Mark Lescroart
@neuromdl.bsky.social
Computational cognitive neuroscientist at University of Nevada, Reno. Uses fMRI to study visual representations of objects, scenes, and bodies, and how attention affects them. Dad. Nerd since before it was cool. Likes pretty science pictures & puns. He/him
IT’S STONE FRUIT SEASON MOTHERFUCKER
June 25, 2025 at 5:50 AM
IT’S STONE FRUIT SEASON MOTHERFUCKER
Finally, we didn’t stop at pretty pictures - we generated images to elicit strong responses from specific regions in the brain (FFA, EBA, RSC) and showed them to people while we scanned their brains with fMRI. This worked; by and large, images elicited high responses where they were intended to.
June 12, 2025 at 4:34 PM
Finally, we didn’t stop at pretty pictures - we generated images to elicit strong responses from specific regions in the brain (FFA, EBA, RSC) and showed them to people while we scanned their brains with fMRI. This worked; by and large, images elicited high responses where they were intended to.
To do this, we measure the brain with fMRI while people watch movie clips. We then use an encoding model based on a DNN (InceptionV3) and fit to the movie-watching data to predict responses to other movie clips. The accuracy of the predictions tells us we can trust the model to act like the brain.
June 12, 2025 at 4:34 PM
To do this, we measure the brain with fMRI while people watch movie clips. We then use an encoding model based on a DNN (InceptionV3) and fit to the movie-watching data to predict responses to other movie clips. The accuracy of the predictions tells us we can trust the model to act like the brain.
In a new paper led by @matthewshinkle.bsky.social, we use a pre-trained deep neural network (DNN) to model brain activity and to make detailed visualizations of feature selectivity across many brain areas. In plain English: we make pretty pictures of what different parts of the brain like to see.
June 12, 2025 at 4:34 PM
In a new paper led by @matthewshinkle.bsky.social, we use a pre-trained deep neural network (DNN) to model brain activity and to make detailed visualizations of feature selectivity across many brain areas. In plain English: we make pretty pictures of what different parts of the brain like to see.
A good thing happened today: congratulations to Dr Matthew Shinkle @matthewshinkle.bsky.social , the first PhD to graduate from my lab!
May 15, 2025 at 9:10 PM
A good thing happened today: congratulations to Dr Matthew Shinkle @matthewshinkle.bsky.social , the first PhD to graduate from my lab!
Spring in Reno, when you get to play “Blossoms or snow?”
April 3, 2025 at 12:06 AM
Spring in Reno, when you get to play “Blossoms or snow?”
Thinking about this meme a lot as I have realized that nearly EVERY option for word processing now has AI autocomplete of some kind built in. How are we supposed to teach students to write in this environment?
December 13, 2024 at 11:00 PM
Thinking about this meme a lot as I have realized that nearly EVERY option for word processing now has AI autocomplete of some kind built in. How are we supposed to teach students to write in this environment?
eLife’s strong position seems to be that using simple heuristics for judging the impact of science is always bad and should always be avoided. Ok. But how else are we to make sense of a complex world with a tiny human brain? @zachweinersmith.bsky.social
October 25, 2024 at 2:51 AM
eLife’s strong position seems to be that using simple heuristics for judging the impact of science is always bad and should always be avoided. Ok. But how else are we to make sense of a complex world with a tiny human brain? @zachweinersmith.bsky.social
August 22, 2024 at 3:30 AM
Somehow a riff on a Manhattan felt like the right beverage this evening.
May 31, 2024 at 1:06 AM
Somehow a riff on a Manhattan felt like the right beverage this evening.