Adam Glaser
@adamkglaser.bsky.social
990 followers 210 following 22 posts
Developing hardware + software tools for large-scale imaging at the Allen Institute for Neural Dynamics
Posts Media Videos Starter Packs
adamkglaser.bsky.social
Latest imaging data from our completely re-designed ExA-SPIM system.

Expansion-assisted imaging (3X expansion, 0.3 NA) enables brain-wide imaging with high-contrast and <400 nm effective resolution.
adamkglaser.bsky.social
That is a good data point to know. We hypothesize (no data yet) that aberrations will reduce quickly with expansion factor (RI gradients proportional to density, density gradients will reduce with expansion factor ^3). Whereas the imaging path length only increases linearly with expansion factor.
adamkglaser.bsky.social
ExA-SPIM2 will provide 35 mm WD with 0.5 NA.

I think there are definitely applications where higher NA is needed. If we designed a 3rd variant, I think aiming for ~0.8 - 0.85 NA (which allows for symmetric 90 deg objectives) with 20 mm WD would be very valuable.
adamkglaser.bsky.social
"Expansion-assisted" is all about balancing light collection (NA^2) with practicality of engineering lenses with both high NA and long WD. We are collecting data now to quantify this, but we also believe aberrations in ExM will scale with NA^3.

So IMO go as low in NA as specimen brightness allows!
adamkglaser.bsky.social
None of the above captures the long WD (35 mm). Perhaps we should introduce a volumetric SBP... that divides the usable WD by either the Rayleigh axial resolution, or an isotropic resolution assumption that divides by the lateral resolution -> would be very high for this lens.
adamkglaser.bsky.social
Thanks for resolving the numbers.

Agreed, this was the primary motivation for ExA-SPIM in the first place. More pixel parallelization with a larger SBP, either in 2D (global shutter) or even along 1D for rolling shutter (i.e. for ASLM). Also removes some scanning bottlenecks with a low FPS sensor.
adamkglaser.bsky.social
That is right, I used the equation from doi: 10.1117/1.ap.3.4.044001. The required number of samples to fully resolve the signal - also their equation uses 0.5 instead of 0.61. Lens has more information than any current standard camera sensor but potentially could be fully used in the near future.
adamkglaser.bsky.social
Totally agree with you 😆 space-bandwidth product > etendue (!)

For 420 - 850 nm light the SBP is 175 - 825 megapixels. Around 500 megapixels at 500 nm.
adamkglaser.bsky.social
We are excited to share our development of a new 10X lens with NA=0.50 over a 7.2 mm FOV and a 35 mm working distance (water), fully corrected from 400-850 nm. This lens will power our upcoming ExA-SPIM "2" and is available for dissemination - please get in touch if interested!
Reposted by Adam Glaser
voigtvision.bsky.social
Looking forward to the @mesospim.bsky.social #lightsheet #Microscopy anniversary symposium in October with @preibischs.bsky.social @adamkglaser.bsky.social @christellelangevin.bsky.social & many more! And to the next mesoSPIM decade! 🔬 Thanks @nvladimus.bsky.social & Co for organizing it!
adamkglaser.bsky.social
New SPIM system underway!
adamkglaser.bsky.social
It’s across the board. All grants from all agencies, future ones on pause and current ones also on pause :/
adamkglaser.bsky.social
Can maybe estimate based on the 35 mm full frame sensor size (36x24 mm). So probably around 1.46 micron pixels? The max rate is 3280 megapixels/sec 🤯