#14 · Systems biology

Multiparametric and scaling readouts define neural state labels from observed dynamics

The dominant thread is the assignment of discrete or scalar state labels to neural dynamics from structured observables, with state definitions dependent on chosen readouts and model structure. A time-resolved, multiparametric framework classifies electrode-level responses in neuronal cultures as excited, inhibited, or unchanged. A generalized Taylor's law with RMS scaling, applied to detrended fMRI, links the scaling exponent to synchrony and reports shifts across lifespan and in ADHD. DyssomniaNN couples two-process sleep regulation, hypothalamic flip-flop circuitry, and corticothalamic LIF networks to model human sleep disorders and perceived restedness. The items are heterogeneous; the third is a constrained mechanistic model. What is new is the use of multiparametric, time-resolved, or generalized scaling readouts to define states, alongside a multi-scale spiking model addressing perceived restedness. Whether such labels and parameters are identifiable from the available observables remains the open question.