1: \begin{abstract}

2: This paper develops a point impact linear regression model in which

3: the trajectory of a continuous stochastic process, when evaluated at a

4: sensitive time point, is associated with a scalar response. The

5: proposed model complements and is more interpretable than the

6: functional linear regression approach that has become popular in recent

7: years. The trajectories are assumed to have fractal (self-similar)

8: properties in common with a fractional Brownian motion with an unknown

9: Hurst exponent. Bootstrap confidence intervals based on the

10: least-squares estimator of the sensitive time point are developed.

11: Misspecification of the point impact model by a functional linear model

12: is also investigated. Non-Gaussian limit distributions and rates of

13: convergence determined by the Hurst exponent play an important role.

14: \end{abstract}