometimes some anatomical features or regions of dust or higher exposure would be wrongly quantified as a plaque.
We would manually optimise the parameters to reduce this, but there needed to be at least one plaque recorded on each section for the alignment code to work and register the image. As it wouldnt recognise the image if it did not have the signal.
Therefore, we would manually remove.
How would you remove the noise plaques?
We would calculate if the mean plaque quantification was greater than 0.5 and if so all values would be set to 1.
Why did you then need to add noise to the 5xFAD- plaque signals?
For correlation analysis we could not provide a vector that had zero values.
To overcome this, we would apply randomly generated values on a low scale and range.
Was this randomly generated noise also added to FAD+?
No.
Why would mice die under urethane?
Because this is a terminal anaesthetic that appears to reduce the function of the lungs of the mice and can make them gaspy, eventually stopping breathing. It seems that males and those with a higher body weight and are older are more susceptible to be unstable under urethane becuase it can be stored in body fat.
What set-up faults occured?
Rarely the fibre would break during implant or the patch cable would reduce output due to twisting.
Why did you not complete correltion analysis on original system data?
As based off fluorescence data it appeared that there was no detection of plaque signalsl.
Therefore, there was no need to confirm the noise.
In what cases was the depth of 4000 not reached?
If the mouse died before the recording was completed.
Why did you use a standard error of 1.5 for exclusion?
approximately 93.3% of the data fall within ±1.5 standard deviations of the mean. This threshold is thus considered a reasonable cutoff for identifying observations that lie significantly outside the expected range.
Stringest approach
n many fields, including statistics, psychology, and biomedical research, using 1.5 SE as a threshold for outlier detection is a common practice or guideline.
Was the experimenter blinded?
They were blinded to the fluorescence data but not to the geneotype.
Why did you do a two way anova?
For fluorescence measurements we had two major dimensions that we wanted to compare which were the genotype and depth.
This was because the goal was to show that the fluorescence was greater in FAD+ mice compared to FAD- mice and that the fluorescence was higher at some depths compared to others.
As data was normally distributed, we used a twi way anova to test these two variables on fluorescence.
Did you do a post-hoc test and what one?
We did the multicomparisons test with bonferroni correction.
We did this because it is conservative and will reduce the risk of false positives. However, because we have a very large group number due to the number of depths, this can perhaps result in an increased number of false negatives.
What is bonferroni test?
Bonferroni test is a post-hoc comparisons test which is used to adjust the significance level when comparing groups statistical tests.
It is completed by dividing the chosen significance value by the number of groups, and comparing each mean and determining if the p value is less than the adjusted p value.
What is the r and p value for pearsons correlation coefficient?
R represents the strength and direction of the linear relationship between two varables.
The p-value tells you the probability of observing the correlation coefficient under the assumption that there is no true correlation in the population.
Why did you use two sample t-tests?
Because I wanted to compare two groups of data that were independent of each other.
They were independent as they were diffferent animals and genotypes.
Why did you show SEM?
Because te same size was different for groups and this shows a standardised way showing the estimated variance surrounding the mean based on the sample size. Additionally, by showing the data points, I am showing the distribution of data.
Why are you comparing against histology?
Because this is the most commonly used approach across various AD papers, it has little error and is well-established, reliable read-out of plaque pathology.
Why do you see increased fluorescence with concentration and irradiance?
Fluorescence –> the number of fluorphore is increased which means that when excited, more photons will be released form the molecule.
Irradiance –> the greater the irradiance, the greater number of photons that will excite the fluorphore which will enhance the energy transfer.
Why do you see nothing at 470?
Because this is not in M04 emission spectra so this wavelength of light does not cause an energy shift, resulting in no fluorescence being emitted.
Why does fluorescence occur at specific wavelengths?
Each color of light has a different amount of energy. Molecules can only absorb light that matches the energy they need to move their electrons to a higher energy level. If the energy doesn’t match, the light is not absorbed and might be reflected instead.
What is sol A?
Control solution with no M04
Why is the signal saturated?
Because the limit of fluorescence measured by this photodetector is 5V and at this higher concentration and irradiance, the fluorescence output is out of this range.
Why do you still see fluorescence even with the control solution?
Because this solution will contain molecules that can provide autofluorescence, just like distilled water or PBS would.
Also, the delivery of this light into the systsem will result in some background fluorescence being recorded.
Why do yous ee some increase at 470?
This occurs with increasing irradiance as by increasing the power of light entering system, it increases the background light that will reach the detector.
However, its important to note that there is no change between concentrations.
There is not a massive difference between collecting emission between 525 nm and 550 nm. Why do we not see such an increase in signal using the 550 nm?
The bandwidth is higher on the 525nm filter (39 nm compared to 10) which means more light will be picked up and this will reach to ~480 nm which has a stronger signal intesntiy in M04 emission spectra.
Also, while only 25 nm difference, 550 nm is right at tail end of emission spectra so the signal output will be minimal.
You say that the fluorescence was increased at 50nm for all irradiances, but some who some increases before this. What is this?
This is noise which can be detected from variations in light power or the magnetic stirrer
What dies the time bin axis mean on 3.3.1D?
We took the average signal over 2-min time bins.
Therefore 60 represents the 60th 2-min time bin.
Why are the irradiances so random?
In these initial experiments, we monitored fluorescence by supplying voltages and later calculated what irradiance this would have been based on the fibre quality.
Later we optimised this.
What does the colour map represent?
As the colour gets brighter it represents a stronger concentration.
What do the error bars in 3.3.1B represent?
The SEM of the fluorescence across illumination repetitions.
Why is it saying 200 nM in methods but showing 300 nM in figs?
For sensitivity experiments on the original system we went up to 300 nM, but after we seen such an increase at lower values, we reduced this to 100 nM.
Mistake –> needs changed in methods and maybe modify figs to 100 nM.
What are the time bins? Why are the linear trends less than the incubation time with the increase starting before addition of M04? (3.3.1)
Here, the incubationw as actually 10 mins and then the data was binned int o 2 min time binds,
Then a bleaching trend was set between data points 1-5 which show even after addition of M04 here, bleaching is maintaining the dominent trend.
Then from points 6-14 there is a balance which show that aroud 20 nM there is change in signal occuring.
After we see a prominent increase where the signal is freater then belaching.
Why were DP recordings for original not completed on both sites?
When first starting in vivo procedures, we aimed to first target the SUB. Then, to make most use from the animal we later increased the number of implant sites.
You say there is an increase in fluorescence from 2000-3000 for site 1 ODP, what is this?
For site 1, the fibre should be entering the septum at this depth which has dense plaque pathology which is shown in image but is hard to see with the lighting.
Why would you see stronger signals at fad-?
We believe the signals recorded were due to autofluorescence of anatomical structures within the brain. At these implant sites, the fibre will pass white matter which can provide an increase n fluorescence.
This may be higher in these examples due to the mouse itself, or behaps the fibre used for fad- was greater functionality as this was before we controlled for irradiance.
Why do you see signals at 470 in ODP? Why are they weaker occasionally?
Because these signals may not be m04 and instead are autofluorescnece.
The prpfiles are not as prominent because 405 nm is lower wavelength which are known to be absobed by tissue greater and therefore increase AF.
If the sensitivity looked good in vitro, why did you not get in vivo signals?
The brain will scatter the light in different ways and reduce the light delivered to the brain tissue.
Also, the concentration of M04 may be lower than what reaches the brain compared to in vitro concentrations.
Also, autofluorescence absorbed and collected at 525 nm will reduce signal to noise and oerhaps mask the M04 signals.
In 3.3.3 what are the white dots over the images?
This was an artefact at the time with the epifluorescent microscope where the opjectives and light were slightly misaligned which resulted in these dots over the centre of the image.
Why can you not see the full fibre track in some images?
Because it will be spread across several images. This was the best image to show the depth of the fibre track.
Why do the FAD- and septum show strong fluorescence at regions which don’t show plaques?
Tissue Composition: Brain tissue contains various molecules that can exhibit autofluorescence. For example, lipofuscin, a pigment composed of lipid-containing residues of lysosomal digestion, is known to fluoresce in the blue spectrum. This can lead to background fluorescence even in areas devoid of cells.
Extracellular Matrix: The extracellular matrix (ECM) of brain tissue contains proteins like collagen and elastin, which can contribute to autofluorescence. These proteins may be distributed throughout the tissue and emit fluorescence upon excitation.
Lipids and Membrane Components: Lipids present in cell membranes and other cellular structures can also contribute to autofluorescence. Brain tissue contains lipid-rich structures such as myelin sheaths, which can emit fluorescence when excited by the appropriate wavelength.
Why is there some damage on the sections?
Tissue processing resulted in small damage. Disadvantage of histology and free floating analysis.
What does dF stand for and what are the units?
dF shows the change in fluorescence from the baseline or minimum signal.
However, as I dont transfer this to a percentage, this is dimentionless and therefor units is just units.
Why do you show the profiles of FAD+/- rather than wavelength?
Because my goal is to show that there is an increase in fad+ compared to fad-. The different wabelenfths were to show that 405 is better than 470.
Why did you calculate the moving median over 500 um?
As this data set has a length of 41 measures. 5 was a perfect balance of reducing the noise of the signals, while maintaining the profile we wanted to see.
Why did you want a 550 nm channel of AF?
Because in the original DP experiments we found that AF was going to contribute to our findings. Therefore, we hypotehsised that by having a channel of autofluorescence may present itself as an appropriate channel to normalise our signal channel against.
550 was chosen as this is a peak fluorescence emission for several molecules in the brain like lipofuscin.
However, we found that the 440 nm channel alone provided such strong channels devoid of AF, that 550 nm channel was not necessary.
Why is the fluorescence measured lower in the new system calibrations?
The bandwidth on the filters are much lower which will reduce the level of background fluorescence that is recorded.
Why did you put the irradiance x-axis in log scale?
Because they are widely sampled so by putting the log scale it allows us to estimate better the trend of fluorescence over irradiance.
Why are the signals in 3.3.4 noisy at lower wavelengths and why are there some large outliers?
Thw light power was too low to cause a sufficient energy transfer so signals are a bit noisier.
Some outliers exist due to the magentic stirrer.
Why is the values negative in 3.3.4?
This is data that has not been offsetted. This was introduced for all in vivo recordings.
Why did you choose these irradiances?
These provided strong in vitro and in vivo singals.
Also, they were relatively low to minimise bleaching of M04.
Why is there no balanced linear trend in 3.3.4?
Why hypothesise that the experimental proceudres we implemented such as better bleaching before recordings and reduced light power has minimised the bleaching of the system.
Also, we believe the signals are much stronger so they are overcoming bleaching much faster.
Did you look into sex/age differences?
No, the sample size was too low in these experiments.
Also, the age range was fairly similar as its known that males especially plaque load can plateua around this age.
Why do you not see any change in signal intensity for septum, wwhere there aer plaques? Does this not mean that this is not feasible for all brain regions?
This implant site if inserted slihglty lateral can pass trhough a ventricle which means that the fbre would need to reenter the brain tissue.
Due to the angle of tissue at this region, there is little force against the fibre which may prevent reinsertion.
No, it means that care would need to be considered when choosing an implant site. Also, that fibres with a sharp tip are desirable to ease insertion.
The ranges in the text do not always reolicate what is seen in fig 3.3.6.
Agree, needs changed.
What does F(40 = *) mean?
40 is the degrees of freedom which is the number of variables that are being compared -1.
F is the statistical output of the 2way anova which can represent the ratio of variance between the variables. A F of 1 suggests little difference, and greater than 1 often represents bigger variance.
Why do you think there was a significantly stronger fluorescence for FAD- for site 1?
According to some histology it would suggest that in fad+ there was more recordings where the fibre may remain in ventricle and not be close to reentering, while in fad- its trying to reenter and pressed up against the tissue which would suggest that there may be increased tissue AD.
However, its hard to be completely sure as some may be damage.
Why do you not mention the post-hoc results for 3.3.6?
We completed post-hoc tests but did not show significnat differences.
We believe this is because the 2way ANOVA was strong enough to detect differences between groups but due to the large group number, the statistical power was too low in cases to detect significant difference between different variables.
Why do you think you dont get significance across depth for site 3?
A major limitation of this analysis is the fact that we have grouped all recordings together. However, it is impossible to get the exact same implant site across mice, which will result in a shift of the fluorescence profile, resulting in a larger error.
How can you confirm that these profiles are all from the same brain regions?
We align these images to the Allen brain atlas and complete histology to achieve a list of brain regions that the fibre passes.
However, this is not the same across each mouse and therefore results in variance of profiles which could be contributing to some lower signal profiles.
An approach could be taking the fluorescence level for each brain region across recordings, but this was out of the scope of this report.
Why in the FAD- images does the fluorecence appear stronger?
Due to the autofluorescence at this wavelenfths, in FAD- the autoexposure of the images can result in enhancement of these structures.
There seems to be fluorescence surrounding the implant site, can this interfere witht he recording?
This occurs due to the tissue damahe which may have some overlap on the sections causing a stronger signal.
In vivo this wouldnt occur as the area we record from has not been damaged yet.
Why do you still see fluorescence profile at 550nm? Why would you see this here but not in the 525 nm ODP? Is this really AF?
Becuase this is at the tail-end of the M04 emission spectra so it can be expected that it may detect some low M04 fluorescence.
The 525 nm filter had a much wider bandwidth which meant that the noise would have been larger so it can be expected that these signals were perhaps in that signal but masked by the larger background fluorescence.
Maybe not as the bandwidth is very low so the target fluorescence is minor.
Why do you see a small increase in fluorescence at 1000 um at site 3 in fad?
Maybe white matter and fibers at the border of cortex and hippocampus.
How did you get a good profile in site 1 for this analysis?
For site 1, we for 1/2 good recordings where it appeared that the fibre passed the target site, and resulted in a profile that was reflective of the plaque pathology. However this was not consistent across samples like the other sites were.
Why across examples do the FAD+/- appear to have similar values of dF?
Noise
Why in the FAD- is the correlation coefficient sometimes quire high?
Limitation with this analysis approach.
As the photometry signal is noisy as remains consistent across depth, this is also reflective of the histological noise we add to signals for correlation analysis which may provide some slight false correlation.
Why are there some FAD+ in site 3 with low correlation?
Because of the differences in depth of the subiculuim so subtle differences in the alignment on the contralateral hemisphere can result in some fasle postives/negatives.
Why in this example in site 3, does the fluorescence measure not replicate the bumps in the histological measure?
As we are quantifying on the contralateral hemisphere, it is expected that we will lose some hemisphere specific pathology.
Also, some small changes in pathology may not be detectable.
You have some bimodal distribution in the summary FAD- correlation coefficient, why?
The histology data should remain the same as a flat profile.
However, the photometry could be flat or slightly decreasing with depth due to decreased background noise.
Depending on small variation in these values, a false correlation (+/-) may arise.
If mice are under terminal anaesthesia will this affect the M04 time profile?
Potentially, however urethane allows as much functionality as possible.
But as we did not get any positive data here we cannot directly compare with our later data.
Why were your implants off target?
Optimisation, variation in skull structure across mice, manual error, dificiulty with one depth with FF.
Why was there this decline in fluorescence in APK recordings?
It is thought that because I am off target in an area with minimal plaque pathology, I am bleaching the background autofluorescence.
Or, I am bleaching the FP system due to the high power.
Why did you use 5V for APK experiments, when this was not the case for DP experiments?
This was the case for original exps, but was changed for later. However, no profile was seen in this case either. Also, the highest irradiance was always mid 4V anyway so not much difference.
Why does the damage look much greater in APK than in DP?
the first three may have been completed before I optmised the implantation protocol by completing a durectomy so improve ease of implant. Therfore, due to dimpling of the tissue this may have caused bigger damage trying to insert.
Why was the last implant for APK much more posterior?
In attempt to get a stronger signal I moved the implant slightly posterior as this region has a greater plaque load over a bigger area which mau have helped minimise the implantation error.
What do the papers cited mention on the influence of AF for collecting a true signal?
Formosov quantifies the percentage of signal recorded that is autofluorescence by completing experiments in aie and measuring the signal intensity directly at a photodetector. They found patch cable has high AF as well as brain tissue and the implant to a lesser extent.
Bianco mention that the materials in a implant can contribute to AF. They measure the AF of patch cable and with FF or TF. They find that all contribute to AF with TF highest rate.
Simpson is a review that discusses the importance of AF in FP recordings from various aspects
What do these papers show about risk of AF at 405?
These show emission spectras of lipofuscin and senile plaques showing that this can overlap with our wavelength.
What do chen and colleagues show?
They test the emission spectra of M04 at various excitation wavelengths. They show strong emission at 440 nm, with minimal at 550 nm.
What did you use the 550 nm channel as now?
We did not end up using it for analysis.
As we saw strong signals using the 440 nm channel alone, we have initially focussed on using this data.
What do these other papers show about tissue damage when implanting?
They abalyse the neuroinflammatory effects of implanting various different foriegn recording materials to the brain.
They note increased inflammation surrounding the implat site at 24h which enhanced with time.
Also, they note metabolic changes, neuronal death and differences in nueronal signals,.
Woodroofe showed IL-1 is barely detectable immediately after implantation but over a 24-48 h period a 15-fold increase is seen. In contrast IL-6 levels at day 0 are high, increasing slightly (10%) by day 1 but decreasing to 40% by day 2.
How does anaesthesia affect plaque pathology?
sevoflurane increased activity of APP and resulting amyloid –> perhaps due to energy inhibition or other moleclar channels.
Other inhalation anaesthetics also increase AB: may involve presence of interpeptide hydrophobic cavities in the oligomerized peptides that are anesthetic binding sites where occupancy by anesthetics are expected to stabilize this, increasing their population and leading to an enhanced rate of larger aggregate formation.
Anaesthetic can increase caspases and BACE.
Been shown in vitro, in vivo and in humans.
Another showed a role of change of expression iin RAGE and LRP recpetors controlling the influx and efflux of AB.
How does urethane not indluence neuronal signals? How have they shown this?
it produces less prominent impairment of cortical neuronal activity
What is meant by warping and how can this be avoided? Did other alignment methods have this effect?
When aligning the software will stretch the section in order to match the atlas image. Therefore, if there is some damage of the section where tissue is missing, this can be stretched resulting in tissue in the wrong area.
Some methods do have a way to do manual alignment and I have seen some presented at conferences where they want to have both manual and automated. But at this point, for out project this was the best option.
How can you be sure this signal is actually plaques and not due to other things detected in the brain?
We also have worked on a model that can predict if the signals alone are form a FAD+ or FAD-
This is a work in progress but so far can detect FAD+ with success greater than 80%
Did you try the same dF over all experiments?
Yes,
the baseline adjusted provided the best approach to show the comparison for FAD+/- as it prevented a minimum signal standaridsing FAD- in a great way.
the minimum scoridng proved best to get histology profiles.
