Efficient estimation of compressible state-space models with application to calcium signal deconvolution

In this paper, we consider linear state-space models with compressible innovations and convergent transition matrices in order to model spatiotemporally sparse transient events. We perform parameter and state estimation using a dynamic compressed sensing framework and develop an efficient solution consisting of two nested Expectation-Maximization (EM) algorithms. Under suitable sparsity assumptions on the innovations, we prove recovery guarantees and derive confidence bounds for the state estimates. We provide simulation studies as well as application to spike deconvolution from calcium imaging data which verify our theoretical results and show significant improvement over existing algorithms.

[1]  E.J. Candes,et al.  An Introduction To Compressive Sampling , 2008, IEEE Signal Processing Magazine.

[2]  G. Kitagawa.  A self-organizing state-space model , 1998 .

[3]  S. Frühwirth-Schnatter.  Data Augmentation and Dynamic Linear Models , 1994 .

[4]  Michael A. Saunders,et al.  Atomic Decomposition by Basis Pursuit , 1998, SIAM J. Sci. Comput..

[5]  Rafael Yuste,et al.  Fast nonnegative deconvolution for spike train inference from population calcium imaging. , 2009, Journal of neurophysiology.

[6]  Namrata Vaswani,et al.  Kalman filtered Compressed Sensing , 2008, 2008 15th IEEE International Conference on Image Processing.

[7]  C. Stosiek,et al.  In vivo two-photon calcium imaging of neuronal networks , 2003, Proceedings of the National Academy of Sciences of the United States of America.

[8]  Richard M. Leahy,et al.  MEG-based imaging of focal neuronal current sources , 1996, IEEE Transactions on Medical Imaging.

[9]  Jong Chul Ye,et al.  Motion estimated and compensated compressed sensing dynamic magnetic resonance imaging: What we can learn from video compression techniques , 2010, Int. J. Imaging Syst. Technol..

[10]  Ludwig Fahrmeir,et al.  State Space Models: a Brief History and Some Recent De- Velopments , 2022 .

[11]  Catie Chang,et al.  Time–frequency dynamics of resting-state brain connectivity measured with fMRI , 2010, NeuroImage.

[12]  Namrata Vaswani,et al.  Time Invariant Error Bounds for Modified-CS-Based Sparse Signal Sequence Recovery , 2015, IEEE Trans. Inf. Theory.

[13]  Martin J. Wainwright,et al.  A unified framework for high-dimensional analysis of $M$-estimators with decomposable regularizers , 2009, NIPS.

[14]  Pini Gurfil,et al.  Methods for Sparse Signal Recovery Using Kalman Filtering With Embedded Pseudo-Measurement Norms and Quasi-Norms , 2010, IEEE Transactions on Signal Processing.

[15]  Namrata Vaswani,et al.  Time invariant error bounds for modified-CS based sparse signal sequence recovery , 2013, 2013 IEEE International Symposium on Information Theory.

[16]  Gerhard Tutz,et al.  State space models , 1994 .

[17]  W. J. Nowack.  Neocortical Dynamics and Human EEG Rhythms , 1995, Neurology.

[18]  Namrata Vaswani,et al.  LS-CS-Residual (LS-CS): Compressive Sensing on Least Squares Residual , 2009, IEEE Transactions on Signal Processing.

[19]  Philip Schniter,et al.  Dynamic Compressive Sensing of Time-Varying Signals Via Approximate Message Passing , 2012, IEEE Transactions on Signal Processing.

[20]  Emery N. Brown,et al.  Exact and Stable Recovery of Sequences of Signals with Sparse Increments via Differential _1-Minimization , 2012, NIPS.

[21]  R. Shumway,et al.  AN APPROACH TO TIME SERIES SMOOTHING AND FORECASTING USING THE EM ALGORITHM , 1982 .

[22]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[23]  G. Peters,et al.  Monte Carlo Approximations for General State-Space Models , 1998 .

[24]  M. Pitt,et al.  Likelihood analysis of non-Gaussian measurement time series , 1997 .

[25]  L. Knorr‐Held.  Conditional Prior Proposals in Dynamic Models , 1999 .

[26]  S. Frühwirth-Schnatter.  Applied state space modelling of non-Gaussian time series using integration-based Kalman filtering , 1994 .

[27]  Le Song,et al.  Estimating time-varying networks , 2008, ISMB 2008.

[28]  R. Yuste,et al.  Detecting action potentials in neuronal populations with calcium imaging. , 1999, Methods.

[29]  Min Wu,et al.  Fast and Stable Signal Deconvolution via Compressible State-Space Models , 2016, bioRxiv.

[30]  Emery N. Brown,et al.  Convergence and Stability of Iteratively Re-weighted Least Squares Algorithms , 2014, IEEE Transactions on Signal Processing.