# MI-014: Optimizing Spatial Filters for Robust EEG Single-Trial Analysis

> Generated from a local paper card. Do not edit this page directly; edit the source card and rerun `vp run docs:generate`.

> Internal wiki boundary: paper-card evidence and user-provided PDF downloads are available for private reading. Extracted full-paper text is not published.

> Source: `library/paper_cards/MI-014.md`

## Paper Access

* Internal PDF: <a href={"/papers/MI-014.pdf"} download style={{ display: "inline-flex", alignItems: "center", justifyContent: "center", minHeight: "2.25rem", padding: "0.45rem 0.8rem", borderRadius: "6px", backgroundColor: "#047857", color: "#ffffff", fontWeight: 700, lineHeight: 1, textDecoration: "none", boxShadow: "0 1px 2px rgba(15, 23, 42, 0.22)" }}>Download Paper</a>
* DOI / official page: [10.1109/MSP.2007.909009](https://doi.org/10.1109/MSP.2007.909009)
* Deployment boundary: these PDF links are intended only for a private/protected internal wiki.

## MI-014: Optimizing Spatial Filters for Robust EEG Single-Trial Analysis

## Metadata

* ID: MI-014
* Title: Optimizing Spatial Filters for Robust EEG Single-Trial Analysis
* Year: 2008
* DOI / URL: 10.1109/MSP.2007.909009
* Local PDF: see Paper Access section above
* Text artifact: local-only path withheld from docs site
* Review status: `extracted`

## Study Type

* Track: MI
* Task: explain and optimize CSP-type spatial filters for single-trial EEG
* Participants or dataset: tutorial/review with example BCI datasets; exact dataset details need confirmation
* Device/electrode setup: local text includes examples with 55 Ag/AgCl electrodes and sensorimotor channels C3/C4
* Protocol/task: sensorimotor rhythm modulation and left/right imagery classification examples

## Methods

* Signal processing or analysis: common spatial patterns, variants, regularization, and practical tuning
* Training/calibration: supervised calibration trials are central to the CSP pipeline
* Online/offline: method-focused tutorial and examples

## Key Results

* The paper gives the theoretical and practical foundation for CSP as a robust single-trial EEG spatial-filter method.
* It explains why spatial filtering helps recover discriminative sensorimotor rhythms from volume-conducted EEG.

## Limitations

* Does not solve nonstationarity or no-control by itself.
* Not a robotic grasping study.
* Detailed numerical outcomes should be extracted only if cited from the example sections.

## Relevance To Current Review

* Core CSP lineage citation between Ramoser 2000 and FBCSP/regularized CSP variants.
* Supports MI as a spatially filtered sensorimotor-rhythm decoding paradigm.

## Evidence Status

| Claim | Status | Evidence Note |
| --- | --- | --- |
| CSP is a classical spatial-filter method for MI/SMR EEG classification. | verified | Abstract and introduction explicitly focus on CSP for BCI. |
| Sensorimotor channels and mu/beta rhythms are important for MI decoding. | verified | Local text discusses C3/C4, SMR, and imagined movement. |
| CSP alone is sufficient for safe stop/cancel control. | needs confirmation | No no-control or robot safety loop is evaluated. |

## Open Questions

* Should Exp2 use plain CSP as the minimal baseline before FBCSP and EEGNet?
* Which regularized CSP variant is most appropriate for small local calibration data?