Section 2.4.1.3 of the lossless paper describes a step to flag channels/epochs based off a 50 microvolt fixed threshold. something like:
input: channel x epoch x times array
take std across the times axis (return a channel x epochs array)
if std value of a cell is > 50, it is "above threshold".
flag chs & epochs:
For each channel, if the std val of > 20% of epochs are "above threshold" . Flag channel..
For each Epoch - if the std val for > 20% of channels are "above threshold", flag epoch.
The paper called this "staging" but given that 1) We are not convinced we should rei-implement allowing a staging script 2) The very Paper itself uses this step and says this step may be a common procedure users want to do... and 3) We need to do it just to validate pylossless against MATLAB losless... I propose we include this as an optional step of the pipeline.
Section 2.4.1.3 of the lossless paper describes a step to flag channels/epochs based off a 50 microvolt fixed threshold. something like:
The paper called this "staging" but given that 1) We are not convinced we should rei-implement allowing a staging script 2) The very Paper itself uses this step and says this step may be a common procedure users want to do... and 3) We need to do it just to validate pylossless against MATLAB losless... I propose we include this as an optional step of the pipeline.
EDIT: I am going take a stab at this today.