Different effects of internal representations on decisions and eye movements
Poster Presentation 23.453: Saturday, May 16, 2026, 8:30 am – 12:30 pm, Pavilion
Session: Decision Making: Perception 1
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Jozsef Fiser1,2, Saleha Siddiqui3,4, Barnabas Molnar1,2, Rana Banai-Tizkar3,4, Frederic Chavane3,4, Anna Montagnini3,4; 1Central European University, 2Center for Cognitive Computation, 3Institut de Neurosciences de la Timone, 4Aix-Marseille Université
Every conscious and involuntary action we make is strongly influenced by our internal model of the monetary environment and by the actual context. In our previous investigation of perceptual decision-making, we used response biases to identify which alternative internal representation observers used during their decisions. We found that, depending on the representation they adopted, observers could make opposite choices and, in particular, could follow decisions that appear irrational from the standpoint of a “standard” internal representation. In the present study, we explored whether the internal representation that sets the bias of conscious decisions also controls anticipatory eye movements, or conscious and involuntary actions are governed by different internal models. We used a modified version of the standard 2-AFC decision-making paradigm in which observers (N = 20) judged the direction of random-dot kinetograms moving either right or left. First, we identified each observer’s direction-sensitivity threshold with an adaptive staircase method, and then, in the main phase of the experiment, we used randomly chosen coherence levels on each trial. Importantly, when transitioning from threshold measurement to randomized trials, we also changed the appearance probability of the right/left motion directions from 50/50% to 80/20%, thereby imposing a general bias in the probability of motion across trials. Confirming our earlier results from a shape-classification task, observers’ decisions under these conditions showed a strong bias toward selecting the less frequent direction during low-coherence trials. This decision bias is long-lasting (over hundreds of trials) and contradicts the common expectation that observers should choose the more frequent direction. In contrast, measurements of anticipatory eye movements showed a significantly stronger bias toward the more frequent direction. Thus, the internal representations governing conscious decisions and the mechanisms of eye-movement control are at least partially dissociable.
Acknowledgements: This work was supported by Grant ANR-FWF I 6793-B