How to properly stream band power values

[sccn]

How to properly stream band power values

[gerardklee]

Sorry, the code is here.

[cboulay]

We can't tell you what's in the stream. It's up to the sender to tell you. Try browsing the inlet.info().as_xml() contents to see if there's any information there.

If not, then ask whoever wrote the code to send the stream to please add that information to their stream metadata.

[gerardklee]

Thank you @cboulay. I actually wrote the code. Would you explain more about "It's up to the sender to tell you what's in the stream"?

Basically, I did this trial for three minutes and got the values from the power bands expecting some increase over time in alpha waves (I had my eyes closed the entire trial). I plotted the csv file and could not really find any pattern. Could it be the electrodes being worn out that could lead to very random values?

[agricolab]

Dear @gerardklee, regarding

"It's up to the sender to tell you what's in the stream"?:

The outlet broadcasts information about itself. Besides the defaults (e.g. name, type), the creator can add additional fields. A suggestion list of additional infos for EEG is e.g. https://github.com/sccn/xdf/wiki/EEG-Meta-Data. By calling inlet.info().as_xml() you can inspect these fields.

Considering https://github.com/OpenBCI/OpenBCI_LSL/blob/master/lib/streamerlsl.py#L144, it seems the LSL Outlet will contain additional info, e.g. channel labels and units (i.e. µV). This suggests the data is samples. Looking at https://github.com/OpenBCI/OpenBCI_LSL/blob/master/lib/streamerlsl.py#L100 and https://github.com/OpenBCI/OpenBCI_LSL/blob/3134a1513c204834d66084491f104e6f236e440b/lib/open_bci_v3.py#L133, it also seems to be raw samples of EEG data, and not band power. This means, you would have to calculate the spectrum yourself. But you might get a definite answer if you ask there.

[gerardklee]

Thank you very much @agricolab. I will try your suggestions!