Schedule of Events | Search Abstracts | Invited Symposia | Symposia | Poster Sessions | Data Blitz

What Gets Decoded in Frontoparietal Cortices?

Poster Session F - Tuesday, March 10, 2026, 8:00 – 10:00 am PDT, Fairview/Kitsilano Ballroom

Adem YAZICI¹ (), Gulsum Ozge Sengil, Ausaf A. Farooqui; ¹Bilkent University

Numerous studies have claimed that conscious, task-relevant, attended, working memory (WM) items may be decodable from frontoparietal regions. However, their designs unwittingly had a key confound. The exemplars being decoded (e.g., different WM contents) corresponded to specific and unique trials. Since trials are the task units participants execute, this ends up linking the distinct exemplars being decoded with distinct task or goal identities, leaving it unclear if these studies decoded anything other than these identities. Here, we removed this confound. Instead of the usual design of one exemplar occurring on a single trial and different exemplars occurring on distinct trials, we had participants execute an extended task episode that had them attend to a series of 6 pictures, presented one at a time. Participants were given the goal of detecting up to 3 instances of pictures repeating after two steps. This required participants to encode and keep every presented picture in their WM across two steps. Since goal completion involved going over multiple pictures, individual pictures could no longer be linked to distinct goal identities. We found that despite being task-relevant, attended, and encoded into WM, such pictures could not be decoded from frontoparietal regions and were only decodable from visual regions. Three other multivariate methods confirmed that distinct pictures did not elicit distinct activity patterns in frontoparietal regions. In contrast, goal-related information – the number of repeats detected so far – was readily decodable in frontoparietal regions, suggesting that only higher-level, goal-related entities are represented in them.

Topic Area: EXECUTIVE PROCESSES: Other