Abstract
Signals from brain functional magnetic resonance imaging (fMRI) can be efficiently represented by a sparse spatiotemporal point process, according to a recently introduced heuristic signal processing scheme. This approach has already been validated for relevant conditions, demonstrating that it preserves and compresses a surprisingly large fraction of the signal information. Here we investigated the conditions necessary for such an approach to succeed, as well as the underlying reasons, using real fMRI data and a simulated dataset. The results show that the key lies in the temporal correlation properties of the time series under consideration. It was found that signals with slowly decaying autocorrelations are particularly suitable for this type of compression, where inflection points contain most of the information.
Highlights
The large-scale dynamics of the brain exhibit a plethora of spatiotemporal patterns
We revisited the heuristic point process approach originally introduced by Tagliazucchi et al [1,2,3] to represent brain spatiotemporal dynamics in terms of the relatively high-amplitude inflection points of the BOLD signal
Why it works: The present results show that the point process (PP) approach works due to a rather trivial fact: in any case of rather strongly autocorrelated signals, the most informative points are the inflection points; the remaining samples are more or less interpolated by straight lines which can in principle, and for certain applications, be ignored
Summary
The large-scale dynamics of the brain exhibit a plethora of spatiotemporal patterns. An important methodological challenge is to define adequate coarse-graining of the brain imaging data which comprises thousands of the so-called BOLD (“blood oxygen level dependent”) time series. A qualitative comparison of how well it works: It has already been established, in different circumstances [2,3,4], that the co-activation matrix obtained with the point process method is very similar to the correlation matrix computed from the full (i.e., continuous) BOLD signal. We ask how much of the raw signal is left out if these inflection points are used to extrapolate a piece-wise linear time-series To answer this we analyze BOLD time series from the brain of a subject during an experiment in which fMRI data are collected at rest [3]. Panel D shows that as the BOLD signal autocorrelation increases, the similarity between the piece-wise linear and the raw signals increases, evaluated in two ways: by the root mean squared error (rmse) and by the linear correlation r between the two time series. As expected, raising the threshold ν above zero produces an increasing loss of information about the signal, which is reflected in a monotonic increase in the rmse and a decrease in the r values (see Panel E)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
