INVESTIGADORES
KAMIENKOWSKI Juan Esteban
congresos y reuniones científicas
Título:
“Brain settings across free-viewing tasks: from Exploration to Visual Search and Hybrid Search”
Autor/es:
KAMIENKOWSKI JE; CARE D; GONZALEZ J; RIES AJ; ISON MJ
Reunión:
Congreso; Visual Science Society meeting; 2023
Resumen:
When we look at a scene in the real world, we typically do it with a goal in mind, for instance searching for our keys or looking for one of many cups to prepare our breakfast. These goals can shape how we go over the scene. Here, we aimed to study how ecologically-relevant variables such as the particular task performed, the progression of the task, and the memory load required by the task influence brain activity. We performed two experiments with co-registered EEG and eye-tracking to investigate task-related effects on fixation-related potentials (FRPs). In the first experiment, participants were asked to observe (EX) some images or actively search for a target among an array of faces and objects (VS) in other. In the second experiment, participants memorized potential targets (of a memory set size -MSS-) and searched for any of them (hybrid search, HS). We applied a deconvolution analysis approach to estimate the contribution of the different elements embedded in the task. In both cases, we obtained robust FRPs for specific fixation content such as object category, consistent with classical fixed-gaze experiments and a handful of free-viewing studies. We also report a robust task effect, both comparing EX vs VS, and MSS=1 vs MSS>1 in the HS task. Moreover, when further investigating the dynamics along the trial, we found significant effects of trial progression and its interaction with the task effect. These effects occur early after fixation onset (100-200ms). These results shed light on top-down influences associated with different goals, not only in the pattern of eye movements but also in the brain response after each fixation. Overall, these studies show how combining experimental and analytical approaches allow us to discern among multiple overlapping neural processes, preserving key attributes of real-world tasks.