Understanding the unit of measurement used on the Chart

[jaycode]

First of all, thank you for the amazing app. I have purchased three copies to help with my research.

On the chart, the absolute average of all four items is calculated with the following function:

Code: Select all

      function scaleBrainwaves(value){
        return (value+1)*50;
      }

May I know the reasoning behind this calculation instead of just averaging the values from all four electrodes? Also, what is the unit of measurement for the vertical axis of the menu "Brainwaves"?

Thank you.

[James]

I do average the four sensors. The above is run after that.

It's just a linear scaling function to convert from a -1:+1 range to a more understandable 0:100 range.
The reason for this, is that it's just more understandable for the laymen when given numbers that go from 0 to 100.
The units are bels.

[jaycode]

Ah, got it, thank you for your explanation.

By the way, I am looking to reproduce the research in this paper:
http://www.scielo.br/scielo.php?script= ... so&tlng=en

But the paper uses microvolts squared/Hz (μV^2/Hz) as its units of measurements. How do I convert bels to this unit?

[jaycode]

I do average the four sensors. The above is run after that.

It's just a linear scaling function to convert from a -1:+1 range to a more understandable 0:100 range.
The reason for this, is that it's just more understandable for the laymen when given numbers that go from 0 to 100.
The units are bels.

Wait, there is something I do not quite get here. If you changed the range of values, does that mean the bels units can still be used? Or should I remove the range conversion if I want to see the graph in its actual units?

[James]

Power Spectral Density doesn't have an official unit and the values are only relative to one another, so you can convert to any range you like provided you are consistent.

Bels are used as a unit for anything which is derived from the result of a logarithm, which PSD is.

[debra peri]

James,

Could you add some of this info to the tech docs? It's helpful but I only found it accidentally just now by looking over the forum topics.


Thanks!

[giraffepigcat]

Hi James,

Thank you so much for the app and website! And I'm looking through the issues concerning scales and units of measurements.

1. The absolute band power for a given frequency range (for instance, alpha, i.e. 9-13Hz) is the logarithm of the sum of the Power Spectral Density of the EEG data over that frequency range.

2. The EEG PSD values as read from the sensors are commonly in the {-1:+1} range. For the text value display these are shown in a more legible {0:100} range.　　　display_value = (value+1)/0.02

So does it mean the range without logarithm is (0.1~10) Bel, or (1~100) dB?
Thanks in advance!

[James]

I say "dB" just because it's recognizable and most people don't know what "bel" is.
The units don't matter, they just indicate that the value is derived from a logarithm. Try not to think about it too much, because it doesn't matter

[giraffepigcat]

Thank you again, James!

I just got dB or Bel is a unit that measures the ratio of the quantity of two identical units, say P1 and P0 ---> take log.

Asuming that P1 is for the brainwaves bands we are watching, i.e, Delta, Theta, Alpha, Beta, or Gamma, may I ask where is P0 from? Is P0 from low_freqs (2.5-6.1Hz)?

The accroding link specifies,
https://web.archive.org/web/20181105231 ... and_Powers
low_freqs 2.5-6.1Hz
delta_absolute 1-4Hz
theta_absolute 4-8Hz
alpha_absolute 7.5-13Hz
beta_absolute 13-30Hz
gamma_absolute 30-44Hz

Yi-Ting Chen

[James]

Sorry, I don't know what you mean by P0 or P1?
If you're asking how the FFT works, take a look at these code examples: https://www.mathworks.com/help/matlab/ref/fft.html