Science of Emotion and Sensibility

In the automotive aftermarket, user experience (UX) is typically measured using numerical ratings; however, such rating-based assessments rarely capture the emotional tone and contextual dissatisfaction expressed in user reviews. We introduce a mixed analytical framework that combines numerical ratings with text-based sentiment analysis and topic modeling to comprehensively interpret UX in the automotive aftermarket. To examine how numerical ratings, sentiment scores, and latent topics reflect user experience, we analyzed online review data from the four major aftermarket categories. At the aggregate level, sentiment scores were positively correlated with ratings. However, at the category level, this relationship became weak or unstable when rating distributions were highly concentrated. Topic modeling also showed that reviews with similar ratings differed in their experiential contexts (e.g., usability problems, functional inconvenience, login failures, and connectivity issues). Specifically, usability and system instability were rated lower and received more negative sentiment, while satisfaction-oriented topics received higher ratings and positive sentiment. Regression-based analyses showed that sentiment and topic-derived features can complement rating-only evaluation, indicating that automotive aftermarket UX should be interpreted through numerical ratings and the emotional and contextual information captured in review text.

keyword : Automotive Aftermarket, Online Review Analysis, Sentiment-Based Text Analysis, Topic Modeling, Satisfaction Prediction Model, 자동차 애프터마켓, 온라인 리뷰 분석, 감성 기반 텍스트 분석, 토픽 모델링, 만족도 예측 모델

This study explored whether tactile and auditory stimuli share modality-general affective representations. To investigate this, tactile and affective responses to tactile and auditory stimuli were analyzed and compared using a tactile model based on roughness and hardness. Based on previous and newly collected data, a representational similarity analysis (RSA) was conducted on evaluation ratings for tactile (roughness × hardness) and auditory (roughness × hardness) stimuli. The RSA results indicated that tactile and affective ratings were similar across both tactile and auditory stimuli. Notably, roughness played a significant role in both auditory and tactile stimuli, whereas hardness was crucial for tactile stimuli but was relatively less important for auditory stimuli. A mixed ANOVA revealed that tactile responses were more similar for tactile stimuli, whereas affective responses were more similar for auditory stimuli. Additionally, the roughness + hardness model was provided the best fit for tactile stimuli, whereas the roughness-only model was more appropriate for auditory stimuli. These findings suggest the potential existence of modality-general affective experiences and that the perceptual dimensions essential for tactile stimuli may not apply in the same way to the tactile perceptions of auditory stimuli. In other words, while affective experiences can be consistent across modalities, perceptions may differ. This study extends prior research by empirically examining modality-general affective representations across tactile and auditory stimuli, demonstrating the need for further research to explore multisensory affective representations.

keyword : Modality-General Affective Representation, Representational Similarity Analysis, Roughness, Hardness, 감각보편적 정서 표상, 표상 유사성 분석(RSA), 거칠기, 경도

South Korea is facing a serious mental health crisis, with a suicide rate approximately 2.4 times higher than the Organization for Economic Co-operation and Development average, while mental health service utilization remains low at 12.1%. As Large Language Models (LLMs), including ChatGPT, are increasingly filling this gap, concerns have emerged that LLM sycophancy may foster emotional dependency in a psychological counseling context. This study examines the perceived dependency reflected in user discourse rather than directly measuring sycophancy. A total of 1,476 Naver blog posts were collected using Selenium-based web crawling and analyzed via Latent Dirichlet Allocation topic modeling and KoELECTRA-based sentiment analysis. Five major topics were identified, with discussions primarily focusing on emotional support, empathy, and sharing personal concerns. The dependency-related discourse primarily appeared among users who explicitly recognized their reliance on ChatGPT. Moreover, sentiment analysis indicated that positive sentiment accounted for 65.6% of the data, while the “ChatGPT dependency” keyword group exhibited a relatively high proportion of negative sentiment (37.7%). These findings suggest that although sycophantic LLM responses can provide emotional support, they may also contribute to emotional dependency among some users, thereby providing empirical evidence to support the design of safer AI-based psychological counseling services.

keyword : LLM Counseling, Perceived Dependency, User Discourse, LDA Topic Modeling, Sentiment Analysis, LLM 심리상담, 의존 인식, 사용자 담론, LDA 토픽모델링, 감성분석

This study investigated the effects of activities involving interaction with companion plants on the psychological well-being of adults experiencing depressive symptoms. Adults with moderate or severe depressive symptoms (PHQ-9 ≥ 10) were randomly assigned to either an experimental group (n = 17) or a control group (n = 16). Participants in the experimental group cultivated four types of companion plants and participated in structured interaction activities at least once per week for three months, whereas the control group received no intervention. Depression (CES-D), anxiety (GAD-7), stress (PSS), loneliness (UCLA), self-esteem (RSES), and self-efficacy (SES) were assessed at three time points (baseline, mid-intervention, and post-intervention) using a mixed-design repeated-measures ANOVA. Results demonstrated significant Time × Group interaction effects for loneliness and self-efficacy, indicating differential intervention-related effects in the experimental group. Significant main effects of time were observed for depression, anxiety, stress, self-esteem, and self-efficacy. Furthermore, polynomial trend analysis showed that significant linear improvement trajectories for self-esteem and self-efficacy were observed exclusively in the experimental group, suggesting that companion plant interaction specifically enhanced positive psychological resources. These findings suggest that interaction with companion plants may serve as a promising non-pharmacological mental health intervention for adults experiencing depressive symptoms, particularly young adults who have limited access to formal mental health services.

keyword : Mental Health, Agro-Healing, Companion Plant, Depression, 정신건강, 치유농업, 반려식물, 우울

Although generative AI has reshaped human-AI collaboration, empirical research has primarily focused on outcome-based evaluations, neglecting interaction processes. A clear understanding of user-AI interaction mechanisms requires investigating the linguistic speech acts of prompts. Therefore, this study examined how personality traits and task types shape prompt speech act tendencies. Controlled experiments were performed with 40 college students across three tasks. Users’ personalities were clustered using the Big Five Inventory, and prompt logs were classified using GPT, Gemini, and Claude. Binary logistic regression revealed that both personality and task types significantly affected speech act tendencies. Remarkably, conscientiousness and neuroticism emerged as key determinants of these differences. This study provides academic value by shifting the analytical focus to the interaction process using empirical conversation data, serving as a foundational reference for designing personalized human-AI systems.

keyword : Generative AI, Big Five Personality, Task Type, Human-AI Collaboration, Interaction, Prompt, Speech Act, 생성형 AI, 빅 파이브 성향, 과업 유형, 인간-AI 협업, 상호작용, 프롬프트, 화행

Due to the aging global population, there is increasing interest in aging-related changes in hand function and motor control. Nevertheless, conventional research has focused primarily on aggregate metrics such as maximum grip strength. Therefore, this study investigated age-related differences in grip control ability, individual finger contributions and how practice effects from repetitive execution manifest across generations using 60 male participants divided into three age groups (20s, 40s, and 60s). Their grip control function was determined using a multifinger dynamometer. Performance, finger contributions, and their relationships were evaluated using one-way ANOVA and a general linear model (GLM). A linear mixed model (LMM) was used to determine practice effects over repeated trials. Significant age group differences were observed in both tracking accuracy and tracking RMSE, with older age groups showing lower accuracy and higher RMSE. Significant age-related differences were found in finger contributions in the middle and ring fingers, with middle finger contribution being higher in the 40s and 60s, whereas ring finger contribution being higher in the 20s. Over repeated trials, the tracking accuracy increased and RMSE decreased, confirming clear practice effects. These findings shift the focus from simple maximum grip strength to individual finger characteristics in aging, and the provide foundational data for the early assessment of hand function decline and for designing age-specific hand training and rehabilitation programs.

keyword : Age, Grip Control, Finger Contribution, Aging, Motor Practice, 연령, 그립 컨트롤, 손가락 기여도, 노화, 운동 연습효과

This study investigated the associations among work characteristics, safety practices, and accident experiences in delivery workers with and without assigned delivery times. Survey data were collected from 285 delivery workers in South Korea. Participants were divided into assigned-time (n = 146) and nonassigned-time groups (n = 139), based on whether the platform assigned them a specific delivery time. A multigroup path analysis was conducted using the same path model for both groups. Traffic signal violations were positively associated with accident experiences in both groups. In the assigned-time group, average daily travel distances and customer complaints were positively correlated with traffic signal violations. These violations, along with the number of daily deliveries and customer complaints, were also positively associated with experiencing accidents. Comparing individual paths revealed a potential between-group difference in the link from customer complaints to accidents. These findings highlight traffic signal violations as a key safety-related factor across delivery scheduling contexts and emphasize the importance of travel exposure, workload, and customer service demands in the assigned-time setting. Consequently, platforms and delivery companies should consider operating conditions and worker safety when determining delivery times, assigning workloads, and handling complaints.

keyword : Delivery Time Setting, Delivery Workers, Safety Behavior, Traffic Signal Violations, Accident Experiences, Multi-Group Path Analysis, 배달시간 설정, 배달종사자, 안전행동, 신호위반, 사고경험, 다집단 경로분석

This study investigated how students’ use of generative AI differs by design thinking literacy in digital experience planning education. Design thinking literacy is understood not as a level of performance but as an orientation through which learners’ value and connect problem definition with solution development. A self-diagnostic survey was administered to 36 undergraduate students, and principal component analysis and k-means clustering identified two major learner types, user-situation-oriented and visual-narrative-oriented. Students then completed a 5-week digital experience planning project using generative AI, after which their practice diaries and reflections were examined thematically. Both types employed AI to structure problems and develop solutions but they approached AI differently. User-situation-oriented students used AI to deepen users’ understanding and refine solutions from the user’s perspective, whereas visual-narrative-oriented students used AI to verify analysis and implement expressive intentions. In conclusion, generative AI education should address learners’ design thinking orientations and AI role-setting.

keyword : Generative AI, Design Thinking Literacy, Digital Experience Planning, Design Education, Digital Content Design, 생성형 AI, 디자인씽킹 리터러시, 디지털 경험 기획, 디자인 교육, 디지털 콘텐츠 디자인