turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- NI Discussion Forums
- :
- Special Interest Boards
- :
- Additional Interest Boards
- :
- Life Science
- :
- Using FFT for EEg signal monitoring

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic to the Top
- Bookmark
- Subscribe
- Printer Friendly Page

Highlighted
# Using FFT for EEg signal monitoring

Options

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report to a Moderator

07-14-2009 01:08 PM

I saw in old thread some posted:

*Spectral monitoring of EEG signals is pretty straightforward. Normally, the time domain signal is broken into short epochs of a few seconds and an FFT is performed on that data array. The data from the FFT is binned into frequency ranges according to standard EEG definitions: Delta, (0-4Hz), Theta (4-8Hz), alpha (8-12Hz) , etc. The Peak power value in that frequency range and/or the cumulative power (total power in that bin) is reported. Then the next segment of EEG data is processed and this continues either in real-time or on a prerecorded file of EG. This process is duplicated for each channel of data corresponding to different locations on the brain.*

I need to get a ratio of the power levels of two bands, 0-5 Hz and 20-40Hz, of a single channel EEG. How do bin the output of FFT? Am I using the correct Vi for this (FFT Spectrum (Mag-Phase) VI )? The Spectal Measurements Express VI look like it might be able to do this too but I cannot figure how to bin the ranges. Can I get the area under the plot some how?

We are are using a NI PCI-6221 with a NI BNC2090A box and of course the proper saftey precautions for the subjects.

Thanks for your help.

Solved! Go to Solution.

Solution

Accepted by topic author Richard_Z.

08-27-2015
06:09 PM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report to a Moderator

07-14-2009 03:44 PM

All you need to do is sum the values in the output array that correspond to the frequency band of interest. The output cluster includes a "df" value (the spacing or interval/resolution of the output values in Hz) so its easy to index into the output array and get the array subset needed for your frequency bin. Then just sum the array subset. This is effectively the area under the curve. So if the df is 0.1Hz then for your 0-5Hz band you would just grab the array subset from the start of the output array to the 50th value and sum all these values. For the 20Hz-50Hz bin, grab the 200th point to the 500th point and sum all these. Then compare the two sums (bins).

Steve

Visit the NI Biomedical User Group at:

www.ni.com/biomedusers

www.ni.com/biomedusers

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report to a Moderator

07-15-2009 12:42 PM

Thanks that worked.