Choose sensing configurations by target metric and deployment burden, favoring compact five-stream or chest-ECG setups for lightweight deployments¶

Strategy · All strategies
Evidence · 3 claims (3 for) · 2 studies (1 review, 1 design), q2 · 0 of 2 report an effect size · 3 claims rest on one study

Description¶

The article recommends treating sensing configuration as a design decision balancing predictive value against practical burden. It reports that "the preferred configuration depends on the performance metric and the practical cost of its sensors", with a five-stream setup (EEG, EDA, eSense IMU, HR, Ring Temperature) achieving the highest Within-1 and binary accuracy and chest ECG alone achieving the highest binary Macro-F1, so lighter deployments can drop streams without uniform loss.

Design Implications¶

Context¶

Requirements¶

• Screen candidate sensor combinations under the same participant-grouped evaluation protocol so performance changes reflect the simplified sensing setup

Constraints¶

• Differences between independently selected configurations at adjacent subset sizes should not be interpreted as the effect of adding or removing a single stream; performance is not monotonic in stream count

Target Learners¶

• college students in self-guided computer-based video learning

Target Learning Goals¶

• fine-grained estimation of attention difficulty under practical wearable deployment constraints

Claims¶

• Predictive performance of sensor configurations does not change monotonically with the number of sensing streams; preferred configuration depends on target metric and sensing cost [+W]
• Chest-worn ECG provides the strongest cardiac-sensing performance for attention-difficulty prediction, and fusing all cardiac streams does not improve upon chest ECG alone [+W]
• Combining more modalities does not always improve detection of learner mental states [+W]

Related Strategies¶

Examples¶

-

Key Sources¶

• Zikang Leng, Edan Eyal, Yingtian Shi, Jiaman He, Yaqi Liu, and Thomas Plötz. (2026). EduGage: A Multimodal Dataset and Benchmark for Sensor-Based Momentary Assessment of Engagement in Self-Guided Video Learning. Proc. ACM Interact. Mob. Wearable Ubiquitous Ubiquitous Technol. https://arxiv.org/abs/2605.01238