École d'été | 9 juin 2026 : Affective brain computer interfaces par Dongrui Wu

Value of the Information & Strength of the Argument

The value of the information is high: the tutorial synthesizes a large body of research and provides a clear, structured overview of the field. The speaker’s argumentation is solid, grounded in established literature and his own research. He critically discusses methodological pitfalls, such as the block design pitfall, which adds depth and credibility. The presentation is well-organized, moving from basic concepts to advanced topics, and includes practical considerations for data acquisition and processing. The speaker also highlights open challenges, such as the need for faster calibration and the limitations of current emotion elicitation methods. Overall, the argumentation is coherent and evidence-based, making it a valuable resource for both newcomers and experienced researchers.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is high: the talk is based on a peer-reviewed tutorial published in the Proceedings of the IEEE, a reputable journal. The speaker cites key historical papers and datasets (e.g., DEAP, SEED) and discusses methodological issues with reference to recent literature. The sources are credible and relevant. The title accurately reflects the content, and the talk stays on topic throughout. The presentation is well-structured, and the speaker acknowledges limitations and open questions, which enhances its scientific credibility. The only minor weakness is that some concepts are simplified for the sake of brevity, but this is appropriate for a tutorial format.

The title accurately reflects the content: a tutorial on affective brain-computer interfaces, focusing on EEG-based emotion recognition and regulation.

8/10

The talk is based on a peer-reviewed tutorial published in the Proceedings of the IEEE (2023), and the speaker is a recognized expert in the field. The content is well-structured, covers fundamental concepts and recent advances, and includes critical discussions on methodological pitfalls (e.g., block design pitfall). However, the talk is a tutorial and does not present original experimental data, and some claims (e.g., specific band-power associations) are simplified.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction to affective computing and its history, citing Picard and Minsky.
• 05:00Explanation of emotion representation in 2D and 3D spaces, and the six basic emotions.
• 10:00Overview of BCI types: non-invasive, invasive, and partially invasive, with their pros and cons.
• 15:00Discussion on EEG signal acquisition, including electrode types and frequency bands.
• 20:00Introduction to public affective BCI datasets, focusing on DEAP and SEED.
• 25:00Signal processing steps: filtering, re-referencing, artifact removal, resampling, and epoching.
• 30:00Feature extraction techniques, including differential entropy and connectivity features.
• 35:00Machine learning for emotion recognition, including within-subject and cross-subject scenarios.
• 40:00Explanation of the block design pitfall and its implications for EEG classification.
• 45:00Transfer learning and deep transfer learning approaches for cross-subject emotion recognition.
• 50:00Multimodal and cross-modal learning, and emotion regulation via brain stimulation.

• Affective Brain-Computer Interfaces: A Tutorial — The talk is based on this tutorial published in the Proceedings of the IEEE, 2023.
• DEAP: A Database for Emotion Analysis using Physiological Signals — Mentioned as one of the most popular affective BCI datasets.
• SEED: SJTU Emotion EEG Dataset — Mentioned as a family of datasets for emotion recognition.

• Affective Computing and Sentiment Analysis — General reference for affective computing methods and applications.
• EEG-Based Emotion Recognition: A Comprehensive Review — A review article that aligns with the tutorial's content on EEG-based emotion recognition.

• A Critical Review of EEG-Based Emotion Recognition — This review questions the reliability of EEG-based emotion recognition, highlighting issues such as individual differences and lack of standardized protocols, which contrasts with the tutorial's more optimistic view.