How does familiarity change visual object representations ?

Abstract

Familiarity with something implies having a close acquaintance with that item, generally acquired over a period of time. Apart from single items, we have experienced getting familiar with a pair of items (like a mouse and keyboard, iPad and apple pencil, mortar and pestle, etc) based on their co-occurrence in our visual world, with time we develop an expectation of seeing them together. Alongside long-term familiarity, we also develop familiarity with new items as we come across them. How and in what circumstances our behavior gets modified as we get familiar with novel objects or object-pairs, are fundamental questions to ask for unravelling the mystery of object representation in the brain. In a series of three studies, I have sought to uncover the answers to the following questions: 1) How does our behavior change as individual novel objects get familiarized? 2) What is the benefit of learning the co-occurrence statistics of novel object pairs? And 3) how can we apply the knowledge of familiarity in real-world scenarios? In the first study, the goal is to understand the behavioral advantage of recent familiarity. We keep developing familiarity with novel items, like a new painting after buying it, after receiving a new gift, etc. We asked 1) how quickly we develop familiarity with the novel items, and 2) in what situations we see any behavioral benefit for the recently familiarized items. The main finding is, that though familiarization with novel items is quick (within a single session) and long-lasting, the advantage of recently acquired familiarity is not robust, and depends highly on task difficulty. In the second study, we intended to understand the behavioral benefit of learning co-occurrence statistics of novel shape pairs. Earlier studies have shown that participants learn the statistical regularities of novel shape pairs quickly, but whether this learning infers any advantage in an unrelated behavioral task has not been studied before. The main finding of this study is that, with statistical learning participants’ sensitivity to the whole pair, as well as to the individual parts creating the pair increased, which leads to better performance when the associated pair, or individual parts of that pair changes in a change detection task, compared to when a non-cooccurred pairs or parts changes. To explore real-world implications of familiarity, in the third study, we investigate the effect of letter familiarity in the children of the emergent literacy phase. We wanted to understand the source of better discrimination of upright letters/akshara (compared to the inverted ones) in adults. One possibility is that associating letter shapes with their corresponding sounds leads to improved letter discrimination. Alternatively, familiarity with the letter shapes could lead to improved letter shape discrimination. We investigated these two possibilities and found that performance on akshara familiarity but not akshara recognition predicts the improved discrimination of upright akshara during an oddball visual search. Taken together, these above findings uncover how visual representation gets modified with recent familiarity and how familiarity is important for real-world situations like learning to read.