| name | psychosis-scaling-critical-regime |
| description | 精神病早期阶段脑动力学临界性scaling偏差研究方法论。结合重整化群(RG)框架与多种scaling分析方法,揭示临界 regime内的动力学重组而非临界性丧失。 |
| platforms | ["linux","macos","windows"] |
| tags | ["neuroscience","criticality","psychosis","fMRI","renormalization-group","scaling-analysis","brain-dynamics"] |
| category | neuroscience |
Early Psychosis Scaling Behaviour in Critical Regime
Paper: arXiv:2606.06290v1 - "Early psychosis shows deviations in scaling behaviour within a critical regime"
Authors: Irem Topal, Paola Moreno Ancalmo et al.
Published: 2026-06-04
核心发现
精神病早期阶段不是简单的临界性动力学丧失,而是在保持的 scaling regime 内的系统性重组。通过 phenomenological renormalization group (PRG) 框架结合 PSD 和 DFA 分析,揭示:
- 健康对照组:静息态活动展现与临界组织一致的非平凡 scaling 行为
- 早期精神病:保持相同的 scale-invariant 组织总体现象学,但多个可观测量上有系统性 scaling exponent 偏移
- 关键结论:早期精神病特征化的是集体动力学在保持的 scaling regime 内重组,而非临界性简单丧失
方法论框架
1. Phenomenological Renormalization Group (PRG)
PRG 是一种 coarse-graining 方法,用于研究跨尺度的集体动力学:
def prg_coarse_graining(data, scale_factor):
"""
Apply phenomenological renormalization group coarse-graining
Parameters:
- data: fMRI time series
- scale_factor: spatial/temporal coarse-graining factor
Returns:
- coarse_grained_data: renormalized data preserving critical structure
"""
pass
2. Power Spectral Density (PSD) Analysis
检测 1/f scaling 特征:
def power_spectral_analysis(fmri_signal):
"""
Compute PSD and estimate scaling exponent
PSD(f) ~ f^(-β) for critical dynamics
- β ≈ 1-2: near-critical regime
- β deviations indicate altered collective dynamics
"""
pass
3. Detrended Fluctuation Analysis (DFA)
量化时间序列的自相似性:
def detrended_fluctuation_analysis(signal, window_sizes):
"""
DFA for quantifying temporal scaling
F(n) ~ n^α
- α ≈ 0.5: uncorrelated (white noise)
- α ≈ 1: 1/f noise (critical)
- α > 1: non-stationary
- α < 0.5: anti-correlated
Returns fluctuation scaling exponent α
"""
pass
4. Combined PRG + Scaling Analysis Workflow
def combined_scaling_analysis(fmri_data, subject_groups):
"""
Full workflow combining PRG with temporal scaling analyses
Steps:
1. Apply PRG coarse-graining at multiple scales
2. Compute PSD at each scale
3. Compute DFA at each scale
4. Track scaling exponent evolution across scales
5. Compare exponent trajectories between groups
"""
results = {}
for scale in [1, 2, 4, 8, 16]:
coarse_data = prg_coarse_graining(fmri_data, scale)
psd_exp = power_spectral_analysis(coarse_data)
dfa_exp = detrended_fluctuation_analysis(coarse_data)
results[scale] = {
'psd_beta': psd_exp,
'dfa_alpha': dfa_exp
}
return results
神经科学意义
临界性理论背景
脑网络临界性假说认为大脑在 near-critical regime 运行,支持:
- 长程相关性 (long-range correlations)
- 高效信息处理 (efficient information processing)
- 集体组织涌现 (emergence of collective organization)
精神病中的临界性改变
传统观点认为精神疾病是临界性丧失,本研究揭示更 nuanced 的现象:
- Scaling regime 保持:整体 scale-invariant 现象学未丧失
- Exponent 偏移:scaling exponent 系统性改变指示动力学重组
- 跨尺度一致性:多个可观测量展示一致的偏移模式
临床应用潜力
1. 早期精神病诊断标志物
def psychosis_scaling_marker(fmri_data, reference_controls):
"""
Compute scaling-based biomarker for early psychosis
Returns:
- deviation_score: quantification of scaling deviation
- confidence: statistical significance
"""
subject_exponents = combined_scaling_analysis(fmri_data)
control_distribution = compute_control_exponents(reference_controls)
deviation = compute_multivariate_deviation(subject_exponents, control_distribution)
return deviation
2. 动力学重组量化
系统性偏移而非临界性丧失为干预策略提供新视角:
- 动力学调节:调整网络动力学回到健康 exponent 范围
- 尺度特定干预:针对特定 coarse-graining scale 的偏移
实现细节
数据要求
- fMRI 数据:静息态 BOLD 信号
- 时间分辨率:TR ≈ 2-3 秒
- 空间分辨率:ROI 或 voxel-level 分析
- 扫描时长:建议 > 10 分钟以捕获长期 scaling
统计分析
def statistical_comparison(group_A, group_B, exponents):
"""
Compare scaling exponents between groups
Statistical tests:
- Mann-Whitney U for non-parametric comparison
- Permutation tests for robust inference
- Effect size: Cohen's d
"""
from scipy.stats import mannwhitneyu
for exp_name in exponents:
a_values = [combined_scaling_analysis(s)[exp_name] for s in group_A]
b_values = [combined_scaling_analysis(s)[exp_name] for s in group_B]
stat, p = mannwhitneyu(a_values, b_values)
effect_size = compute_cohens_d(a_values, b_values)
print(f"{exp_name}: p={p:.4f}, d={effect_size:.2f}")
理论框架扩展
重整化群在神经科学的应用
RG 方法源于统计物理,用于研究相变和临界现象:
- 空间 RG:coarse-graining 空间区域,保留临界结构
- 时间 RG:积分时间窗口,研究动力学跨尺度行为
- 脑网络 RG:研究从微观神经元到宏观脑区的动力学传播
Scaling Universality
临界系统的 scaling exponent 具有 universality:
- 不同系统(物理、生物)可能共享相同 exponent
- Exponent 偏移指示动力学 regime 改变而非简单噪声增加
Pitfalls and Solutions
Pitfall 1: fMRI 时间序列非平稳性
问题:fMRI 信号包含缓慢漂移,影响 DFA 分析
解决:
def robust_dfa(signal, window_sizes, detrending='linear'):
if detrending == 'linear':
signal = linear_detrend(signal)
Pitfall 2: Scaling 拟合区间选择
问题:Scaling exponent 估计依赖于拟合区间选择
解决:
- 使用多个拟合区间验证 exponent 稳定性
- 报告 exponent 不确定性估计
- 使用 robust regression 方法
Pitfall 3: 样本量限制
问题:精神疾病研究通常样本量较小
解决:
- 使用 permutation tests
- Bootstrap for confidence intervals
- Combine multiple scaling measures for robust inference
Future Directions
- 多模态整合:结合 EEG、MEG 的 scaling 分析
- 纵向研究:追踪 scaling exponent 沿疾病进展的演化
- 干预效果:评估药物/治疗对 scaling 行为的影响
- 机器学习分类:使用 scaling features 作为 psychosis 预测特征
Activation
触发词:psychosis, critical dynamics, scaling analysis, renormalization group, fMRI, brain criticality, psychosis scaling, DFA, PSD, neuroimaging marker
References
- arXiv:2606.06290v1 - Primary paper
- Beggs & Plenz (2003) - Neuronal avalanches and criticality
- Linkenkaer-Hansen et al. (2001) - Long-range temporal correlations in brain oscillations
- Fraiman & Chialvo (2012) - fMRI scaling and brain criticality