The representation and recognition of icons play a crucial role in interface interaction efficiency and user experience within human–computer interaction. However, the intricate relationship between product icon types, feature types, and abstraction in cognitive contexts has yet to be clarified. This study aimed to delve into the cognitive mechanisms concerning practical and hedonic product icons across varying abstraction levels using EEG analysis. Moreover, it investigated how the explicitness and implicitness of these icons and their abstraction levels influence recognition efficiency via eye-tracking studies. In Experiment A, a high abstraction level led to prolonged reaction times (RTs), reduced accuracy rate (ACC), more negative N400, and decreased late positive potential (LPP) amplitudes. These outcomes suggested increased effort, heightened semantic conflict, and negative emotional engagement. Moreover, at middle abstraction levels, RTs were consistently longer for hedonic product icons compared to practical ones. Additionally, both N400 and LPP amplitudes were notably larger for hedonic product icons. In Experiment B, eye-tracking results revealed that compared with implicit features, the change of abstraction degree of explicit features is more likely to increase the number of fixation points and RTs of users. Specifically, functional explicit features wielded the greatest impact on overall icon cognitive efficiency. Synthesizing both experiments revealed significant differences in abstraction requirements for practical vs. hedonic product icons. A practical implication arises: moderate overall abstraction suits practical product icons, while higher abstraction suits hedonic ones. Among features, it’s essential to avoid excessive abstraction in functional and symbolic aspects, while simplifying or removing area line features and chamfered elements can be beneficial. This study contributes to Internet of Things (IoT) icon research, offering potential guidance for graphic designers in crafting user-friendly icons.
Read full abstract