Neuronal loss reshapes survivor dynamics and limits mechanism inference in excitatory inhibitory neural fields

Neuronal loss can reshape the behavior of surviving neural networks, beyond merely reducing measured activity, according to a new study. The researchers utilized a next-generation excitatory-inhibitory neural field model to differentiate between these effects. By expressing the viable population measure as q_a = lambda_a f_a, with lambda_a representing viable population mass and f_a denoting the normalized survivor distribution, they found that normalization commutes with the Ott-Antonsen/Montbrio-Pazo-Roxin reduction under specific conditions.

This means that while mortality terms disappear from conditional transport, viable mass remains in recurrent coupling, demonstrating that loss can alter survivor dynamics. Conversely, when parameters are homogeneous, viability, pathway integrity, and compensation can lead to conjugate conditional deterministic dynamics, indicating that identical conditional activity does not necessarily imply the same biological mechanism.

The study also reveals that equilibrium and oscillatory bifurcations, finite-population escape, and delayed propagation are consequences of these principles. Finite-population escape and delayed propagation, in particular, highlight the impact of viability-dependent propagation and phase relaxation. The researchers emphasize that attributing activity changes to neuronal loss requires additional information, such as tissue-level measurements or independent structural constraints, which must be interpreted through an observation model.