1: \begin{abstract}

2: In this paper, the key objects of interest are the sequential covariance matrices $\mathbf{S}_{n,t}$ and their largest eigenvalues. Here, the matrix $\bfS_{n,t}$ is computed as the empirical covariance associated with observations $\{\bfx_1,\ldots,\bfx_{ \lfloor nt \rfloor } \}$,  

3: for $t\in [0,1]$. The observations $\bfx_1,\ldots,\bfx_n$ are assumed to be i.i.d. $p$-dimensional vectors with zero mean,

4: and a covariance matrix that is a fixed-rank perturbation of the identity matrix.

5: Treating $\{

6: \mathbf{S}_{n,t}\}_{t \in [0,1]}$ as a matrix-valued stochastic process indexed by $t$, we study 

7: the behavior of the largest eigenvalues of $\mathbf{S}_{n,t}$, as 

8: $t$ varies, with $n$ and $p$ increasing simultaneously, so 

9: that $p/n \to y \in (0,1)$. As a key contribution of this work, we establish the weak convergence of the stochastic process corresponding to the sample spiked eigenvalues, if their population counterparts exceed the critical phase-transition threshold. Our analysis of the limiting process is fully comprehensive revealing, in general, non-Gaussian limiting processes.

10: 

11: As an application, we consider a class of change-point problems, where the interest is in detecting structural breaks in the covariance caused by a change in magnitude of the spiked eigenvalues. For this purpose, we propose two different maximal statistics corresponding to centered spiked eigenvalues

12: of the sequential covariances. We show the existence of limiting null distributions for these statistics, and prove consistency 

13: of the test under fixed alternatives. Moreover, we compare the behavior of the proposed tests through a simulation study.

14: \end{abstract}

15: