ahah no worries! Terminology overlap between subfields is always tricky 🙂

Here “encoding” refers to how stimulus information is represented in spike timing on single trials, not to memory formation or synaptic modification. These are distinct uses of the term

BTSP is an important related topic, but it falls outside the scope of this review, which focuses on spike timing in stimulus coding and behavioral readout rather than on plasticity rules.

Contribution of spike timing to the neural code: from fast to slow timescales - Biological Cybernetics The publication of Mainen and Sejnowski’s 1995 seminal paper strongly renewed interest in how spike timing contributes to the neural code. In the 3 decades since then, considerable experimental and th...

New review out in Biol Cybern!🧠
How does spike timing contribute to the neural code - and at which timescales?
We define and operationalize 6 timescales, from ms encoding precision to behavioral readout.
Thanks to N.M. Engel & S. Panzeri!
link.springer.com/article/10.1...

4/4
Thanks so much to my collaborators
@jwastrachan.bsky.social and Oriana Pansardi for their amazing work on this, to Cristina Becchio and Stefano Panzeri for leading the project, and to all the other collaborators for their valuable contributions! 🙌

3/4
The key point is in the errors: GPT-4o’s mistakes are highly structured and informative, but differ from humans’, especially for inverted faces.
Our results suggest that superhuman mental-state reading does not imply human-like internal computations. 🤖🚫🧠

2/4
We found a striking pattern: GPT-4o outperforms humans for upright faces, but its performance drops much more sharply when faces are inverted.🙃
So strong mindreading-like performance can coexist with a marked vulnerability to visual perturbation.

GPT-4o reads the mind in the eyes Humans possess a sophisticated ability to read the mind in the eyes of other people. Here we tested whether this ability is also present in GPT-4o, a …

1/4 🧵
Can GPT-4o “read the mind in the eyes”? 👀
In our new study, we compared humans and a multimodal LLM on inferring complex mental states from eyes using theory-of-mind tasks developed for humans.
www.sciencedirect.com/science/arti...

Such an amazing week diving into open science and working on cool data reanalysis projects! Met so many interesting people and had great discussions in an incredible setting.
Big thanks to @bendichter.com and all the organizers for putting together this awesome event!