Sensory and semantic interference during working memory in the human brain

• Applicant: Dr. Thomas Christophel
• Subject Area: Biological Psychology and Cognitive Neuroscience
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 554623481

Human cognition relies on the retention of information in working memory. Limits in cognitive abilities derive from limitations in this ability to retain information and understanding these limits is the central goal of working memory research. These limitations are believed to be the consequence of interference between concurrent representations in the human brain. Thus, understanding how concurrent representations interfere with each other is critical to understanding limitations in working memory capacity. Here, we aim to understand when, why, and how interference by intermittent stimulation can change a currently held working memory representation. First, to delineate when intermittent stimulation during working memory delays alters mnemonic representations and memory recall, we will use a large-scale online study combining data from hundreds of subjects in a between-subject design to finely demarcate distracting effects of different stimuli under varying presentation conditions. Importantly, we will include verbal, abstract distractors that are meant to interfere with the visual working memory representation not on a level of low-level visual feature maps but by targeting higher-level representations. In this way, we expect to derive a clear set of indicators allowing us to predict when an intermittent stimulus is affecting working memory. Second, we will aim to understand why some cortical memory representations are more affected by intermittent stimulation than others. Using fMRI and decoding methods, we will test two alternative hypotheses: The first alternative is that representations in sensory areas are generally more susceptible to distraction including abstract non-visual distraction, demonstrating that distraction effects are centrally mediated. Reversely, under the second alternative, anterior and posterior regions could be similarly distractor-susceptible but to different kinds of distractors. This would suggest that distraction effects can occur in feature-dependent manner across the cortex, meaning that multiple concurrent representations across multiple regions help shielding working memory from distraction. Third and finally, we ask how cortical memory representations are changing under the influence of intermittent stimulation. We will overcome limitations in prior work which could only observe distractor and memory representations in an intermixed, indivisible fashion using a carefully designed multi-feature paradigm and precise fMRI encoding modelling. This approach will allow us to precisely quantify the alterations of the representational geometry of a remembered item in different cortical regions as an incoming distractor is processed.

• DFG Programme: Research Grants