fcollman (@fcollman) 's Twitter Profile
fcollman

@fcollman

Neuroscientist @AllenInstitute studying synaptic and artificial networks

ID: 15063795

calendar_today09-06-2008 18:33:17

297 Tweet

449 Followers

212 Following

Andy Shih (@andyshih_lab) 's Twitter Profile Photo

We highlight brain #vasculature in large-scale 3D electron microscopy studies. There is a gold mine of information waiting to be explored, with relevance to the BBB, blood flow, brain clearance, and metabolism. Amazing resources for the vascular community. biorxiv.org/content/10.110…

Nicholas Turner (@nicholasturner0) 's Twitter Profile Photo

Our paper online today in Cell (authors.elsevier.com/c/1eeR~L7PXinZ5) combines EM neural circuit reconstruction, organelle reconstruction, and neural activity in mouse visual cortex released here microns-explorer.org/phase1. Princeton Neuroscience Institute @BCMHouston Allen Institute IARPA

Our paper online today in <a href="/CellCellPress/">Cell</a> (authors.elsevier.com/c/1eeR~L7PXinZ5) combines EM neural circuit reconstruction, organelle reconstruction, and neural activity in mouse visual cortex released here microns-explorer.org/phase1. <a href="/PrincetonNeuro/">Princeton Neuroscience Institute</a> @BCMHouston <a href="/AllenInstitute/">Allen Institute</a> <a href="/IARPAnews/">IARPA</a>
Allen Institute (@alleninstitute) 's Twitter Profile Photo

🔬 ⚡️ Using beams of electrons — our #electronmicroscopy team is capturing the highest possible resolution images of neurons to map the brain. Peek into their lab with this day-in-the-life #brainweek feature. 👀 alleninstitute.org/what-we-do/bra… #neuroscience

fcollman (@fcollman) 's Twitter Profile Photo

My first act of in person lab science in years... any time you can pour liquid nitrogen into something you gotta do it.

Sven Dorkenwald (@sdorkenw) 's Twitter Profile Photo

The results of a fun connectomics project with Viren Jain, @[email protected] and collegues are finally out. Starting from an automated segmentation and EM, self-supervised learning can produce informative embeddings that make downstream tasks more attainable / possible.

Amit Zeisel (@amitzeisel) 's Twitter Profile Photo

I'm happy to share a somewhat unpredicted first manuscript out of my lab with you. Browsing the invaluable Allen Gannett Mouse Brain Connectivity Project's images we wondered, "Hey, anyone noticed these black dots in the red channel?"

I'm happy to share a somewhat unpredicted first manuscript out of my lab with you. Browsing the invaluable <a href="/Allen/">Allen Gannett</a> Mouse Brain Connectivity Project's images we wondered, "Hey, anyone noticed these black dots in the red channel?"
Viren Jain (@stardazed0) 's Twitter Profile Photo

Excited to bring TensorStore to a wider audience after several years of iterative refinement. TensorStore was developed by the Connectomics team at Google and is now used for other applications such as managing checkpoints of huge neural networks (e.g. Google's PALM).

fcollman (@fcollman) 's Twitter Profile Photo

We should replace the Turing test with the “Turing time”, or how long can an interaction be and not be distinguishable from human. Clearly the Turing time has passed a minute at this point, but maybe not 24 hours?

Viren Jain (@stardazed0) 's Twitter Profile Photo

The idea behind SegCLR is simple: embeddings are learned using a contrastive loss where patches from nearby locations on the same cell should yield similar representations, whereas patches chosen from different cells should yield dissimilar representations. (1/5)

The idea behind SegCLR is simple: embeddings are learned using a contrastive loss where patches from nearby locations on the same cell should yield similar representations, whereas patches chosen from different cells should yield dissimilar representations. (1/5)
Sven Dorkenwald (@sdorkenw) 's Twitter Profile Photo

Our analysis of the connections between excitatory L2/3 pyramidal cells in mouse V1 just came out in eLife! Amazing team work between Princeton Neuroscience Institute BCMHouston Allen Institute Previous tweet thread: x.com/tmacrina/statu… elifesciences.org/articles/76120

Zhihao Zheng (@zhihaozheng) 's Twitter Profile Photo

Excited to share beam deflection Transmission Electron Microscope (bdTEM), a fun and productive collaboration with Project Voxa. biorxiv.org/content/10.110… (1/5)

Tyler Sloan (@quorumetrix) 's Twitter Profile Photo

In reality, neurons are packed in so tightly that they make sardines look comfortable. Here are 364 neurons packed into a box roughly the size of a grain of sand. This dense reconstruction is from the Allen Institute's MICrONS layer 2/3 dataset. Rendered in #Blender3d

In reality, neurons are packed in so tightly that they make sardines look comfortable. Here are 364 neurons packed into a box roughly the size of a grain of sand. This dense reconstruction is from the Allen Institute's MICrONS layer 2/3 dataset. Rendered in #Blender3d
fcollman (@fcollman) 's Twitter Profile Photo

Happy holidays from the brain! This is an awesome remix of the original I made. I love seeing how open science and data also can lead to open art and inspiration. Thanks to the broader team from Allen Institute Sebastian Seung and Andreas Tolias Lab @ Stanford University to make things like this happen.

Tyler Sloan (@quorumetrix) 's Twitter Profile Photo

Happy to share my latest #neuro animation, a remix of previous #MICrONS videos, while I took some #SciArt license and had some fun, it remains quantitative. Possibly inspired by me having recently ridden Smuggler’s Run. A video🧵 1/4 #neuroscience #b3d

Allen Institute (@alleninstitute) 's Twitter Profile Photo

Our scientists are using millimeter-scale #volumeEM to map the brain w/ single-synapse resolution. In a recent preprint, using data spanning millions of synapses across more than 1300 neurons, they reveal new organizing principles for cortical inhibition. biorxiv.org/content/10.110…

Our scientists are using millimeter-scale #volumeEM to map the brain w/ single-synapse resolution. In a recent preprint, using data spanning millions of synapses across more than 1300 neurons, they reveal new organizing principles for cortical inhibition. 
biorxiv.org/content/10.110…
fcollman (@fcollman) 's Twitter Profile Photo

We posted our pre-print today describing how the Connectome Annotation Versioning Engine (CCAVE) supports analyzing and editing the worlds largest connectomics datasets biorxiv.org/cgi/content/sh…! Sven Dorkenwald @csdashm Allen Institute Sebastian Seung!

fcollman (@fcollman) 's Twitter Profile Photo

Excited to get started working on this! Will also aim to do 10 mm^3 at synaptic resolution across the basal ganglia, including NaC, VP, MD thalamus and VTA and the connections between, with segmentation from Zetta AI.