I am reading the paper "Classical and Bayesian Inference in Neuroimaging" by Friston. In the introduction he says:

"[...] a trivially small activation can be declared significant if there are sufficient degrees of freedom to render the variability of the action's estimate small enough."

I have heard the terms "degrees of freedom" in the context of neuroimaging numerous times but I never understood what it actually means. I sometimes also encounter the same term in a statistical context but I assume it's not the same as what Friston is referring to here. Or am I utterly mistaken?

  • $\begingroup$ In your quote they talk about significance, that is, the probability that some effect is present/absent. I believe they are talking about the statistical degrees of freedom. Here is a post from Cross Validated about the topic. However, I don't exactly know (and am interested to know) about the DF in the neuroscientific domain. $\endgroup$ – Robin Kramer Sep 22 '16 at 6:43
  • $\begingroup$ I've just quickly read the paper, and believe in this context degrees of freedom can be defined the same way as it is in a statistical context $\endgroup$ – queenslug Sep 29 '16 at 6:51

In a statistical context, any independent datum adds a degree of freedom. So when Friston talks about "sufficient degrees of freedom to render the variability of the action's estimate small enough" he simply means "sufficient sample size to cancel out noise". At least that's my take on it, based only on the single quoted phrase(!).

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.