Investigate behaviour of `.mat` logger saving

Recently I found a strange behaviour in the `.mat` logger. If the acquisition is longer than a fixed interval (probably hardcoded in the code with an autosaving function), the logger split into several `.mat` files the acquisitions. Actually the problem doesn't lie on this aspect. The problem lies on the fact that all the fields inside the `.mat` struct are consistent **only** in the case in which the acquisition consists of one file, i.e, all the elements are the same number of samples and the same timestamp. When the acquisition is split into two or more `.mat`, elements inside each struct are no longer consistent, i.e., they have different number of samples and timestamps do no coincide.
A patch solution for me was to concatenate all the `.mat` files and I've verified that, **after concatenation**, both number of samples and timestamps coincide.