Trouble understanding "Lags" for Cross Correlation

I've read through the help page for cross correlations but I'm still having trouble understanding what lags actually represent.

The help page refers to them as "Lag indices, returned as a vector," but I'm having trouble understanding what they actually means in terms of my data.

Let's say I have two sets of data, each equals a 10 second trial of data and contain 1000 data points.

Does the lag represent a unit of time? e.g. 1 lag = 1 second? So if my correlation coefficient is highest at lag = 1, would it mean the relationship is strongest with a phase shift of 1 second?

Thank you.

Answer 1

[Honglei Chen]

1 lag is essentially 1 sample, so how much time it represents depends on your sample rate. In your case, since for 10 sec data you have 1000 data points, it means that your sample rate is 100 Hz, so 1 lag represents 0.01 sec.

HTH