The NOW model (Riddle & Schooler, 2024) proposes that consciousness is a hierarchy of observer windows, systems that integrate information within spatial and temporal bounds, nested across multiple brain scales with a unitary experience at the apex.
The Three Signature Mechanisms
Synchrony
Zero phase lag within a window. Elements align in phase to form a coherent observer window. Measured by the Kuramoto order parameter.
Coherence
Non-zero phase lag between peer windows. Enables communication within the same level through stable phase relations with delays.
Cross-frequency coupling
Communication across scales. Low-frequency phase at higher levels couples to higher-frequency amplitude at lower levels (PAC).
The Mosaic Metaphor
Schooler describes consciousness as a mosaic photograph: every pixel is itself a photograph. This represents a fractal system of windows nested in windows. The Apex Window is the top-level integrator that what we typically call I or my consciousness.
Signal-Level Formalization
The Kuramoto order parameter measures phase synchrony within a group of oscillators:
r=N1j=1∑Neiθj
Where r=1 indicates perfect synchrony (all phases aligned) and r=0 indicates uniform phase distribution.
Testable Predictions
Window identification: Observer windows should be recoverable as high internal synchrony ensembles that reconfigure with task demands.
Vertical information flow: Changes in conscious access (sleep onset, anesthesia) should show structured degradation of cross-frequency coupling across levels.
Horizontal dialogue: Competing interpretations should show coherence patterns consistent with lateral negotiation among same-level windows.
What to Try
Set noise high and observe how synchrony degrades.
Increase within-synchrony to see windows become more coherent.
Adjust cross-frequency coupling to see how parent-child modulation affects report stability.
Watch how all three mechanisms must work together for stable apex report.
Model provenance
claude opus 4.8January 2026·first cut. builds the Nested Observer Windows sandbox: a hierarchy of phase oscillators grouped into windows across levels, with within-window synchrony, lagged peer coherence, a cross-frequency-coupling summary, and a heuristic report-stability output. adds a calibration of the deterministic core (Kuramoto endpoints and the report-stability law), seven assumptions separating mechanism from interpretation, and a research companion.
Calibration
Assumptions
Model changelog
v1.0January 2026
oscillator hierarchy: numLevels levels, each with windowsPerLevel windows of phase oscillators, slower intrinsic frequency toward the apex, a single window at the apex.
within-window synchrony pulls phases toward their circular mean; the Kuramoto order parameter r = |mean(exp(i*theta))| reads out window coherence.
peer coherence pulls window means toward a non-zero preferred lag, separating zero-lag synchrony (within) from lagged coherence (between).
cross-frequency coupling scalar: parent mean phase modulates a child amplitude envelope, averaged over parent-child level pairs.
report stability composite: clamp01(syncApex*(0.55+0.45*bandwidth)*(0.6+0.4*avgCoherence)), encoding joint necessity of all three mechanisms.
live time series of apex synchrony, cross-frequency coupling, and report stability.
calibration targets the deterministic, noise-free core (Kuramoto order at exact endpoints, report-stability law at hand-worked operating points) since the live run is stochastic.
framing kept honest: the toy illustrates the NOW hypothesis and does not test it; the report-stability law is marked speculative.