Tutorials
Tutorials at the IEEE Conference on Games (CoG) bring researchers and practitioners together to explore emerging methods, tools, and ideas in game AI and interactive entertainment. Covering topics such as game-playing AI, procedural content generation, player modeling, and generative systems, they provide hands-on learning opportunities and expert insights while fostering knowledge exchange across academia and industry.
- From Zero to Platformer Game with Godot
- Instrumenting Player Experience:
The Body as Interface for Physiological Game Analytics and Biofeedback - Neurogame Development Tutorial (Unity/Windows/Muse)
- Statistical Analysis for Game AI: To the p-value and beyond
From Zero to Platformer Game with Godot
Organizer
Dr. Carmine T. Guida, Pace University - Seidenberg School of Computer Science and Information Systems
cguida@pace.edu
Description
This is a great hands-on tutorial for first-timers / those who have zero experience with video game development. Godot is a free open-source game engine for creating 2D and 3D games. We will start from downloading Godot (no complicated install, no account signup needed at all) as well as some free art, sound effects and music.
We will create a simple platformer level and bring in our player character. There will be follow-along coding in GDScript (it is like Python) to make our player move and jump. We will also add in a camera to follow the player, as well as some sound effects and background music. Time permitting, we can add some fun extras to our game such as coins to collect, simple enemies/obstacles to avoid, and transitioning to another scene. All you need to do is bring your laptop with you. Mac, Windows, and Linux are all good.
Tutorial Structure
The tutorial will begin with a brief introduction (slides) and then proceed to hands-on. The timeline estimate is as follows:
- Slides: 5 minutes
- Download Godot, other assets, and create a new project: 5 minutes
- Make Level 1 of our game: 20 minutes
- Set up our player (including scripting): 30 minutes
- Add camera/scrolling: 5 minutes
- Sound effects and music: 10 minutes
- Special feature (coins/basic enemy/scene change): 15 minutes
- Total: 90 minutes
Instrumenting Player Experience: The Body as Interface for Physiological Game Analytics and Biofeedback
This tutorial introduces practical methods to instrument player experience using physiological signals as both (1) an analytics layer for game user experience (UX) and player modeling and (2) a design substrate for biofeedback mechanics. Participants will learn how to acquire and interpret physiological time series (e.g., heart rate/HRV, electrodermal activity, respiration; optionally EMG/EEG), align them with gameplay events, and translate them into actionable insights and real-time feedback loops. We cover an end-to-end workflow: sensor selection and study setup, synchronization with game telemetry, segmentation and baseline normalization, feature extraction, inference strategies (descriptive analytics), and biofeedback design patterns (closed-loop feedback and intelligent adaptation). The tutorial highlights common pitfalls, motion artifacts, individual differences, over-interpretation, and biofeedback placebo, and provides a responsible practice checklist for consent, privacy, and safe biofeedback game design.
Description
Physiological signals offer a high-resolution view of player experience that complements self-report, performance, and in-game telemetry. Beyond passive measurement, physiology can also become an interactive channel: games can reflect the player's state back to them and support skills such as self-regulation, attention control, or stress management. However, physiological data are noisy, highly individual, and easy to misinterpret if used as direct emotion detectors, and biofeedback mechanics can unintentionally reduce player agency or introduce confounds if not designed carefully. This tutorial provides a practical, game-focused, evidence-based foundation for physiological game and simulation analytics and biofeedback game design, positioning the body as an interface that both reacts to and shapes gameplay.
The tutorial is organized as an end-to-end workflow that participants can apply to entertainment games, serious games, and extended realities. We begin with a pragmatic overview of sensing options commonly used in game contexts (HR/HRV via PPG/ECG, EDA, respiration; optionally EMG/EEG), including trade-offs in accuracy, comfort, cost, and motion robustness. We then cover the core analytics pipeline: synchronization with game telemetry, event-based segmentation (e.g., challenge spikes, decision points, failures, transitions), baseline collection and normalization strategies, artifact detection/handling, and feature extraction suitable for interactive contexts. Participants will learn what physiological features are reliable enough to support design iteration and evaluation, and how to communicate uncertainty.
From analytics, we transition to biofeedback game design: how to turn physiological metrics into mechanics, feedback, and pacing without breaking immersion or undermining player autonomy. We present concrete design patterns and examples for closed-loop systems (direct feedback vs. mediated feedback; continuous vs. discrete; explicit training modes vs. "stealth" biofeedback), mapping strategies (thresholds, adaptive targets, personalization), and reinforcement schedules. We also discuss how to validate whether a biofeedback loop is working (control conditions, manipulation checks, disentangling arousal from physical exertion, and avoiding overclaiming causality).
Throughout, we emphasize the practical questions that matter to the CoG community: Which gameplay moments reliably elicit physiological change? How do we connect signals to mechanics and UX pain points? When should physiology be used for measurement, when for adaptation, and when for player-facing biofeedback? We conclude with a responsible practice checklist covering consent UX, privacy, accessibility considerations, and safety, particularly important when physiology is surfaced to players in real time.
Learning Objectives / Outcomes
By the end of the tutorial, participants will be able to:
- Select appropriate physiological sensors and plan a game study integrating physiology with gameplay telemetry.
- Synchronize signals with game events and perform event-based segmentation and baseline normalization.
- Extract interpretable features (time/frequency-domain metrics; event-related measures) and connect them to gameplay constructs.
- Identify common artifacts and confounds (motion, exertion, missing data, individual differences) and avoid over-interpretation.
- Choose analysis approaches aligned to goals: design iteration, controlled experiments, or player-state inference (descriptive to mixed models to lightweight ML).
- Design biofeedback mechanics using proven patterns (closed-loop mapping, personalization, pacing, reinforcement, transparency/agency).
- Evaluate biofeedback efficacy using appropriate controls and reporting practices.
- Apply responsible practices for consent, privacy, and safe player-facing feedback.
High-Level Agenda (120 minutes)
- Why physiology in games? What it can/cannot infer; CoG-relevant use cases (15 min)
- Sensing primer: HR/HRV, EDA, respiration (and optional EMG/EEG); trade-offs and setup (15 min)
- Pipeline essentials: telemetry logging, synchronization, sampling, timestamps (15 min)
- Physiological game analytics: segmentation, baselines, artifacts, feature extraction, interpretation (20 min)
- Biofeedback game design patterns: mapping strategies, feedback modalities, agency, reinforcement, pacing (20 min)
- Demos: games and interactive systems using biofeedback as main mechanism (20 min)
- Evaluation and responsible practice: controls, confounds, reporting limits, consent/privacy + Q&A (15 min)
Target Audience and Prerequisites
- Researchers and practitioners in game AI, player modeling, game UX/HCI, affective computing, serious games, and XR.
- No prior physiology background required. Familiarity with basic game telemetry and user studies is helpful but not required.
Organizers
- Dr. John E. Muñoz, PhD - Assistant Professor, User Experience Design, Wilfrid Laurier University (Canada); Director, BioAdaptive Research Lab / BioAdaptive Interface Lab. Email: jmunoz@wlu.ca
- Dr. Ifingenia Mavridou - Assistant Professor, Tilburg University, School of Humanities and Digital Sciences, The Netherlands. Email: I.Mavridou@tilburguniversity.edu
- Diego Saldivar - Tech Game Designer, Neurogames, Spain. Email: theneurogamedev@gmail.com
Neurogame Development Tutorial (Unity/Windows/Muse)
Organizer
Diego Saldivar, Busy Eye Games
Description
Neurogames are videogames that use neurotechnology as part of their game mechanics. This tutorial will demonstrate how to make a neurogame for Windows using Unity and the Interaxon SDK. The main focus is to translate mental states to in-game interactions using tools that lower barriers of entry for classical game developers.
This tutorial includes:
A Quick History of Neurogames
What has actually been done in the field of neurogame development in the 21st century and what the game development community can learn from it.
Neurology for Dummies
A crash course in the very basics of neurology. Just the bare minimum to understand the neurotechnology used in this tutorial.
How to Use a Ready-Made Template
This tutorial will demonstrate the use of a neurogame template supporting Unity and Interaxon headsets to build Windows neurogames. The template was designed to be plug-and-play. Its features and toolsets will be showcased so that attendees can further explore the template during the workshop time.
Workshop
Attendees will be given about an hour to explore the template on their own computers. A limited number of Muse headsets will be available for testing. Attendees may ask questions to further understand the use of the technology. A couple finished neurogames will be at hand to demonstrate possible use cases.
Prerequisites
Attendees should already have an intermediate to advanced knowledge of Unity and C#. Attendees must bring their own Windows laptop with Unity LTS and Visual Studio pre-installed. Due to logistical constraints, Mac/iOS will not be supported in this tutorial. No previous neurology knowledge is necessary. A brief crash course in neurotechnology will be given. There will be a limited number of Muse EEG headsets in the workshop, so it is recommended to bring your own if you can. Attendees may download the Unity project which will be demonstrated beforehand.
Statistical Analysis for Game AI: To the p-value and beyond
Organizer
James Goodman, Queen Mary University of London
Description
In many computational intelligence and games venues, such as CoG and ToG, the analysis of results is often a secondary thought. This tutorial stems from a recurring observation in the peer-review process: while the AI agents are cutting-edge, the statistical evidence supporting their superiority is often, er, sparse. This may partially be a byproduct of classic Computer Science curricula, which emphasize algorithmic over experimental design.
This tutorial attempts to construct a rope-bridge over that gap, providing a non-comprehensive toolkit of basic to intermediate techniques for the statistical analysis of results. While we focus on Game AI, some of the principles should be more universally applicable.
What to Expect
- A lightning tour of useful parametric and non-parametric methods.
- What are error bars?
- Bootstrapping as a robust alternative for estimating confidence intervals and effect sizes, especially when your data does not fit neatly into traditional theoretical distributions.
- How to avoid accidental (or intentional) p-hacking. Learn how to properly use Bonferroni or Holm-Bonferroni adjustments when performing multiple comparisons to ensure your significant result is not just a statistical fluke.
- The Wall of Shame: we will review anonymized examples of common errors found in literature to help you identify what not to do in your next submission, or what constructive feedback you can provide in your next paper review.
- Interactive Knowledge Exchange: the session concludes with a collaborative segment for participants to share experimental hurdles and successful analysis pipelines used in their own research.
Prerequisites
No advanced background in statistics is required. This tutorial is designed especially for PhD students and practitioners who want to move beyond simple mean-vs-mean or win-rate comparisons, and add a grain or two of statistical rigour.