Signal Studio is in public beta. Usable, but still moving. Join the beta

Publications

Work that used Signal Studio. Nothing is listed without the author asking, and each entry links the pipeline where one was shared.

Loading…

Adding yours

Open a thread on the forum titled publication: followed by the reference. Include the DOI, the version of Signal Studio you ran, and the pipeline file if you can share it. Entries appear here once confirmed.

Why bother: a citation is what makes a project like this legible to funders and to other labs. It is also the only way anyone finds out that the thing they are considering has been used in anger.

Reproduced analyses

Separately from papers that used the software, these published analyses have been rebuilt as card pipelines, end to end, on public data. They are how the card library is tested against something real.

CCEP to seizure onset zone

11 cards / ds004457 sub-1

Bipolar re-referencing, epoching, variance rejection, artefact removal, averaging by stimulation site, Morse wavelet, FDR-corrected significance, zone fractions, graph centrality, then comparison against the clinical label.

The recipe

CCEP response metrics

4 cards / current, pulse width, charge

N1 and N2 peak amplitude and latency with adjustable windows, root-mean-square response metrics, waveform correlation, and a linear mixed model over stimulation parameters.

The card

Network model

4 cards / interictal segments

A linear time-varying model fitted over sliding windows, virtual stimulation through that model, an excitability matrix, and a cross-validated classifier.

Validation