REVIEW 3 major objections 3 minor 25 references
A new integral equation for the first passage time density of the Ornstein-Uhlenbeck process
T0 review · 3 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The paper derives a Volterra integral equation for the first-passage-time density of the Ornstein–Uhlenbeck process, with a parabolic-cylinder-function kernel, that reduces to the Fortet renewal equation at q=0 and covers time-dependent…
desk verdict The claimed new Volterra integral equation for OU first passage densities is invalid: equation (14) does not follow from (13) because the exponential terms are algebraically inconsistent, so the central result fails. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The mechanism is a Laplace-inversion identity for products of two parabolic cylinder functions (equation (9)), combined with the convolution theorem and the time-scaling property of the Laplace transform. The first-passage Laplace transform (a ratio of two parabolic cylinders) is multiplied by a chosen parabolic-cylinder function whose inversion is known; rearranging the resulting convolution yields the Volterra equation (14)/(17).
What would settle it
Evaluate the ratio of the exponential factor inside the inversion integral in (13) to the exponential kernel of (14) at generic parameter values (for instance x0=0, S=1, x=2, θ=1, σ=1, t=1, τ=0.5); if the ratio is not identically 1, the simplification claimed between those equations is false. Equivalently, compute both sides of (17) numerically for the exponential-threshold density (16) at q=1 and a non-degenerate parameter set and compare the discrepancy.
Extended reading notes
Core claim
The central claim is that equations (14) and (17) form a Volterra integral equation of the first kind for the first-passage-time density g(S(t), t | x0) of the Ornstein–Uhlenbeck process. The kernel is built from exponentials and a parabolic cylinder function D_q, with the order q available as a free parameter. For q = 0 the equation reduces exactly to the Fortet renewal equation. The paper also claims that the kernel is regular for q ≤ −1 at x = S(t), and that for x = S(t), q = 1 the equation is of the second kind and directly yields the known closed-form first-passage density for the constant threshold S = µθ, avoiding the usual transformation argument.
Load-bearing premise
The derivation of equation (14) from (13) depends on an algebraic simplification of the exponential prefactors that must hold as an identity over all parameter values; if that equality is not exact, the stated Laplace-convolution argument does not produce the new integral equation.
Editorial extensions
If this is right
- For q ≤ −1, the new equation's kernel is regular at x = S(t), so numerical evaluation may avoid the weakly singular kernel of the Fortet renewal equation.
- At q = 0 the equation collapses to the classical Fortet renewal equation, making the Fortet equation a member of a one-parameter family.
- For x = S = µθ and q = 1, the equation reduces to a Volterra equation of the second kind whose solution reproduces the closed-form density (19) without the usual transformation argument.
- The same integral equation applies to time-dependent thresholds; the paper reports numerical agreement for the exponential threshold family (15) across values of q.
Reading between the lines
- If the algebraic step between (13) and (14) does not hold, the integral equation itself could still be true; a direct proof by differentiation or renewal arguments, independent of Laplace inversion, would settle whether equation (17) stands.
- The free parameter q suggests a family of integral equations; different choices of q might be equivalent by direct transformation of the transition density, which would give a way to tune the kernel's singularity profile for numerical work.
- A numerical benchmark comparing the claimed regular kernel (q ≤ −1) against existing regularized forms of the Fortet equation on the exponential-threshold example would reveal whether the regularity translates into better accuracy or conditioning.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a new Volterra integral equation for the first-passage-time density of the Ornstein--Uhlenbeck process. The starting point is the Laplace transform of the first-passage-time density for a constant threshold, which contains a ratio of parabolic cylinder functions. Using inverse Laplace transforms for products of parabolic cylinder functions previously derived by the author, the paper claims to invert the product of this Laplace transform with a suitably chosen factor and thereby to obtain a new integral equation whose kernel contains a parabolic cylinder function. The Fortet renewal equation is recovered when the parabolic cylinder order q is zero, and a Volterra equation of the second kind is obtained for x=S and q=1. The paper further claims that the equation extends to time-dependent thresholds, supported by a numerical check for one exponential threshold, and that the kernel is regular for q≤−1. Section 4 uses the q=1 case to retrieve a known closed-form density for the threshold S=μθ.
Significance. If the proposed integral equation were correct, it would give a new representation of first-passage-time densities for the Ornstein--Uhlenbeck process and could be of numerical interest, especially because the Fortet equation appears as a special case. The paper also aims to unify several known results through a single kernel containing parabolic cylinder functions. However, the central equation is not consistent with the paper's own known closed-form density: at q=1 and x=S=μθ, Eq. (14) yields a result larger by a factor of √2 than the standard formula (19). Since this is a numerical-algebraic contradiction at a parameter point explicitly treated in the paper, the main claim cannot be accepted. Credit is due for the q=0 reduction, which is internally consistent, and for the attempt to connect the inverse Laplace transform literature with first-passage-time equations, but these strengths do not compensate for the incorrect central equation.
major comments (3)
- [Section 3, Eq. (14)] The reader's stated objection to the step from (13) to (14) -- that the exponential changes from exp(-((μθ−x0)+(x−μθ)e^{-t/θ})^2/(2σ²θ(...))) to exp(-((x−μθ)−(x0−μθ)e^{-t/θ})^2/(2σ²θ(...))) and that a factor is therefore dropped -- does not by itself invalidate the derivation. When the kernel is rewritten with the threshold appearing in both the exponential and the D_q argument, the relation F_u(t)=(1/θ)exp(((x−μθ)^2−(u−μθ)^2)/(2σ²θ))K_u(t) holds, so the prefactor in (13) cancels against the same prefactor generated on the left-hand side. The equation (14) is nevertheless wrong. Consider q=1 and x=S=μθ. Then the left-hand kernel of (14) vanishes because D_1(0)=0, and (14) reduces to g(μθ,t|x0)=(2/(θ√π))K_{x0}(t), where K_{x0}(t) is the right-hand kernel. Using D_1(z)=z exp(−z²/4), this gives g(μθ,t|x0)=2√2(μθ−x0)/(σ θ^{3/2}√π) e^{-t/θ}(1−e^{-2t/θ})^{-3/2} exp(−(x0−μθ)^2 e^{-2t/θ}/(σ²θ(1−e^{-2t/θ}))). This is √2 times the known formula (19) quoted in the paper. Since (19) is standard and is also the paper's own target, Eq. (14) cannot be correct.
- [Section 3, Eqs. (15)-(17)] The extension from constant to time-dependent thresholds is not derived. After specializing to q=0, the paper notes that the Fortet equation holds for S(t) and then asserts that the same extension holds for q≠0, supported only by the statement that 'numerical evaluation reveals' equality for the exponential threshold (15). No parameter values, quadrature method, error tolerance, or code are provided, so the verification is not reproducible; moreover, because Eq. (14) is false, this numerical check cannot provide evidence for Eq. (17). A proof, or at minimum a fully specified and reproducible numerical experiment, is required before Eq. (17) can be accepted.
- [Section 4, Eq. (20)] The printed Eq. (20) does not follow from Eq. (14). Carrying out the q=1, x=S inversion in (14) gives a factor √2 in both the inhomogeneous term and the integral kernel, whereas Eq. (20) as printed has coefficient 2 and agrees with the known result (19). Thus Eqs. (14) and (20) are mutually inconsistent. The paper should identify which of the two is intended and correct the derivation accordingly; as presented, the contradiction is load-bearing because Section 4 claims to retrieve (19) from the new integral equation.
minor comments (3)
- [Section 4 heading] The section heading contains a repeated-text typo: 'S = µθS = µθS = µθ' should read 'S = µθ'.
- [Section 3, numerical paragraph] The numerical verification of the time-dependent threshold should state the parameter values, the quadrature or discretization scheme, and the error tolerance; the current one-line statement is not reproducible.
- [Section 3, between Eqs. (13) and (14)] The phrase 'after some straightforward simplifications' hides a nontrivial normalization step in which the kernel is rewritten so that the threshold appears symmetrically in the exponential and in the D_q argument. Showing this intermediate step would make the algebra checkable and might have revealed the normalization error.
Circularity Check
No significant circularity: the derivation applies independent inverse-Laplace identities to the known Laplace transform of the first passage time density, rather than restating its own inputs.
full rationale
The paper's derivation chain begins with inverse Laplace transforms for products of parabolic cylinder functions quoted from the author's earlier papers [14] and [15]. These are self-citations, but they are parameter-free mathematical identities about parabolic cylinder functions and do not already contain the first passage time renewal equation or the target integral equation; they are independent inputs. The paper then combines these transforms with the Laplace transform of the Ornstein-Uhlenbeck first passage time density, given in Eq. (2), using the convolution theorem to obtain Eq. (13), and rearranges to the claimed integral equation Eq. (14). No step defines the first passage time density in terms of the claimed integral equation, and no fitted parameter is renamed as a prediction. The q=0 reduction to the Fortet equation is an external consistency check, not an input used in the derivation. The time-dependent threshold claim in Eq. (17) is supported numerically by inserting a known closed-form first passage time density; this is a weak verification rather than circular reasoning. There is a separate algebraic correctness concern in the simplification from Eq. (13) to Eq. (14), where the exponential factors do not obviously coincide for general x and S, but that is a potential derivation error, not a circularity: the claimed equation is not equivalent to its own inputs by construction. Overall, the central derivation does not reduce to its own assumptions, so no circularity is found.
Assumptions & free parameters
assumptions (2)
- standard math Inverse Laplace transform identity (9) from [15] and [14] for products of parabolic cylinder functions, including the special term for y+z=0, c+q=1.
- standard math Standard properties of parabolic cylinder functions, such as D_0(z)=exp(-z^2/4), D_1(z)=z exp(-z^2/4), and the limit D_nu(z)->0 as z->infinity, are used without proof.
Cite this review
Pith. "Pith review of A new integral equation for the first passage time density of the Ornstein-Uhlenbeck process." pith.science (2026). https://pith.science/paper/UPKLZQC3
@misc{pith2026190802071,
author = {Pith},
title = {Pith review of: A new integral equation for the first passage time density of the Ornstein-Uhlenbeck process},
year = {2026},
howpublished = {\url{https://pith.science/paper/UPKLZQC3}},
note = {Machine review of arXiv:1908.02071}
}
read the original abstract
The Laplace transform of the first passage time density of the Ornstein--Uhlenbeck process for a constant threshold contains a ratio of two parabolic cylinder functions for which no analytical inversion formula is available. Recently derived inverse Laplace transforms for the product of two parabolic cylinder functions together with the convolution theorem of the Laplace transform then allow to derive a new Volterra integral equation for this first passage time density. The kernel of this integral equation contains a parabolic cylinder function and the Fortet renewal equation for the Ornstein-Uhlenbeck process emerges as a special case, namely when the order q of the parabolic cylinder function is set at 0. The integral equation is shown to hold both for constant as well as time dependent thresholds. Moreover, the kernel of the integral equation is regular for q<=-1.
Reference graph
Works this paper leans on
-
[1]
Gluss, B. (1967). A Model for Neuron Firing with Exponential Decay of Potential Resulting in Diffusion Equations for Probability Density. B. Math. Biophys. 29, 233–243
work page 1967
-
[2]
Sacerdote, L., Giraudo, M. T. (2013). Stochastic Integrate a nd Fire Models: A Review on Mathemat- ical Methods and Their Applications. In: Bachar, M., Batzel, J. J., Dit levsen, S. (eds.). Stochastic Biomathematical Models (Lecture Notes in Mathematics 2058). Springer–Verlag, Berlin
work page 2013
-
[3]
Smith, P. L., Ratcliff, R. (2004). Psychology and neurobiology of s imple decisions. TRENDS Neurosci. 27, 161–168
work page 2004
-
[4]
Ditlevsen, P. D., Kristensen, M. S., Andersen, K. K. (2005). The Recurrence Time of Dansgaard– Oeschger Events and Limits on the Possible Periodic Component. J. Climate 18, 2594–2603
work page 2005
-
[5]
Kurzy´ nski, M., Che/suppress lminiak, P. (2003). Mean First-Passage Timein the Stochastic Theory of Biochem- ical Processes. Application to Actomyosin Molecular Motor. J. Stat. Phys. 110, 137–181
work page 2003
-
[6]
G¨ oing-Jaeschke, A., Yor, M. (2003). A clarification note about hitting times densities for Ornstein– Uhlenbeck processes. Finance Stoch., 7, 413–415
work page 2003
-
[7]
Alili, L., Patie, P., Pedersen, J. L. (2005). Representations of th e First Hitting Time Density of an Ornstein–Uhlenbeck Process. Stoch. Models 21, 967–980
work page 2005
-
[8]
Fortet, R. (1943). Les fonctions al´ eatoires du type de Markoff associ´ ees ` a certaines ´ equations lin´ eaires aux d´ eriv´ ees partielles du type parabolique.Journ. de Math. 22, 177– 243
work page 1943
Show all 25 references
-
[9]
Ditlevsen, S., Ditlevsen, O. (2008). Parameter estimation from o bservations of first-passage times of the Ornstein–Uhlenbeck process and the Feller process. Probabilist. Eng. Mech. 23, 170–179
2008
-
[10]
L´ ansk´ y, P., Ditlevsen, S. (2008). A review of the methods for signal estimation in stochastic diffusion leaky integrate-and-fire neuronal models. Biol. Cybern. 99, 253–262
2008
-
[11]
M., Sacerdote, L., Sato, S
Ricciardi, L. M., Sacerdote, L., Sato, S. (1984). On an Integra l Equation for First-Passage-Time Prob- ability Densities. J. Appl. Probab. 21, 302–314
1984
-
[12]
G., Ricciardi, L
Buonocore, A., Nobile, A. G., Ricciardi, L. M. (1987). A New Integ ral Equation for the Evaluation of First-Passage-Time Probability Densities. Adv. Appl. Probab. 19, 784–800
1987
-
[13]
G., Ricciardi, L
Giorno, V., Nobile, A. G., Ricciardi, L. M., Sato, S. (1989). On the E valuation of First-Passage-Time Probability Densities via Non-Singular Integral Equations. Adv. Appl. Probab. 21, 20–36. 16
1989
-
[14]
Veestraeten, D. (2015). On the inverse transform of Laplac e transforms that contain (products of) the parabolic cylinder function. Integral Transforms Spec. Funct. 26, 859–871
2015
-
[15]
Veestraeten, D. (2017). An alternative integral represent ation for the product of two parabolic cylinder functions. Integral Transforms and Spec. Funct. 28, 915–922
2017
-
[16]
K., Smith, D
Roy, B. K., Smith, D. R. (1969). Analysis of the Exponential Dec ay Model of the Neuron Showing Frequency Threshold Effects. B. Math. Biophys. 31, 341–357
1969
-
[17]
P., Perkel, D
Sugiyama, H., Moore G. P., Perkel, D. H. (1970). Solutions for a S tochastic Model of Neuronal Spike Production. Math. Biosci. 8, 323–341
1970
-
[18]
M., Ricciardi, L
Capocelli, R. M., Ricciardi, L. M. (1971). Diffusion Approximation an d First Passage Time Problem for a Model Neuron. Kybernetik 8, 214–223
1971
-
[19]
Ricciardi, L. M. (1977). Diffusion Processes and Related Topics in Biology (Lecture Notes in Biomath- ematics 14). Springer–Verlag, Berlin
1977
-
[20]
Veestraeten, D. (2014). On Transition and First Hitting Time De nsities and Moments of the Ornstein– Uhlenbeck Process. Stoch. Models 30, 143–161
2014
-
[21]
Oldham, K., Myland, J., Panier, J. (2009). An Atlas of Functions. Second Edition, Springer–Verlag, Berlin
2009
-
[22]
Debnath, L., Bhatta, D. (2015). Integral Transforms and T heir Applications. Third Edition, CRC Press, Boca Raton
2015
-
[23]
Abramowitz, M., Stegun, I. A. (1972). Handbook of Mathemat ical Functions with Formulas, Graphs, and Mathematical Tables. Dover Publications, New York
1972
-
[24]
Doob, J. L. (1942). The Brownian Movement and Stochastic Eq uations. Ann. Math. 43, 351–369
1942
-
[25]
Magnus, W., Oberhettinger, F., Soni, R. P. (1966). Formulas an d Theorems for the Special Functions of Mathematical Physics (Die Grundlehren der mathematischen Wiss enschaften 52). Third Edition, Springer–Verlag, Berlin. Amsterdam School of Economics, University of Amsterdam, ...
1966
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.