Pritha Ghosh supervised by Dr. Vishnu Sreekumar received her Master of Science by Research in  Computer Science and Engineering (CSE). Here’s a  summary of her research work on Why do we not perceive jealousy? An ethological perspective of emotion perception & Event segmentation and recognition memory

Certain emotions, such as fear or joy, are recognized more readily than others, such as jealousy or loneliness, across modalities including auditory (speech and vocalization), visual (facial expression and bodily movement/gestures), and music. Despite there being multiple theories of emotions, existing accounts do not adequately explain why some emotions are more easily recognized across modalities than others even when there is no contextual information. To address this gap, the present work proposes an ethological framework for understanding emotion perception. We hypothesized that emotional displays can be broadly categorized as signals or cues. Emotions associated primarily with signal-like displays are expected to be more readily perceived in both auditory and visual modalities than emotions that are not. Building on this distinction, we further propose that these emotions are also more easily perceived in instrumental music. We conducted four surveys and one experiment to test our hypotheses. Indeed, emotions that were more frequently associated with signals were the ones that were not only judged to be more easily perceivable but were indeed more easily perceived in real life. Even though we could not directly test the multimodality hypothesis due to the extremely high correlation obtained in the collected data, based on exploratory analysis, it can be speculated that these emotions are actually better perceived in both modalities. These emotions were also more readily perceived in instrumental music. Overall, the findings support the proposed framework and are consistent with existing theories of emotion and emotion perception in music. The second part of the thesis deals with a different aspect: memory. How do different forms of contextual change segment ongoing experience into discrete events? How do such boundaries influence memory for information encountered around them? While event boundaries are known to affect recognition memory, their impact on fine-grained mnemonic discrimination remains largely unexplored. Prior work [1] reported boundary-related memory effects when multiple sources of contextual change co-occurred, leaving open which type of change is critical. Building on this work, we used a design that separates two sources of contextual change that were previously confounded: shifts in stimulus domain and shifts in task context. We contrasted these by introducing stimulus-domain shifts (image category changes; Experiment 1) and task-context shifts (changes in the encoding judgment; Experiment 2). Following encoding, participants completed recognition and lure discrimination tests. Reliable boundary effects emerged only for task-context shifts, which produced a post-boundary decline in recognition accuracy. Stimulus-domain shifts did not yield comparable effects, and neither manipulation influenced lure discrimination. These findings indicate that shifts in task context alone are sufficient to induce boundary-related memory effects, highlighting a central role for changes in goals and attentional set in event segmentation. 

June 2026