iiinova — where your digital self keeps evolving

ya agree. tbh it was more that i did not think you’d be a classical models guy, and i think the reference class of models is a sneaky but v important part of navigating this space, so the negative prediction error btwn my expectation & the post drove me to seek the relevant info lol

010

ahh gotcha okay i read too much into the initial post, you were saying that a criticism of classical models isn’t solved by ANNs either

110

perhaps more directly: what features of classical models make them appropriately/sufficiently(?) “brain-y” to you? fwiw i am asking bc i like to know where ppl fall on spectrums of takes lol, not bc i want to argue 🥲

100

Dan Mirea @danmirea.bsky.social · 01/10/2026

im really curious what you have in mind as a classical model here, and why you think that classical models in general are the relevant contrast case. i’m generally sympathetic to this line of reasoning, but also see ANNs used in several ways that map to multiple diff construals of “brain like”

100

Reposted by ari ☀️

Thrilled to announce that I will be joining the University of Miami as an Assistant Professor of Psychology next year! I am recruiting a PhD student to start in Fall 2027 — please reach out if you are interested in joining the lab! For more info check out my website: dan-mirea.com

56912

Reposted by ari ☀️

aaron bornstein @aaronbornstein.bsky.social · 29/09/2026

New lab preprint! @khoudary.bsky.social presents a tour-de-force: a new model of how uncertain expectations should help resolve uncertain sensations, supported by re-analysis of four experiments in monkeys (behavior, ephys) and humans (behavior, fMRI, EEG, intracranial) @meganakpeters.bsky.social

23810

Reposted by ari ☀️

Tim Verstynen @tdverstynen.bsky.social · 29/09/2026

Looking for a postdoc in comp neuro and human brain imaging? Interested in foraging? Join our international collaboration with Jon Rubin, @jyotikab.bsky.social, & @davidrobbe.bsky.social be.bsky.social to study the neural basis of foraging behavior! cmu.wd115.myworkdayjobs.com/CMU/job/Pitt...

cmu.wd115.myworkdayjobs.com

Postdoctoral Research Associate – Cognitive Axon Lab – Dietrich College

The Cognitive Axon Lab, in the Department of Psychology and Carnegie Mellon Neuroscience Institute, seeks applications for a multi-year Postdoctoral Research Associate at Carnegie Mellon University. T...

02222

Kevin O’Neill @kevingoneill.github.io · 29/09/2026

thank you!!!! it’s been so fun thinking with you about this for the past several years 🥲

010

Reposted by ari ☀️

big congrats to my mentor @felipedebrigard.bsky.social for being recognized for his incredible work on memory & forgiveness! also props to his student @gabrielafm.bsky.social who is leading this research and to lab manager @kmiceli.bsky.social

2235

additionally, our model predicts that reliance on expectations ebbs & flows throughout a single decision: increasing at the moments in time when sensory evidence is uncertain, and decreasing as sensory uncertainty diminishes 〰️ this makes sense but has been difficult to test in lab---stay tuned!

Exciting Roller Coaster Ride

Alt: roller coaster going down and back up

030

critically, the findings we synthesize *cannot* be explained (or predicted!) by existing normative models. on our model, however, they emerge naturally from the temporally-extended, probabilistic nature of memory retrieval + optimal integration of that signal with incoming evidence from vision

Regular Show: Mordecai - That Is Pretty Cool

Alt: gif of a cartoon bluejay saying "that is pretty cool"

140

we show that our model synthesizes neural & behavioral findings across four (!) experiments w/ diff manipulations of expectations (x&y axes below). also, we show that one of its key latent quantities---a time-varying weight on expectations--matches frontal theta oscillations in humans 🧠📈📉

a quadrant depicting simulated studies. the x axis spans from trials to blocks, and the y axis spans from instruction to learning. in instruction x blocks, we simulate model & single-unit recordings from Hanks et al. (2011). in the blocks x learning quadrant, we simulate theta amplitude in EEG from Rungratsameetaweemana, Itthipuripat et al (2018). in trials x learning quadrant, we simulate model & behavior & fMRI from Bornstein et al. (2023). and in instruction x trials, we simulate frontoparietal activity in iEEG from Hachisuka, Shor, Liu et al. (2026)

140

these assumptions come from models that optimize speed-accuracy tradeoffs, but totally elide the role of memory systems in acquiring and deploying expectations (😱) we present a new model jointly grounded in empirical properties of memory systems & normative principles of inference under uncertainty

text describing our model, which reads: perceptual decisions reflect dynamic reliability-weighted integration of expectations with sensory evidence. expectations are a probabilistic source of evidence sampled in parallel to vision, and each source is weighted by its relative reliability as estimated during sampling.

130

lots of work has investigated how what we _expect_ to see influences how (& what) we decide what we’re seeing. but it has largely rested on two assumptions: (1) observers have no uncertainty about their expectations (2) expectations exert static (i.e., time-constant) effects on decisions

Spike Spiegel's Big Assumption on Netflix's Cowboy Bebop

Alt: man resting on couch saying "that's a big assumption"

130

additionally, our model predicts that reliance on expectations ebbs & flows throughout a single decision: increasing at the moments in time when sensory evidence is uncertain, and decreasing as sensory uncertainty diminishes 〰️ this makes sense, but has been difficult to test in lab---stay tuned!

ALT:

000

gif of an animated bluejay saying “that is pretty cool”

Alt: gif of an animated bluejay saying “that is pretty cool”

100

we show that our model synthesizes neural & behavioral findings across four (!) experiments w/ diff manipulations of expectations (x & y axes below). we also show that one of its key latent quantities---a time-varying weight on expectations--matches frontal theta oscillations in humans 🧠📈📉

100

text & figures describing our model. text: perceptual decisions reflect dynamic reliability-weighted integration of expectations with sensory evidence. expectations are a probabilistic source of evidence sampled in parallel to vision, and each source is weighted by its relative reliability as estimated during sampling.

figure: three timeseries of accumulated evidence. the key quantity driving choice (decision variable) exhibits a nonlinear increase over time, reflecting the dynamics prescribed by our model.

100

many studies have examined how what we _expect_ to see influences how we decide what we’re seeing. but they largely rest on two assumptions: (1) observers have no uncertainty about their expectations (2) expectations exert static (i.e., time-constant) effects on decisions