Rank Vector Entropy: Role of a Leaky Integrator

[sinbi]

Hi All,

I am not sure if this is the right section to post this question but it does involve probability..so please redirect me if necessary.

I am currently looking at the Robinson et al. (2013) paper on rank vector entropy in MEG (doi: 10.3389/fncom.2012.00101). Due to my lack of mathematical knowledge I am struggling to understand what role the 'leaky integrator' performs in this algorithm (or what a leaky integrator does in general?).

Essentially a state histogram is produced counting all the occurrences of the rank vector states and the probability of a state occurring is calculated...however, this just produces the absolute entropy across time once Shannon's entropy has been applied. Am I right in thinking by introducing a time constant using the leaky integrator we are able to measure the relative fluctuations in entropy over time as opposed to the absolute entropy? And if so, how does the introduction of a time constant or leaky integrator achieve this? (I'm not sure mathematically how this gives us temporal information).


Apologies if this is posted in the wrong section or exposes my complete lack of knowledge!

Many thanks.

 

[mmwave]

Thank you for your question about the Robinson et al. (2013) paper on rank vector entropy in MEG. The leaky integrator plays an important role in this algorithm, as it helps to measure the relative fluctuations in entropy over time rather than just the absolute entropy.

A leaky integrator is a mathematical concept that describes a system in which the output is dependent on both the current input and the previous output. In the context of this paper, the leaky integrator is used to calculate the time constant, which is a measure of how quickly the system responds to changes in the input. This time constant is then used to calculate the relative fluctuations in entropy over time, providing a more dynamic measure of entropy rather than just a static value.

To better understand the role of the leaky integrator, it may be helpful to think of it as a filter that allows for a more nuanced analysis of the data. By introducing the time constant, the algorithm is able to capture the changes in entropy over time and provide a more comprehensive understanding of the data.

I hope this helps to clarify the role of the leaky integrator in the algorithm. Please let me know if you have any further questions or concerns. Thank you for your interest in this topic and for your contribution to the scientific community.