CNS*2026 Halifax

Introduction
In a recent study on wide-field calcium images in mice before and after stroke [1] singular value decomposition (SVD) was used to perform a spatiotemporal analysis of movement-evoked global cortical events. Indicators such as angle and smoothness of the propagation were defined that allowed to compare different stroke rehabilitation therapies. This method worked because the events were truly global (e.g., basically all pixels in the field of view were participating). By contrast, in a follow-up study on a mouse model of autism we now also include non-complete events but for such events SVD does not work. Thus, novel and more general approaches are needed and this need is addressed here.

Methods
Adaptive coincidence detection and the SPIKE-synchronizarion and SPIKE-order framework [2] are used to identify (global or non-global) events and to sort the participating pixels of each event from leader to follower. From the resulting two-dimensional propagation patterns, we define various indicators such as completeness (fraction of participating pixels), connectedness (clustering of these pixels) and correlation (similarity of rank order among neighboring pixels). The angel of propagation is defined from the resultant length [3] of the direction vectors of all pairs of participating pixels and its carefully renormalized amplitude is an indicator of the strength of the propagation.


Results
We illustrate the new methods using both simulated data for verification and experimental data for exploration. These are typically non-global events
recorded using wide-field calcium imaging to monitor cortical activity in a Shank3b mouse model of autism from late development through adulthood, and under isoflurane anesthesia to manipulate the brain state [4].
We here show that these new methods, and in particular the angle and the strength of the propagation, generalise and improve on the original SVD method in terms of both accuracy and speed.

Discussion
Together, the five indicators completeness, connectedness, correlation, angle and amplitude provide a full characterization of the spatiotemporal activity. Importantly, this new approach is so far the only method that allows calculating the angle and the strength of the propagation for non-complete global events.\u2028
The corresponding scientific article is currently under preparation.

References
[1] Cecchini, G., ... Kreuz, T. (2021). Cortical propagation tracks functional recovery after stroke. PLoS Comput Biol 17: e1008963. https://doi.org/10.1371/journal.pcbi.1008963 
[2] Kreuz, T., ... Mulansky, M. (2017). Leaders and followers: Quantifying consistency in spatio-temporal propagation patterns. New Journal of Physics 19, 043028. https://doi.org/10.1088/1367-2630/aa68c3
[3] Andrzejak, R. G., ... Schindler, K. (2023). High expectations on phase locking: Better quantifying the concentration of circular data. Chaos 33, 091106. https://doi.org/10.1063/5.0166468
[4] Montagni, E., ... Allegra Mascaro, A. L. (2025). Age-dependent cortical overconnectivity in shank3 mice is reversed by anesthesia. Translational Psychiatry 15 (1), 154. https://doi.org/10.1038/s41398-025-03377-5

Acknowledgement
This work has been funded by Telethon Seed Grant Spring Renewal 2025 PHEM (GSA25E002), the Italian Ministry of Universities and Research on the project THE Tuscany Health Ecosystem (ECS_00000017), MUR_ PNRR.