Describe white noise intensity

[Wei, Jingjin]

Hi all,

I was wondering if anyone has used the noise mech (noise.mech) in dynasim. How should I describe its intensity of it? Thanks!

% Gaussian noise with time scaling

noise_amp=1000 % units: 1/sqrt(ms)

Inoise(t) = noise_amp.*randn(1,N_pop).*sqrt(dt)

@current += Inoise(t)

Best,

Jingjin

[Jason Sherfey]

Hi Jingjin,

One way to describe the noise intensity is using the magnitude of the average current fluctuation, something like mean(abs(Inoise)). The below example demonstrates that approach.

data = dsSimulate('E: dv/dt=@current; {noise}; monitor functions','vary',{'noise_amp',[1 2 3]*1e3}); 

intensity = mean(abs([d.E_noise_Inoise]),1);

Output:

intensity =
  1×3 single row vector
   79.4017  160.7184  240.3138

There may be other analytical approaches as well. This is just the first that came to mind.

Alternatively, if your dt is constant, you could parameterize it using noise_amp. 

Hope this helps.

Best,

Jason

--
You received this message because you are subscribed to the Google Groups "DynaSim Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dynasim-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dynasim-users/BL0PR03MB41482BA52592CCC5CF10E172A7790%40BL0PR03MB4148.namprd03.prod.outlook.com.

[Jason Sherfey]

Correction ("data", not "d"):

intensity = mean(abs([data.E_noise_Inoise]),1);

[Wei, Jingjin]

Hi Jason,

Thanks! I've been thinking about using noise_amp directly, but not sure what should be the unit of it. For instance, we can say the I in Hoghkin-Huxley equation is uA/cm^2. When I say the white noise intensity is noise_amp=1000, what is this unit?

Best,
Jingjin

---

From: dynasi...@googlegroups.com <dynasi...@googlegroups.com> on behalf of Jason Sherfey <jssh...@gmail.com>
Sent: Wednesday, July 22, 2020 10:49 AM
To: Wei, Jingjin <jing...@bu.edu>
Cc: DynaSim Users <dynasi...@googlegroups.com>
Subject: Re: [DS] Describe white noise intensity

 

To view this discussion on the web visit https://groups.google.com/d/msgid/dynasim-users/CADYGBjQRYN4Pa-aKXktut-c_9WT4iSSb%2BPkFDqJo83wDOV3dyA%40mail.gmail.com.

[Jason Sherfey]

Hi Jingjin,

The units of noise_amp are 1/sqrt(ms), so it's probably not most meaningful to report it directly. One option is to use noise_amp to index relative change in noise level: vary noise_amp across simulations, plot (% of baseline noise level) in figures, and report baseline <intensity> in uA/cm^2 in the Methods section. This works as long as you don't change dt. Does that make sense?

Best,

Jason

[Julia Chartove]

Popping in with a general comment that it would be great if the Dynasim documentation (or some feature of Dynasim itself) made it easier to tell which quantities have which units, because I've repeatedly run into this specific issue myself and never found a systematic way to address it

To view this discussion on the web visit https://groups.google.com/d/msgid/dynasim-users/CADYGBjRq3mKrVJYnXCeMoj_s6Pi2BducR23Bzhf5NOpKKg3ppg%40mail.gmail.com.

You are subscribed as char...@bu.edu Manage your subscriptions | Unsubscribe Mailing list powered by GroupSpaces - Manage your group online