# SSVEP-027: Short-length SSVEP data extension by a novel generative adversarial networks based framework > 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/SSVEP-027.md` ## Paper Access * Internal PDF: Download Paper * DOI / official page: [needs confirmation](https://arxiv.org/abs/2301.05599) * Open-access page: [Open access page](https://arxiv.org/abs/2301.05599) * Deployment boundary: these PDF links are intended only for a private/protected internal wiki. ## SSVEP-027: Short-length SSVEP data extension by a novel generative adversarial networks based framework ## Metadata * ID: SSVEP-027 * Title: Short-length SSVEP data extension by a novel generative adversarial networks based framework * Year: 2023 * DOI / URL: https://arxiv.org/abs/2301.05599 * Local PDF: see Paper Access section above * Text artifact: local-only path withheld from docs site * Review status: `extracted` ## Study Type * Track: SSVEP * Task: short-window SSVEP data extension using GANs * Participants or dataset: two public SSVEP datasets, including a 4-class and a 12-class dataset * Device/electrode setup: local text mentions public datasets and 8 Ag/AgCl electrodes for one dataset context; exact mapping needs confirmation * Protocol/task: SSVEP classification with short-length signals ## Methods * Signal processing or analysis: TEGAN transforms short-length SSVEP signals into longer representations * Comparators: traditional frequency recognition methods and deep baselines * Training/calibration: evaluated on public datasets ## Key Results * The abstract states that TEGAN improved recognition performance when assisting traditional frequency recognition methods. * The paper addresses short data length and limited calibration data, both central to practical SSVEP BCIs. ## Limitations * Synthetic data benefits may not transfer to dynamic visual targets. * GAN-generated EEG requires careful validation to avoid inflated offline claims. * Published DOI/version needs confirmation. ## Relevance To Current Review * Represents the data augmentation trend for short-window SSVEP. * Useful for future decoder experiments after establishing classical baselines. ## Evidence Status | Claim | Status | Evidence Note | | --- | --- | --- | | Short SSVEP windows and limited calibration data are recognized deployment barriers. | verified | Abstract motivates TEGAN around data length and calibration limitations. | | GAN-based extension is a recent SSVEP trend. | verified | The paper proposes a GAN-based TEGAN framework. | | GAN augmentation should be used as a first prototype requirement. | inferred | It may help later, but fixed-target public datasets differ from the project loop. | ## Open Questions * Which short-window target length is realistic for object selection before user fatigue rises? * Should synthetic SSVEP be evaluated only offline until online validation exists?