HELLO THERE!
I'M VIPIN VERMA.

PUBLICATIONS

AFFECT SENSITIVE SERIOUS GAMES AND PLAYER SKILLS

Under review

Developed and evaluated an affect-sensitive serious game called Chem-o-crypt that teaches chemical equation balancing. The game automatically detects the players' affective states of boredom, flow, and frustration using the Affdex SDK from Affectiva. The sensed affective state is then used to adapt the game play in an attempt to engage the player in the game. An experiment comparing the affect-sensitive and non-affective game with the help of Dynamic Bayesian Network revealed that the affective game improved learning for low domain-knowledge players.

PREDICTING REAL-TIME AFFECTIVE STATES BY MODELING FACIAL EMOTIONS CAPTURED DURING EDUCATIONAL VIDEO GAME PLAY

Dec 2020

In an attempt to predict the cognitive-affective states of a player during an educational video game session, this study used a self- emote procedure in which participants' facial expressions and emotions were continuously recorded along with self-reported data about their emotional states. Participants' facial expressions and emotions were captured using Affdex SDK from Affectiva. The captured data were used for binomial logistic regression to predict the cognitive-affective states of flow, frustration, and boredom. The binomial logistic regression un- covered that expressions and emotions could be used to predict these cognitive-affective states of a player. We discuss these predictors and their potential to adapt an educational video game session with non- intrusive and affect-sensitive personalization capabilities. The current study provides a pathway for the educational play design and suggests that it should be non-intrusive while being adaptive to a player's capabilities.

HOW TO PREVENT PLAYERS FROM TEAMING UP IN FREE-FOR-ALL GAMES: THE IMPORTANCE OF INITIAL CONTACT

Nov 2019

It has been reported many times that gamers temporarily team up with their enemies in games which prohibit teaming in pursuit of accomplishing the final objective of the game. This behavior referred to as “teaming” is frequently observed in free-for-all games and spoils the enjoyment of other players. The current study analyzed data collected from an iterated economic game with a payoff structure and player constraints similar to those of the free-for-all games. The result indicated that cooperative behaviors equivalent to the teaming behavior emerged when cooperative intent between participants was successfully signaled and attended in the early stages of the game. Suggestions are made based on these findings which aim to prevent the teaming behavior observed in the free-for-all games.

EVALUATING GAMESCAPES AND SIMAPPS AS EFFECTIVE CLASSROOM TEACHING TOOLS

Oct 2019

GameScapes and SimApps, are new narrative and game mechanic influenced tools developed for improving the classroom learning experience for students. In Fall 2015, GameScapes and SimApps were tested with 16 American colleges and universities and 26 members of their faculties to improve learning outcomes in first-year composition classes, many including English as secondary language students, specifically focusing on writing skills and grammar abilities. Several GameScapes and SimApps developed for the course were tested with over 1000 students. Faculty members could select at least two games from a set of 17 such games that would best align with their courses’ learning objectives. Scores earned in student gameplay were converted to points that applied to students’ grades in the courses. This paper presents a complete evaluation of the GameScapes and SimApps that were used by the teachers and students and discusses best practices, classroom observations, learning outcomes and future scalability of this methodology to other courses.

EMERGING PRACTICES IN GAME-BASED ASSESSMENT

June 2019

I wrote a chapter in the book Game-based Assessment Revisited. The final manuscript is available on Springer.

MOTOR INDICATORS OF COGNITIVE LOAD: A MOUSE-TRACKING APPROACH

Nov 2018

Presented a poster at the 59th Annual Meeting of the Psychonomic Society, New Orleans, LA.
The study involved finding out if the hand motion can be used to make inferences about cognitive load, and to make predictions about the level of cognitive load.
The supervised classifiers, trained using hand motion data, showed prediction accuracy that is about 20% greater than the chance level (33%) and could predict about 10% ~ 20% of the variance of the cognitive load.

USE OF MOUSE-TRACKING METHODS TO MEASURE COGNITIVE LOAD

Sep 2018

Integrated with computerized education platforms, the mouse-tracking technique could provide an inexpensive and less intrusive tool for assessing cognitive load. The present study examined whether mouse-tracking can quantify changes in cognitive load. Participants performed a dual-task, which required them to perform primary tasks of moving a computer mouse cursor along the vertical and the horizontal axes to a target, and secondary arithmetic tasks designed to impose different levels of cognitive load. Analyses of the mouse-tracking data indicated that slower mean response time and less trajectory deviation were observed when participants were given secondary tasks imposing a greater cognitive load, whereas slower mean response time and greater trajectory deviation were observed when participants moved a cursor toward a smaller-sized target. The cause behind the quantitative difference between the cognitive load effect, and the motor task difficulty (target size) is discussed, as are implications of these results for computerized education platforms.

GIVE AND TAKE IN A MOUSE-TRACKING CHOICE PARADIGM

Nov 2017

Do “gender/race – valence” biases interact with approach- avoidance movements: Giving an object or taking it away?

Developed a mouse tracking application deployed on heroku, to gauge the implicit bias of participants.

Results indicated that the approaching trajectory was modulated by the cursor shape and the gender of target face, whereas avoidance trajectory was modulated only by the cursor shape.