Same Brain, Different Prediction: How Preprocessing Choices Undermine EEG Decoding Reliability
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914543034171392 |
|---|---|
| author | Hou, Dengzhe Wu, Zihao Jiang, Lingyu Li, Zirui Lin, Fangzhou Yamada, Kazunori D. |
| author_facet | Hou, Dengzhe Wu, Zihao Jiang, Lingyu Li, Zirui Lin, Fangzhou Yamada, Kazunori D. |
| contents | Electroencephalography (EEG) is a cornerstone of brain-computer interfaces and clinical neuroscience, yet deep learning models are typically trained and evaluated under a single, unreported preprocessing pipeline. We formalize preprocessing choices as a counterfactual intervention space and show that EEG predictions are surprisingly unstable under this space: across six datasets spanning four paradigms, up to 42% of trial-level predictions flip when only the preprocessing changes, a variability that standard uncertainty methods do not explicitly quantify because they condition on a fixed preprocessing pipeline. We provide three tools to make this instability measurable, decomposable, and reducible. First, a Walsh-Hadamard decomposition of the 2^7 pipeline space reveals that sensitivity is near-additive in practice under the binary intervention design, enabling efficient step-by-step optimization. Second, we introduce Preprocessing Uncertainty (PU), a per-trial diagnostic that captures a dimension of instability complementary to model-based confidence. Third, we study Normalized Adaptive PGI (NA-PGI), a graph-structured regularizer that exploits the compositional structure of preprocessing interventions as one mitigation strategy with clear scope conditions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_07212 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Same Brain, Different Prediction: How Preprocessing Choices Undermine EEG Decoding Reliability Hou, Dengzhe Wu, Zihao Jiang, Lingyu Li, Zirui Lin, Fangzhou Yamada, Kazunori D. Machine Learning Artificial Intelligence Human-Computer Interaction Neural and Evolutionary Computing Signal Processing Electroencephalography (EEG) is a cornerstone of brain-computer interfaces and clinical neuroscience, yet deep learning models are typically trained and evaluated under a single, unreported preprocessing pipeline. We formalize preprocessing choices as a counterfactual intervention space and show that EEG predictions are surprisingly unstable under this space: across six datasets spanning four paradigms, up to 42% of trial-level predictions flip when only the preprocessing changes, a variability that standard uncertainty methods do not explicitly quantify because they condition on a fixed preprocessing pipeline. We provide three tools to make this instability measurable, decomposable, and reducible. First, a Walsh-Hadamard decomposition of the 2^7 pipeline space reveals that sensitivity is near-additive in practice under the binary intervention design, enabling efficient step-by-step optimization. Second, we introduce Preprocessing Uncertainty (PU), a per-trial diagnostic that captures a dimension of instability complementary to model-based confidence. Third, we study Normalized Adaptive PGI (NA-PGI), a graph-structured regularizer that exploits the compositional structure of preprocessing interventions as one mitigation strategy with clear scope conditions. |
| title | Same Brain, Different Prediction: How Preprocessing Choices Undermine EEG Decoding Reliability |
| topic | Machine Learning Artificial Intelligence Human-Computer Interaction Neural and Evolutionary Computing Signal Processing |
| url | https://arxiv.org/abs/2605.07212 |