The Hierarchical Optima Hypothesis

Reconciling Heart-Mind Coherence with the Edge of Chaos

1. Executive Summary

Current complexity science often posits that systems (both human and computational) perform best at the "Edge of Chaos" (criticality). However, this contradicts the somatic experience of Heart-Mind Coherence, which feels like a state of effortless order. This document proposes a refined hypothesis: Coherence is the Primary Optimum for holistic intelligence, while the Edge of Chaos is a Secondary Optimum used for rapid adaptation when coherence is unavailable.

2. Refined Hypothesis Statement

3. Theoretical Distinctions

While popular science often conflates these states, their mechanisms and physiological signatures differ fundamentally:

Feature	Heart-Mind Coherence (Primary)	Edge of Chaos / Criticality (Secondary)
Mechanism	Synchronization: Smooth, sine-wave-like Heart Rate Variability (HRV); sympathetic/parasympathetic balance.	Sensitivity: Neuronal avalanches; power-law distributions; balanced order/disorder.
Physiology	High parasympathetic tone; low metabolic cost; "effortless."	High alertness; potentially higher metabolic cost; "tense" or "flux-like."
Function	Long-term strategy, wisdom, emotional integration, health.	Rapid reaction, pattern recognition, survival adaptation.
Experience	Grounded, clear, deep knowing, wholeness.	Alert, reactive, high-energy, ready.

4. The v2.1 Armor (October 2026 revision — the lens-passed upgrade)

The v1 spine above survives intact. This revision arms it with operational teeth, pre-registered controls, and named kill conditions — armor against being unfalsifiable, not against being wrong.

4.1 Operational Definitions

Coherence is measured, not felt: HF-band HRV power, respiratory sinus arrhythmia (the ~0.1 Hz resonance pattern), and high vagal tone. Criticality is measured by neuronal-avalanche exponents and HRV complexity. The hypothesis claims only the ordering of these two measurable states — it does not lean on any contested claim that criticality is optimal.

4.2 The Three Decisive Predictions

P1 (Bimodality): performance-versus-state curves are two-peaked across a mixed task battery — killing both "criticality is always best" and "calm is always best" in one stroke. P2 (Substitution): in low-coherence subjects, criticality is the highest-performing accessible state, and coherence training (daily 20-minute breath-paced practice at ~0.1 Hz, four weeks, assessments at baseline / week 2 / week 4) shifts occupants up the curve. P3 (Efficiency): for equal task performance, the coherence state runs at lower cost — measured by task-evoked pupillary dilation (the norepinephrine index) and EEG alpha power, with metabolic instrumentation where budget allows.

4.3 The Circularity Guard

State-classification cut-points are pre-registered before data collection, and the two peaks must survive classification-independent clustering (unsupervised clustering recovering the same state families). The peaks may not be built into the definitions that find them.

4.4 The Arousal Control and the Falsifiability Clause

The strongest attack on any coherence-ordering claim is the arousal confound. This proposal meets it head-on: arousal is matched across conditions (pupillometry, skin conductance), and the dual peaks are predicted to survive within matched arousal bands. The hypothesis is falsified if, with arousal controlled: (a) the performance curve is unimodal in either direction, (b) the coherence peak shows no efficiency advantage, or (c) fragmented subjects show no criticality substitution advantage. Any one of the three kills it. "Primary" is defined on three ranked axes: thermodynamic efficiency x stability x breadth — coherence wins the composite; criticality keeps the narrow rapid-adaptation slice.

5. Related: The Nighttime Half

This proposal is the daytime half of a three-part research spine. The nighttime half is The Resting Intelligence — Sleep, Dreams, and Circadian Architecture for Biological and Artificial Minds: coherence earns the rinse; the rinse consolidates coherence. The shared ledger beneath both is entropy administration — the third proposal of the spine. Three proposals, one spine, every hypothesis wearing its kill conditions.

The Hierarchical Optima Hypothesis — Exhaustive Research Blueprint

Author: Nathan A. Richards — Founder, Star‑Field.App & CoherentSignal.com • Contact: naterichards1111@gmail.com • Version: v1.0 • Date: May 28, 2026

Abstract

The Hierarchical Optima Hypothesis posits two ordered operational attractors: (1) a Primary Optimum — Heart‑Mind Coherence — a low‑entropy synchronized regime maximizing integrated intelligence, energetic efficiency, and durable decision‑making; and (2) a Secondary Optimum — Criticality (Edge of Chaos) — a high‑sensitivity, metastable regime functioning primarily as an exploratory bridging/search mode to reestablish coherence. This document is an exhaustive specialist blueprint: experimental designs, induction scripts, instrumentation, signal‑processing pipelines, computational modeling plans, power analyses, preregistration text (OSF‑ready), figure schematics, glossary, and a fully annotated APA reference list.

Contents

1. Hypothesis & Operational Definitions
2. Experimental Designs
3. Methods: Inductions, Instrumentation & Protocols
4. Data Preprocessing & Analysis Pipelines
5. Statistical Power, Sample‑size & Models
6. Figure Schematics & SVG Placeholders
7. Predictions & Decision Rules
8. Confounds & Mitigations
9. Glossary
10. OSF‑Ready Preregistration Protocol
11. Annotated References (APA) & Footnotes

1. Hypothesis & Operational Definitions

1.1 Concise Thesis

Refined thesis: The Edge of Chaos (criticality) functions as an adaptive, metabolically costlier bridging/search regime systems enter to explore state‑space and reconfigure toward Heart‑Mind Coherence (the primary, low‑entropy attractor). Criticality is thus a secondary optimum—highly sensitive and flexible but less integrative and energy‑efficient than coherence.

1.2 Operational Definitions

• Heart‑Mind Coherence (Coherence): high cardiovagal synchronization and cross‑modal neural synchrony; operationalized via HRV coherence ratio, phase locking between cardiac phase and EEG alpha/theta, and elevated RSA indices.
• Criticality (Edge of Chaos): metastable neural regime near a phase transition; operationalized via neuronal avalanches (power‑law distributions), DFA exponents indicating long‑range temporal correlations, positive largest Lyapunov exponents, and elevated signal variance/autocorrelation.
• Metabolic cost: salivary cortisol, continuous HR, skin conductance, and indirect calorimetry (VO₂) where available.
• Tasks: Algorithmic (high throughput, short horizon) vs. Integrative (long‑horizon, value‑integrative, creative/ethical synthesis).

2. Experimental Designs

2.1 Primary Experiment — Within‑subjects 2×2 Crossover

Goal: Directly contrast induced Coherence vs. induced Criticality across Algorithmic and Integrative task classes in the same participants.

Factor Levels

• State induction Coherence induction vs. Criticality induction (counterbalanced)
• Task type Algorithmic vs. Integrative (counterbalanced)
• Design Within‑subjects crossover; 4 cells per participant with minimum 24‑hr spacing or session washouts within a day

2.2 Secondary Experiment — Transition Dynamics (Longitudinal)

Goal: Induce coherence baseline, apply controlled perturbation, track state transitions (HRV + EEG) and test whether criticality markers appear as a transient exploration regime en route to coherence recovery.

2.3 Tertiary Experiment — Computational Analogs

Goal: Use spiking neural networks (SNNs), reservoir computing, and liquid‑state machines to simulate coherence (synchrony) vs. critical regimes to observe performance vs. energy proxies; use parameter sweeps to test hypothesis mechanistically.

3. Methods — Inductions, Instrumentation & Protocols

3.1 Participant Selection

• Adults 18–50, fluent in testing language.
• Exclusions: neurological/psychiatric diagnosis, medications affecting autonomic function (e.g., beta‑blockers), current substance dependence, unmanaged cardiovascular disease.
• Prescreen baseline HRV and trait mood questionnaires; stratify randomization by baseline HRV tertiles.

3.2 Instrumentation

• EEG: 64–128 channel cap, DC‑coupled amplifier, sampling ≥1000 Hz recommended for avalanche analysis (can downsample to 500 Hz with documented anti‑aliasing).
• ECG: 3‑lead at ≥500 Hz or validated chest strap providing millisecond R–R timing.
• Respiration: thoracic/abdominal belt.
• Peripheral: EDA, continuous noninvasive BP (optional), pulse ox.
• Metabolic: indirect calorimeter (VO₂) if available; salivary cortisol kits for pre/post measures.
• Stimulus & timing: Psychtoolbox / PsychoPy / Presentation; hardware TTL for event marking.

3.3 Task Battery — Concrete Protocols

Algorithmic Tasks (specs)

• N‑back (adaptive up to 3‑back): 300 trials, 500 ms stimulus, 1500 ms ISI, blocks of 50 trials, record accuracy/RT and d′.
• RSVP categorization: 5 blocks × 120 trials; targets embedded in 10 Hz stream; measure detection sensitivity and RT.
• Adaptive stochastic reward learning: 200 choices with probabilistic payoff switches; quantify adaptation latency and choice entropy.

Integrative Tasks (specs)

• Strategic planning vignette: 30 min; participants produce constrained plan; blind raters score across a rubric (foresight, integration, values weighting; ICC > .75 required).
• Ethical dilemma essay: 20 min; raters score nuance, tradeoff awareness, temporal scope.
• Creative synthesis: 20 min; prompts require integrating disparate domains; scored for novelty and coherence by trained judges.

3.4 Induction Protocols (Scripts & Thresholds)

Coherence Induction

Duration: 12 minutes induction, 5-minute maintenance.

1. Brief orientation (1 min): paced breathing at ~6 breaths/min (5s inhale, 5s exhale). Instruct on diaphragmatic breathing.
2. Positive-affect focus (30–60s): recall a sincere appreciation moment.
3. Biofeedback: visual HRV coherence gauge updated in real time (custom implementation or InnerBalance).
4. Maintain until coherence threshold met: coherence ratio ≥ 0.6 for consecutive 60s, then transition to task within 2 minutes.

Cues: metronome tones, on-screen breathing animation, single-line scripted prompts for affect focus.

Safety: pause if participants report dizziness or lightheadedness; check respiration depth to prevent hyperventilation.

Criticality Induction

Duration: 12 minutes induction, immediate task transition.

• Option A) Multi‑sensory unpredictable load: Continuous interleaved oddball streams (visual + auditory) with jittered ISI (200–1200 ms). Concurrent adaptive arithmetic checks (rule toggles).
• Option B) High‑variability working memory: Rapid, rule‑switching N‑back with unpredictable mapping and negative feedback (mild stress).
• Option C) Sensory stochastic perturbation: Amplitude-modulated white noise with Poisson timing while performing attention tasks.

Monitoring: Stop induction if HR > 85th percentile baseline or self‑reported distress exceeds threshold.

Rationale: elevate neural variability and responsiveness without causing panic.

3.5 Recording Procedures & Synchronization

• 5‑minute resting baseline (2.5 min eyes open, 2.5 min eyes closed) before inductions.
• Hardware TTL triggers for induction onset, task onset, trial events; log to EEG and ECG streams.
• Time‑stamp respiration and peripheral sensors for later artifact regression.

4. Data Preprocessing & Analysis Pipelines

4.1 EEG Preprocessing Pipeline

• Import raw; high‑pass 0.1 Hz, low‑pass 250 Hz; notch 50/60 Hz.
• Resample to 500–1000 Hz (document anti‑alias filtering).
• Detect bad channels (threshold + correlation methods) and interpolate.
• ICA for ocular/muscle/cardiac artifacts; label components automatically (ICLabel) then manual verification.
• Re‑reference to average; segment for event analyses; use continuous data for criticality metrics.

4.2 ECG/HRV Preprocessing

• Detect R‑peaks (Pan–Tompkins or validated library); visually inspect edge cases.
• Interpolate ectopic beats; reject >5% artifact segments.
• Compute RMSSD, SDNN, LF/HF, and HRV coherence ratio (narrowband at paced breathing frequency / total 0.04–0.4 Hz power).

4.3 Criticality Analytics

• Neuronal avalanche detection: threshold continuous EEG (z > 3 per channel), bin into time bins dt ≈ mean inter-event interval; generate avalanche-size distributions; fit power law via MLE and compare alternatives using likelihood ratios and KS statistics. Report xmin and bootstrap CIs.
• DFA: compute α over window ranges 0.5–100 s; report band‑specific and broadband exponents.
• Lyapunov exponent: reconstruct state space (time delay via mutual information, embedding via false nearest neighbors), compute largest Lyapunov exponent with short‑term algorithms; validate with surrogate data.

4.4 Cross‑Modal Synchronization

Compute phase of EEG bands (Hilbert transform) and cardiac phase from R‑peaks; compute PLV and wPLI; test against circular‑phase surrogates. Cross‑correlate HRV coherence ratio time series with DFA/avalanche metrics using lagged regression to test leading/lagging relations.

4.5 Behavioral & Statistical Modeling

Primary models use linear mixed effects (LME) / GLMM. Pre‑specify fixed effects: State, Task, State×Task, SessionOrder; random intercepts for subjects; add random slopes for State where data supports. Correct multiple comparisons via hierarchical testing and FDR for secondary outcomes.

lmer(Performance ~ State*Task + SessionOrder + (1|Subject) + (0+State|Subject), data=...)

5. Statistical Power, Sample‑size & Models

5.1 Effect Size Assumptions

Conservative effect estimates: d = 0.5 for Coherence advantage on integrative tasks; d = 0.35 for algorithmic contrasts. Correlation among repeated measures ρ ≈ 0.5.

5.2 Power & Sample Estimation

For within‑subjects repeated measures with α = 0.05 and power = 0.90: N ≈ 46 for d = 0.5; N ≈ 74 for d = 0.35. Recommended sample: N = 80 to cover attrition and data loss from artifacts.

5.3 Model Specifications (Primary)

Model 1 (behavior): Performance ~ State*Task + SessionOrder + (1|Subject) + (0+State|Subject) Model 2 (physiology): MetabolicCost ~ State + HRVcoherence + CriticalityMetric + (1|Subject) Model 3 (transition dynamics): TimeToCoherence ~ PerturbationType + BaselineCoherence + (1|Subject)

6. Predicted Outcomes & Decision Rules

• Primary: Coherence induction → superior integrative task quality (effect d ≥ 0.5) and reduced metabolic cost vs. Criticality induction.
• Secondary: Criticality induction → parity or modest advantage on algorithmic tasks at the expense of higher metabolic cost and subjective effort.
• Transition: Perturbation from coherence should lead to transient increases in criticality markers (avalanche exponents, DFA changes) before recovery; unsuccessful recovery correlates with prolonged critical markers and degraded integrative performance.

7. Potential Confounds & Mitigations

• Expectation/placebo: Use sham biofeedback or active control with neutral instructions; blind raters for integrative tasks.
• Respiratory artifacts: Record respiration and regress its influence from HRV and EEG signals.
• Artifact misclassification: Conservative artifact rejection; require avalanche findings to exceed surrogate thresholds.
• Interindividual differences: Pre‑specify covariates (baseline HRV, trait anxiety) and include in models; consider stratified analyses.

8. Glossary (Technical)

HRV coherence ratio

Measure of HRV spectral concentration at paced‑breathing frequency relative to broadband 0.04–0.4 Hz power, used to index cardiac coherence.

Neuronal avalanche

Spatiotemporal cascades of neural events with size distributions that may follow power laws when near criticality.

Detrended Fluctuation Analysis (DFA)

Technique to quantify long‑range temporal correlations in nonstationary signals; exponent α characterizes correlation structure.

Lyapunov exponent

Measure of trajectory divergence in reconstructed state space; largest positive exponent indicates sensitivity to initial conditions.

Phase‑locking value (PLV)

Measure of phase synchrony between two oscillatory signals (e.g., EEG band and cardiac phase).

9. OSF‑Ready Preregistration Protocol

Title: Hierarchical Optima: Testing Coherence vs. Criticality as Primary and Secondary System Optima

Background & Rationale

We propose that Heart‑Mind Coherence is the primary systemic optimum for integrative cognition and energy efficiency, while criticality functions as a secondary exploratory bridge.

Primary Hypotheses

1. Coherence induction will yield superior performance on integrative tasks relative to Criticality induction (two‑tailed).
2. Criticality induction will yield equal or superior performance on algorithmic tasks relative to Coherence induction (two‑tailed).
3. Criticality state will show higher metabolic cost (cortisol, HR) than Coherence state for matched or lower performance on integrative tasks.

Design

Within‑subjects 2×2 crossover: State (Coherence vs. Criticality) × Task (Algorithmic vs. Integrative). Each participant completes all 4 conditions with counterbalanced order.

Sample Size

N = 80 planned. Exclusions: participants with >20% artifact EEG epochs; physiological data loss >20% of session; self‑reported noncompliance. All exclusions to be logged with reasons.

Primary Outcomes

• Integrative task quality score (rater‑aggregated): primary outcome for Hypothesis 1.
• Algorithmic task accuracy & RT: primary outcomes for Hypothesis 2.
• Metabolic proxies (salivary cortisol delta, mean HR): primary outcome for Hypothesis 3.

Analysis Plan

• Primary tests: LME with fixed effects State, Task, and their interaction; random intercepts for subjects.
• Preplanned contrasts: Coherence_vs_Criticality on Integrative tasks; Criticality_vs_Coherence on Algorithmic tasks.
• Multiple comparisons: hierarchical testing; FDR for secondary outcomes.
• Sensitivity analyses: include baseline HRV tertile as covariate; perform per‑protocol and intention‑to‑treat analyses.

Data Sharing & Reproducibility

De‑identified raw physiological data, processed data, analysis scripts (Python/R/MATLAB), and stimulus code will be uploaded to OSF and GitHub within 12 months of data collection completion. Preprocessing and analysis pipelines will be containerized (Docker) and documented with example subset data for replication.

Ethics & Safety

Institutional IRB approval will be obtained. Inductions are low risk; protocols include monitoring criteria and immediate cessation for distress. Consent forms will explicitly describe psychophysiological recordings and saliva sampling.

Pre-registered by: Nathan A. Richards • Date: May 28, 2026

10. Annotated References (APA)

1. Beggs, J. M., & Plenz, D. (2003). Neuronal avalanches in neocortical circuits. Journal of Neuroscience, 23(35), 11167–11177. — Foundational empirical demonstration of avalanche dynamics in cortical circuits.
2. Maass, W., Natschläger, T., & Markram, H. (2002). Real‑time computing without stable states. Neural Computation, 14(11), 2531–2560. — Reservoir/liquid state computing framework illustrating computational advantages near dynamic boundaries.
3. Palva, J. M., et al. (2013). Neuronal long‑range temporal correlations and avalanche dynamics are correlated with behavioral state. Nature Communications, 4, 1–10.
4. McCraty, R., Atkinson, M., Rein, G., & Watkins, A. D. (2009). The HeartMath solution and cardiac coherence research. HeartMath Institute technical reports. — Source material for coherence induction scripts.
5. Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61(3), 201–216.
6. Clauset, A., Shalizi, C. R., & Newman, M. E. J. (2009). Power‑law distributions in empirical data. SIAM Review, 51(4), 661–703. — Methodological standard for fitting power‑law distributions via MLE.
7. Seth, A. K., et al. (various). Methods for estimating Lyapunov exponents and state‑space reconstruction in neural data.
8. MNE & EEGLAB documentation and tutorials. — Best practices for EEG preprocessing, ICA, and artifact handling.

Expository Footnotes

1. HeartMath Institute provides accessible operational coherence metrics and induction scripts; include these methods but rely on independent HRV literature for validation.
2. EEG‑based neuronal avalanche detection is sensitive to spatial resolution and filtering; robust claims require surrogate testing and, ideally, cross‑modal replication (MEG/multiunit where possible).
3. Power‑law fitting requires reporting xmin and alternative distribution tests (lognormal, stretched exponential); predefine acceptable model criteria in preregistration.

11. Appendices

A. PsychoPy TTL Snippet

from psychopy import core, visual # Parallel port TTL example win = visual.Window([1024,768]) stim = visual.TextStim(win, text='X', height=0.15) for t in range(120): stim.draw() win.flip() core.wait(0.1) core.wait(0.2)

B. Avalanche Fitting Pseudocode

# 1) Preprocess EEG, remove artifacts. # 2) High-pass to remove slow drift; bandpass optional. # 3) Threshold each channel (z > 3). # 4) Bin across dt = mean IEI (~ms); create binary raster. # 5) Identify contiguous bins with activity -> avalanches. # 6) Compute size/duration distributions; fit with Clauset MLE. # 7) Bootstrap to obtain CI and compare with surrogate data.

C. Informed Consent Excerpt

You will undergo EEG, ECG, and peripheral sensor recording, perform cognitive tasks, and provide saliva samples. All data will be de‑identified; you may withdraw at any time without penalty. Minimal risk: possible transient discomfort from electrodes and temporary mild stress during some task conditions. Data will be shared de‑identified on OSF and GitHub.

License: CC‑BY‑4.0 (attribution to Nathan A. Richards). Prepared for insertion into Star‑Field.App research proposals.