The Theory of Visual Attention (TVA) is a computational model widely used to describe the various cognitive factors that contribute to the attentional selection of objects in the visual field. In TVA these factors are reflected in mathematical parameters that can be extracted by fitting empirical data into a set of equations.
Traditional works with the TVA model emphasised the influence of the observer’s goals in biasing sensory signals. More recent developments in TVA have also considered the influence of the physical salience of objects in the visual field on the parameters governing attentional selection. As a result, TVA can now account for two main “drivers” of attentional guidance: goal-driven and stimulus-driven. However, many theoretical advancements suggest that attentional guidance cannot be fully understood without considering other factors related to learning and experience. For example, attention can be biased by our recent and long-term experiences with object locations, identities, reward value, and timing, regardless of our current goals or the physical properties of the stimuli.
In my talk, I will explore what we know about the experience-driven modulation of TVA parameters. I will present my work and discuss the work of others on how temporal expectations modulate visual processing speed. I will also present evidence on how spatial weights can shift in response to changes in the spatial probabilities of goal-relevant information. I will propose that future developments in TVA should consider incorporating the independent contributions of memory-related biases as distinct parameters aligning with contemporary perspectives on visual attention.