关于「双相剂量窗:Arndt–Schulz 曲线的数学形式」,《光剂量双相窗与热边界》给出的定量结论是什么?
结论:第一章写出乘积型双相剂量函数,峰位有闭式解;在按已发表综述定性次序构造的示意读数上,单调模型的残差平方和是双相模型的 4.3 倍,峰值在 0.537 J/cm²,照到峰值时间的 30 倍,模型中高出对照的部分只剩 15.6 %。
数据口径:书中只写物理量(辐照度、能量密度、温升、占空比)与已发表研究报告的剂量范围,所有设备与参数均为示意,全部计算由书中程序生成。 本书不是医疗建议,不宣称任何功效;任何自我试验须先咨询执业医师。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.537, 4.3
| biphasic : A 2.107 Ka 0.1905 Ki 1.514 SSE 0.0681 R2 0.900 | |
| peak D* 0.537 J/cm^2, peak R 2.148 | |
| half-max window 0.0764 to 3.78 J/cm^2 (ratio 49) | |
| monotone : A 0.684 K 0.0347 n 4.0 SSE 0.2922 R2 0.570 | |
| SSE ratio monotone / biphasic | 4.3 |
| dose read biphasic monotone | |
| 0.003 | 1.02 1.033 1.000 |
| 0.030 | 1.25 1.281 1.246 |
| 0.300 | 2.10 2.075 1.684 |
| 3.000 | 1.60 1.664 1.684 |
| 30.000 | 1.35 1.101 1.684 |
核心公式
\[R(D) = 1 + A\,g(D),\qquad g(D) = \frac{D\,K_i}{(D+K_a)(D+K_i)}\]
查看完整验算程序
# Chapter 1. p12: biphasic versus monotone fits to five dose levels.
# Dose levels 0.003, 0.03, 0.3, 3 and 30 J/cm^2 are those of the 810 nm
# fibroblast reporter experiment summarised by Huang et al. (2009).
# The readings below are ILLUSTRATIVE numbers built only to follow the
# qualitative order the review reports (no rise, small rise, maximum,
# lower, still above the 0.03 level); they are not the paper's values.
# Models, fitted by least squares on a log grid (A in closed form):
# biphasic R = 1 + A * D*Ki / ((D + Ka) * (D + Ki)) (as in p11)
# monotone R = 1 + A * D**n / (K**n + D**n) (saturating)
import math
DOSE = [0.003, 0.03, 0.3, 3.0, 30.0]
READ = [1.02, 1.25, 2.10, 1.60, 1.35]
Y = [r - 1.0 for r in READ]
def fit_shape(shape):
# for a fixed shape f(D), least squares gives A = <f,y> / <f,f>
f = [shape(d) for d in DOSE]
a = math.fsum(fi * yi for fi, yi in zip(f, Y)) / \
math.fsum(fi * fi for fi in f)
sse = math.fsum((yi - a * fi) ** 2 for fi, yi in zip(f, Y))
return sse, a
best_b = None
for i in range(-250, 51):
ka = 10.0 ** (i / 50.0)
for j in range(-150, 151):
ki = 10.0 ** (j / 50.0)
if ki <= ka:
continue
sse, a = fit_shape(lambda d: d * ki / ((d + ka) * (d + ki)))
if best_b is None or sse < best_b[0]:
best_b = (sse, a, ka, ki)
best_m = None
for i in range(-250, 151):
k = 10.0 ** (i / 50.0)
for n10 in range(2, 41):
n = n10 / 10.0
sse, a = fit_shape(lambda d: d ** n / (k ** n + d ** n))
if best_m is None or sse < best_m[0]:
best_m = (sse, a, k, n)
sse, a, ka, ki = best_b
dstar = math.sqrt(ka * ki)
gmax = ki / (math.sqrt(ka) + math.sqrt(ki)) ** 2
c = gmax / 2.0
b = c * (ka + ki) - ki
r = math.sqrt(b * b - 4.0 * c * c * ka * ki)
lo, hi = (-b - r) / (2.0 * c), (-b + r) / (2.0 * c)
sst = math.fsum((y - math.fsum(Y) / len(Y)) ** 2 for y in Y)
print("biphasic : A %.3f Ka %.4f Ki %.3f SSE %.4f R2 %.3f"
% (a, ka, ki, sse, 1.0 - sse / sst))
print(" peak D* %.3f J/cm^2, peak R %.3f"
% (dstar, 1.0 + a * gmax))
print(" half-max window %.4f to %.2f J/cm^2 (ratio %.0f)"
% (lo, hi, hi / lo))
sse_m, a_m, k_m, n_m = best_m
print("monotone : A %.3f K %.4f n %.1f SSE %.4f R2 %.3f"
% (a_m, k_m, n_m, sse_m, 1.0 - sse_m / sst))
print("SSE ratio monotone / biphasic: %.1f" % (sse_m / sse))
print("dose read biphasic monotone")
for d, rd in zip(DOSE, READ):
fb = 1.0 + a * d * ki / ((d + ka) * (d + ki))
fm = 1.0 + a_m * d ** n_m / (k_m ** n_m + d ** n_m)
print("%7.3f %5.2f %8.3f %8.3f" % (d, rd, fb, fm))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《光生物调节双相剂量窗、阿伦尼乌斯热边界与脉冲调制照射时长母典》第 1 章「双相剂量窗:Arndt–Schulz 曲线的数学形式」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-287 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026287,
author = {Fan, Yafu},
title = {The Biphasic Dose Window of Photobiomodulation: Arrhenius Thermal Limits and Exposure Times under Pulsed Modulation},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}