Stephan Pohl @stephanpohl.bsky.social · 03/07/2026Great to work on this with: @eywalker.bsky.social @dlbarack.bsky.social @racheldenison.bsky.social @neddo.bsky.social @weijima.bsky.social #neuroskyence #neuroscience #philsky 020
Stephan Pohl @stephanpohl.bsky.social · 03/07/2026Our reply: the framework systematizes empirical practice; it's not an account of the nature of representation. Whether a neural response represents a feature is settled by comparing hypotheses, not an absolute evidence threshold. And this holds independently of questions about consciousness. (3/3) 111
Stephan Pohl @stephanpohl.bsky.social · 03/07/2026Their worry: our framework lacks a lower boundary (when are sensitivity, specificity, invariance, and functionality sufficiently satisfied for a neural response to be a representation?) and an upper boundary (when is a representation conscious?). doi.org/10.1038/s415... (2/3)doi.orgBoundary issues for multidimensional frameworks of representation - Nature Reviews NeuroscienceNature Reviews Neuroscience - Boundary issues for multidimensional frameworks of representation 100
Stephan Pohl @stephanpohl.bsky.social · 03/07/2026Grateful for the chance to clarify our framework: our reply to Overgaard & Kirkeby-Hinrup's "Boundary issues for multidimensional frameworks of representation" — a comment on our Perspective on the conceptual dimensions of representation, in Nature Reviews Neuroscience: rdcu.be/frSNr (1/3) 131
Stephan Pohl @stephanpohl.bsky.social · 23/03/2026I am happy to discuss our paper! I do think the response by Krakauer and Ramsey adds crucial context. However, I think they do not sufficiently appreciate what our paper has to say about the computational role of representations beyond the carrying of information (cf. invariance and functionality). 010
Stephan Pohl @stephanpohl.bsky.social · 21/03/2026Fair caution, but it doesn’t quite apply here. We’re not proposing new metrics. We make explicit how existing approaches (decoding, encoding, RSA, etc.) are already addressing the same questions. Our formalism is a conceptual tool to show how these methods relate, not really an alternative to them. 000
Stephan Pohl @stephanpohl.bsky.social · 20/03/2026What does it take to show the brain represents something? We offer a framework that brings conceptual clarity to representation, systematizing how decoding, encoding, RSA, etc. bear on that question. This makes explicit what findings establish and where interpretations go too far. 0246
