Yesterday at #ICON2025 I got to present our poster with #EttoreAmbrosini, @palencianoap.bsky.social, and @mruz.bsky.social on flexible task representations. I loved the chance to share our preliminary results and hear such thoughtful feedback!

[5/6] RSA revealed that activity patterns coding task demands generally followed a parallel geometry, supporting the transfer of information across contexts. At target onset, there was a shift toward an orthogonal configuration, likely minimizing interference between stimulus and response factors.

[4/6] Instruction content was encoded in distributed EEG signals, with overall task demands displaying more stable patterns. Further, neural codes exhibited both high abstraction, generalizing across conditions,and high and variable dimensionality, suggesting flexible expansion-compression dynamics.

[3/6] Participants followed instructions that involved a novel re-combination of components. These were, the higher-level task demand of integrating or segregating information, the stimulus category, and the relevant feature to attend.