Research

Confidence and Metacognition

Visual processes help us extract information from the environment, but, at a higher level of processing, how do we evaluate the information from our own visual system? In particular, visual signals are not always reliable and often involve uncertainty or "noise". How do we evaluate the (un)certainty of visual information, especially in the context of making confidence judgments about performing a visual task? We focus on studying the metacognitive processes that support perceptual confidence.

Global vs local information processing

Visual information is processed in a multi-stage, hierarchical system. Early stages focus on extracting local, primitive features of an image (e.g., dots and lines). Later stages are responsible for processing complex stimuli (e.g., face and overall motion direction) and forming a global picture. However, it remains unclear how our subjective experiences of seeing, especially the processes underlying learning and metacognition, arise from this hierarchical system. Under this direction, we study global vs local processing in the hierarchical vision system. We are also attempting to relate our findings to differences across individuals (e.g., autistic traits).

Adaptation and Learning

At every moment, our visual perception depends on what we have seen in the past. In shorter terms, viewing a stimulus for only a few seconds can alter our perception. In a longer time frame, our performance in visual tasks can improve over days of training. How do these adaptation and learning experiences lead to changes in our subjective experience of seeing?

Our approach

To address the above questions, we take an interdisciplinary approach. We conduct psychophysical experiments and explain the results using computational models. In collaboration with other research teams, we also use neuroimaging methods in our studies.