“INHIBITED inhibitory feedback theory of autism”
Think Goldberger’s frame first: healthy brains hover in a complex, fractal, near-critical regime; illness pushes them toward either too-regular (rigid, over-synchronous) or too-random (noisy) dynamics—both mean reduced multiscale complexity and lower resilience to perturbations (i.e., you hit tipping points sooner).
Baseline away from criticality (narrower safety margin).
EEG/MEG in autistic cohorts often shows lower Lempel–Ziv complexity and attenuated long-range temporal correlations (LRTCs)—especially outside alpha bands—consistent with being less metastable and further from the sweet spot where systems flexibly absorb shocks. Less metastability ⇒ smaller perturbations push the system across thresholds (attention collapse, sensory overload, shutdowns; in vulnerable subgroups, seizures).
E/I balance and aperiodic slope (control-knob for thresholds).
Atypical excitation–inhibition (E/I) manifests in the 1/f aperiodic slope (flatter ↔ relatively more excitation). Shifts in this knob change the bifurcation distance: tipping can occur at lower input levels, and recoveries are poorer once perturbed. (In epilepsy—more common in autism—approach to ictal transitions often shows critical slowing down, i.e., slower return to baseline.)
Local damping hardware (PV interneurons & perineuronal nets).
Prefrontal PV-interneuron / perineuronal-net abnormalities reported in ASD models (and some human tissue work) imply weaker “brakes” and altered plasticity windows—biophysical ways the system can overshoot and overcorrect and still have low complexity (rigid attractors plus bursts), which makes tipping more likely under load.
Neuromodulatory gain (LC–NE) and precision.
Pupillometry suggests higher tonic and reduced phasic locus-coeruleus activity in autistic children—i.e., elevated baseline gain but poorer adaptive gain control. Elevated tonic gain both steepens responses (easier to cross thresholds) and reduces graceful settling, shrinking the basin of attraction.