Cannot handle `Status` channel recorded in a `bdf` file

[likanzhan]

Hi @ararslan and @rob-luke

Here is description of the problem. We can use the sample bdf data provided in Neuroimaging.jl.

In the sample data named test_Hz19.5-testing.bdf there are 7 channels in total: The first 6 channels are the normal signal channels and the last channel with the label name Status is a trigger channel as being specified in biosemi. In the sample data, there are 3 different triggers and 59 triggers in total. The trigger information can be correctly retrieved with BDF, but not with EDF.

using BDF, EDF, Downloads

url = "https://github.com/rob-luke/Neuroimaging.jl/blob/main/test/data/test_Hz19.5-testing.bdf?raw=true"
dt = Downloads.download(url)

• When the data is read in with readBDF(), the actual data contains only 6 channels. The last channel Status is converted into triggers.

header = BDF.readBDFHeader(dt)
header["chanLabels"] # seven channels, the last one is `Status`

data, triggers, trigger_channel, system_code_channel = readBDF(dt)
size(data) # (6, 237568)
length(unique(triggers["code"])) # 3
length(unique(triggers["idx"]))   # 59

• When the data is read with EDF.read(), the last channel was treated as a normal channel, and the trigger information as being described earlier cannot be retrieved.

edf = EDF.read(dt)
status = last(edf.signals)
status.header.label # "Status"
unique(status.samples)

Related discussions: Switch to EDF.jl #91.

[rob-luke]

You can find another test file at https://github.com/rob-luke/Neuroimaging.jl-example-data

[likanzhan]

Do we have an update on this?

[ararslan]

I've not had time to get around to this yet but I have some ideas for how to implement it.