Daniele Marinazzo

SIMBA: Scalable Image Modeling using a Bayesian Approach, A Consistent Framework for Including Spatial Dependencies in fMRI Studies Bayesian spatial modeling provides a flexible framework for whole-brain fMRI analysis by explicitly incorporating spatial dependencies, overcoming the limitations of traditional massive univariate app...

And the next step? Full voxel-level modeling.

Recent numerical advances cracked the scalability barrier. Voxel-level hierarchical modeling is now feasible, revealing just how punishing traditional multiple-comparison adjustments really are.
arxiv.org/abs/2511.12825

We propose applications to physiological networks, climate, and finance

Unlike traditional methods treating measurements as independent snapshots, PIRD captures temporal structure to reveal whether multiple factors work independently, provide overlapping information, or only show influence when considered together over time.

We developed a mathematical framework called Partial Information Rate Decomposition (PIRD) that accounts for how information evolves in dynamic systems by recognizing that measurements depend on past conditions.

Combo of two papers on partial information rate decomposition now out!

journals.aps.org/prl/abstract...

journals.aps.org/pre/abstract...

Mini thread below 👇

Totally my experience too

It's just yet another journal.. The board moved because of high APC, not because of being Elsevier. Then the same community pays much more to publish in Nature Communications

Indeed comments are clunky, cc-ing @richardsever.bsky.social .

Functional connectivity is dominated by aperiodic, rather than oscillatory, coupling Functional connectivity has attracted significant interest in the identification of specific circuits underlying brain (dys-)function. Classical analyses to estimate functional connectivity (i.e., fil...

And FC is dominated by 1/f, with a comment of mine mentioning that issue, which of course remained unanswered 😭
www.biorxiv.org/content/10.1...

Well but also amplitude modulation = E/I balance 😂😂😂.
Also, brain dark matter!

ok, imagine you have an oscillation that is not symmetric around 0 (small direct current shift) with some amplitude modulation, for instance with 1/f-dynamics.

➡️ then these 1/f-dynamics will show up in low frequency part of spectrum (red); in addition to around oscillation peak (yellow).

A comparison between power spectral density and network metrics: An EEG study Power spectral density (PSD) and network analysis performed on functional correlation (FC) patterns represent two common approaches used to characteri…

See for example www.sciencedirect.com/science/arti...

pubpeer.com/publications...

I find this paper really confusing. There are several measures (power, coherence, network communication, information based statistical dependencies) which are all measures of the "behavior" of the system, and definitely share similarities. Yet some are considered "mechanisms" underlying the others.

We should definitely not blindly talk about "beta connectivity" without asking ourselves whether there's beta activity in the first place, and even more so when there isn't, 1/f is prominent (as it is to be expected). This was (at least I think now 😅) my comment

Functional connectivity is dominated by aperiodic, rather than oscillatory, coupling Functional connectivity has attracted significant interest in the identification of specific circuits underlying brain (dys-)function. Classical analyses to estimate functional connectivity (i.e., fil...

Weird, is the comment to the preprint
www.biorxiv.org/content/10.1...

At short scales all is aperiodic, at long scales, all is periodic
disq.us/p/30ouev0

With the Em Dash, A.I. Embraces a Fading Tradition

Some see these as hallmark of AI generated text, but apparently is not necessarily the case
www.nytimes.com/2025/09/18/m...

You could have rather put the quotation marks in "arousal", given the last sentence of the paper.
Nice and impressive, but it's basically saying that there is a low-dimensional descriptor. Now, it could have been reversed having latent brain dynamics as a predictor, and pupillometry as the target

There are so many assumptions we make, and so many mechanisms that could originate the same spectral features (both "periodic" and "aperiodic"), that pretending to "separate" them (are they separable?) is super difficult if not ill-posed.

thanks!. I appreciate your effort, and I have been thinking a lot about this "separation" too. But outside the framework of a (biased imo) generative model of "pure oscillations" plus 1/f, how can this be applied to real data?

Assessing metadata privacy in neuroimaging The ethical and legal imperative to share research data without causing harm requires careful attention to privacy risks. While mounting evidence demonstrates that data sharing benefits science, legit...

arxiv.org/abs/2509.15278 check that your metadata are 'private' i.e. that they do not leak personal information -- BIDSapp available 😀

Issues | Cerebral Cortex | Oxford Academic Publishes papers on the development, organization, plasticity, and function of the cerebral cortex, including the hippocampus.

A birthday is a good time to reflect! To mark SPM @ 30, this month's issue of Cerebral Cortex features deeply insightful commentaries on neuroimaging analysis from Ed Bullmore, Peter Bandettini, Peter Fox, Pedro Valdes-Sosa, Klaas Enno Stephan, Viktor Jirsa, et al [1/2] academic.oup.com/cercor/issue

a slide mentioning that colliders and confounders (and limits in pairwise and partial correlations) are well known in psychometrics and epidemiology, but much less so in neuro

ha! you wouldn't tell..
I always use this slide

BeyondTheEdge goes to Sicily BeyondTheEdge researchers participated in the School of Complexity on

BeyondTheEdge at the School of Complexity on "Higher-order interactions: mechanisms, behaviors, and networks" by our own @lordgrilo.bsky.social and @gin-bianconi.bsky.social with @aliceschwarze.bsky.social, @danielemarinazzo.bsky.social. Great perspectives! www.beyondtheedge.network/articles/bey...

Congratulations, you'll do great things

It's a self answering question, those are defined by the association, so if they do stuff together, call them the same. Or do as some smart asses who talk about "coactivation"

I'm specifically sharing this paper because I saw a post recently that said people who think about methodology are "attacking authors"

"Brief segments of neurophysiological activity enable individual differentiation", but also when the neurophysiological activity (MEG data) is identical for all the subjects, we have the same differentiation.

Because of the head shape, of course.

pubpeer.com/publications...

Combining Multiple Functional Connectivity Methods to Improve Causal Inferences - PubMed Cognition and behavior emerge from brain network interactions, suggesting that causal interactions should be central to the study of brain function. Yet, approaches that characterize relationships amo...

Exactly! Quite overlooked in neuroimaging, to the point that papers like this one are necessary pubmed.ncbi.nlm.nih.gov/32427070/

www.sciencedirect.com/science/arti...