adam osth

🧠 New preprint alert!

In this study, using a joint modelling method with the Diffusion Decision Model, we offer a mechanistic reinterpretation of the Late Positive event-related potential Component (LPC) as a neural signature of mnemonic strength during evidence accumulation in recognition memory.

📢 New preprint: "Comparing Cue and Target Roles in Cued Recall Reveals Limits of Intrinsic Word Memorability".

See lnkd.in/gXs48uyx for a tweet thread. Congrats, Anuska Maity, for your first preprint (and hopefully published paper soon)!

We want to be able to say stuff that other people, for institutional reasons, can’t say, or they’re too afraid to say. You may be the only CEO in America making that promise. We’re in a unique position, certainly. And other places have to learn from this — being afraid all the time and just constantly making transactional moves. How long can you survive like this? What is even the point of being alive if you’re just gonna continue managing rot? And it seems like that’s what 98 percent of people are doing in these media companies right now. I don’t know, take a chance. Nobody likes what’s happening. Especially if you’re a journalism company or a media company, you’re [meant to be] actually reflecting what people want or what people believe. So get some guts and do something interesting. It’s not that hard.

I had a REALLY lovely time talking to @milesklee.bsky.social at Rolling Stone about why The Onion is thriving.

It's because we're one of the only legacy outlets unafraid to accurately reflect the mood of this country.

That mood? Everyone fucking hates this.

www.rollingstone.com/culture/cult...

Is computation necessarily independent of its physical implementation? Now there’s a whole symposium on this question in the journal Mind & Language!

Please check out this work by @seymaertekin.bsky.social et al. where we studied key empirical benchmarks of working memory in a large-scale sample of elementary school students. This type of data is AWESOME because you can get a lot of clarity - as long as you don't go down any rabbit holes...

Labyrinth, by snailgun track by snailgun

new single new single new single

snailgunband.bandcamp.com/track/labyri...

Gavin Newsom Sits Down For Podcast With Serial Killer Who Targets Homeless SACRAMENTO, CA—Promising his podcast listeners an engaging and enlightening conversation, California Gov. Gavin Newsom reportedly sat down Tuesday with a serial killer who targets the homeless population. “So what do you think Democrats can learn from somebody who, like you, targets the most vulnerable among us?” said Newsom, who acknowledged that while he and the […]

“So what do you think Democrats can learn from somebody who, like you, targets the most vulnerable among us?” said Newsom
theonion.com/gavin-n...

Memory & Cognition Special Issue - Psychonomic Society D. Stephen Lindsay, University of Victoria, Canada

In case you haven't seen it yet, M&C is doing a special issue for Larry Jacoby. It says 4/30 for a due date, but I believe we're allowing for some more wiggle room. www.psychonomic.org/page/MCjacoby

This is extremely interesting and thorough work! But I’m wondering what kind of retrieval mechanisms could actually lead to a Gumble min model in practice?

Most retrieval models use some kind of summed similarity computation, which naturally leads to approximately Gaussian distributions

Extreme-Value Signal Detection Theory for RecognitionMemory: The Parametric Road Not Taken Signal Detection Theory has long served as a cornerstone of psychological research, particularly in recognition memory. Yet its conventional application hinges almost exclusively on the Gaussian…

Honey, we fixed Signal Detection Theory (SDT)! In this preprint, Constantin Meyer-Grant, David Kellen, Sam Harding, and I critically evaluate the (unequal-variance) Gaussian SDT model in recognition memory and pursue the Gumbel-min model as a principled alternative: doi.org/10.31234/osf...
🧵

Sorry but I just don't feel comfortable sharing public toilets with transphobes. At any moment they could just demand to have a look at my gennies. The law should protect women's spaces from these creeps.

Twitter post/pohoto from JK Rowling who sits on what looks like yacht, taking a puff of what looks like a cigar. She holds a cocktail close to her face and looks smug as fuck. Text on tweet: I love it when a plan comes together. #SupremeCourt #WomensRights.

Text on screen: our.moral.imperative British billionaire JK Rowling has posted several gloating tweets today after the UK Supreme Court ruled that only 'biological women' legally count as women, excluding trans women from legal protections and safeguards. Rowling's vicious smear campaign against trans people in Britain has been going on for several years now and has played a major role in cultivating the extremely hostile and toxic environment that exists in the country towards trans people, particularly in the name of liberal 'feminism'. @fight_for_a_future @rowllingofficial

You all need to treat Harry Potter IP like we’re treating the Tesla brand.

I got my old SNES working on a modern tv. I’ve been sick this week, so I spent most of my free time playing the old school sim city. I got this achievement that I was never able to get as a kid

This is now published in the Journal of Mathematical Psychology!

www.sciencedirect.com/science/arti...

with @annalschubert.bsky.social, @gidon-frischkorn.bsky.social, and Klaus Oberauer

how do you get Siri to adopt a Cajun accent?

Large Language Models lack essential metacognition for reliable medical reasoning - Nature Communications Large Language Models demonstrate expert-level accuracy in medical exams, supporting their potential inclusion in healthcare settings. Here, authors reveal that their metacognitive abilities are under...

Metacognition of medical knowledge in LLMs: “Models consistently failed to recognize their knowledge limitations and provided confident answers even when correct options were absent.”

www.nature.com/articles/s41...

BTW, if you run studies on Prolific, I strongly recommend that you spend some time on the Prolific subreddit to see what participants discuss, complain about, etc.
www.reddit.com/r/ProlificAc/

Alright it's here! Where we push back against the exploratory-confirmatory distinction that's become canonized in metascience and talk about what we think exploratory research (including exploratory experimentation and modeling) means and why our view is incompatible with 'unplanned data analyses'.

in “all star” whenever the singer of smash mouth sings “all that glitters is gold” it sounds like he’s singing through a toilet paper roll

Gordon Logan did some cognitive research on guitar players and managed to get some free guitars from a guitar company. Maybe I need to change my research direction and try something like that…

Today I played a Travis Bean guitar. Never played anything like it before.

Sadly it costs 11 grand :(

Reset Power Eject Snailgun · EP · 2024 · 4 songs

we also have it up on spotify!
open.spotify.com/album/33cyrV...

Reset Power Eject, by snailgun 4 track album

hey friends. I've been playing in a post punk band for a little while with @daniellittle.bsky.social. we released our first record recently which is an EP of four songs. this is my first ever recording so it's been pretty exciting to share it. snailgunband.bandcamp.com/album/reset-...

“NIH produces an astounding return on investment to the American taxpayer. In Fiscal Year 2022, NIH research funding supported 568,585 jobs and generated $96.84 billion in economic activity — that’s $2.64 of economic activity for every $1 of research funding.”
www.researchamerica.org/2023-oped-us...

this game is sooo scary

there's gotta be a way we can blame the election loss on the existence of the Costco Guys

spotted in my neighborhood

“simple man” by lynard skynard is the first grunge song

This is really interesting. How do you model that? I generally find that there aren’t many ways to model superior encoding in memory models

To account for hit rates, we rely on the hybrid similarity framework. In this model, self similarities can be augmented through other types of matches. We were able to accomplish this through distinctiveness ratings. Once these were included, the model did pretty well in accounting for hit rates!