New preprint led by Ariel Goldstein provides evidence that IFG (Broca’s area) encodes language using a continuous vectorial representation ("brain embeddings") with strikingly similar geometry to contextual embeddings derived from deep language models: doi.org/10.1101/2022.0…

New manuscript led by Ariel Goldstein and Eric Ham uncovers a surprising similarity between the sequence of internal transformations across layers in deep language models (DLMs) and the sequence of internal processing across time in human language areas: doi.org/10.1101/2022.0…

Happy to announce a new preprint led by Ariel Goldstein: "Deep speech-to-text models capture the neural basis of spontaneous speech in everyday conversations" doi.org/10.1101/2023.0…

I’m thrilled to share the results from my (first) research paper! Grateful to work on an interesting problem, a unique dataset, and with great collaborates. Check it out:

I'm presenting this work as a poster (# 2065) OHBM #OHBM2023 Tue and Wed 12:45-2:45. Come check it out!

During conversation, listener and speaker mirror each other's #brain activity based on the words and their context. Neuron Learn more: cell.com/neuron/fulltex… Zaid Zada Sam Nastase

Happy to see this work led by Zaid Zada now published in Neuron! We use LLM embeddings to capture word-by-word linguistic content transmitted from the speaker's brain to the listener's brain in real-time, face-to-face conversations: cell.com/neuron/fulltex…

ChatGPT and similar #LLMs may process language like people. New work led by Zaid Zada in the lab of @Uri_Hasson found that brain activity aligns with a shared neural code resembling the artificial neural code of #LLMs, as Zada writes in The Conversation U.S.. theconversation.com/ais-encode-lan…

Are you a grad student who wants to give a talk at Princeton, expenses paid? Nominate yourself or a colleague here! forms.gle/WN2ybYMuZiW3re… Priority given to non-ivy schools and students from diverse backgrounds.

Many thanks to Greta Tuckute and Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 for writing a preview of our paper! Check it out for a concise summary and highlights. They also bring up some interesting points (e.g. linguistic context in conversations) relevant to the field.

Really excited to share our new preprint led by Ahmad Samara @ahmadsamara.bsky.social with Zaid Zada, Tammy Vanderwal, and Hasson Lab titled "Cortical language areas are coupled via a soft hierarchy of model-based linguistic features" doi.org/10.1101/2025.0…

How do different languages converge on a shared neural substrate for conceptual meaning? We’re excited to share our latest preprint that specifically addresses this question, led by Zaid Zada

We’re excited to share our unique ECoG dataset for natural language comprehension. The paper is now on Scientific Data (nature.com/articles/s4159…) and data is on OpenNeuro (openneuro.org/datasets/ds005…). Zaid Zada Sam Nastase