Abstract
A number of important data analysis problems in neuroscience can be solved using state-space models. In this article, we describe fast methods for computing the exact maximum a posteriori (MAP) path of the hidden state variable in these models, given spike train observations. If the state transition density is log-concave and the observation model satisfies certain standard assumptions, then the optimization problem is strictly concave and can be solved rapidly with Newton–Raphson methods, because the Hessian of the loglikelihood is block tridiagonal. We can further exploit this block-tridiagonal structure to develop efficient parameter estimation methods for these models. We describe applications of this approach to neural decoding problems, with a focus on the classic integrate-and-fire model as a key example.
Similar content being viewed by others
References
Abarbanel, H., Creveling, D., & Jeanne, J. (2008). Estimation of parameters in nonlinear systems using balanced synchronization. Physical Review. D., 77, 016,208.
Ahmadian, Y., Pillow, J., & Paninski, L. (2009). Efficient Markov chain Monte Carlo methods for decoding population spike trains. Neural Computation (in press).
Ahmed, N. U. (1998). Linear and nonlinear filtering for scientists and engineers. Singapore: World Scientific.
Asif, A., & Moura, J. (2005). Block matrices with l-block banded inverse: Inversion algorithms. IEEE Transactions on Signal Processing, 53, 630–642.
Badel, L., Richardson, M., & Gerstner, W. (2005). Dependence of the spike-triggered average voltage on membrane response properties. Neurocomputing, 69, 1062–1065.
Bell, B. M. (1994). The iterated Kalman smoother as Gauss-Newton method. SIAM Journal on Optimization, 4, 626–636.
Boyd, S., & Vandenberghe, L. (2004). Convex optimization. Cambridge: Cambridge University Press.
Brockwell, A. E., Rojas, A. L., & Kass, R. E. (2004). Recursive Bayesian decoding of motor cortical signals by particle filtering. Journal of Neurophysiology, 91, 1899–1907.
Brown, E. N., Frank, L. M., Tang, D., Quirk, M. C., & Wilson, M. A. (1998). A statistical paradigm for neural spike train decoding applied to position prediction from ensemble firing patterns of rat hippocampal place cells. Journal of Neuroscience, 18, 7411–7425.
Davis, R. A., & Rodriguez-Yam, G. (2005). Estimation for state-space models based on a likelihood approximation. Statistica Sinica, 15, 381–406.
Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society. Series B, 79, 1–38.
Eden, U. T., Frank, L. M., Barbieri, R., Solo, V., & Brown, E. N. (2004). Dynamic analyses of neural encoding by point process adaptive filtering. Neural Computation, 16, 971–998.
Fahrmeir, L., & Kaufmann, H. (1991). On Kalman filtering, posterior mode estimation and fisher scoring in dynamic exponential family regression. Metrika, 38, 37–60.
Fahrmeir, L., & Tutz, G. (1994). Multivariate statistical modelling based on generalized linear models. New York: Springer.
Heskes, T., & Zoeter, O. (2002). Expectation propagation for approximate inference in dynamic Bayesian networks. In A. Darwiche & N. Friedman (Eds.), Uncertainty in artificial intelligence: Proceedings of the eighteenth conference (UAI-2002) (pp. 216–233). San Francisco: Morgan Kaufmann.
Huys, Q., Ahrens, M., & Paninski, L. (2006). Efficient estimation of detailed single-neuron models. Journal of Neurophysiology, 96, 872–890.
Izhikevich, E. M. (2007). Dynamical systems in neuroscience: The geometry of excitability and bursting. Cambridge: MIT.
Jungbacker, B., & Koopman, S. J. (2007). Monte Carlo estimation for nonlinear non-Gaussian state-space models. Biometrika, 94, 827–839.
Koyama, S., Shimokawa, T., & Shinomoto, S. (2007). Phase transitions in the estimation of event rate: A path integral analysis. Journal of Physics. A, Mathematical and General, 40, F383–F390.
Koyama, S., & Shinomoto, S. (2005). Empirical Bayes interpretations for random point events. Journal of Physics. A, Mathematical and General, 38, L531–L537.
Minka, T. (2001). Expectation propagation for approximate Bayesian inference. Uncertainty in Artificial intelligence, 17.
Moehlis, J., Shea-Brown, E., & Rabitz, H. (2006). Optimal inputs for phase models of spiking neurons. ASME Journal of Computational and Nonlinear Dynamics, 1, 358–367.
Olsson, R. K., Petersen, K. B., & Lehn-Schioler, T. (2007). State-space models: From the EM algorithm to a gradient approach. Neural Computation, 19, 1097–1111.
Paninski, L. (2004). Maximum likelihood estimation of cascade point-process neural encoding models. Network: Computation in Neural Systems, 15, 243–262.
Paninski, L. (2005). Log-concavity results on Gaussian process methods for supervised and unsupervised learning. Advances in Neural Information Processing Systems, 17, 1025–1032.
Paninski, L. (2006a). The most likely voltage path and large deviations approximations for integrate-and-fire neurons. Journal of Computational Neuroscience, 21, 71–87.
Paninski, L. (2006b). The spike-triggered average of the integrate-and-fire cell driven by Gaussian white noise. Neural Computation, 18, 2592–2616.
Paninski, L., Pillow, J., & Simoncelli, E. (2004). Maximum likelihood estimation of a stochastic integrate-and-fire neural model. Neural Computation, 16, 2533–2561.
Paninski, L., Brown, E. N., Iyengar, S., & Kass, R. E. (2008). Stochastic methods in neuroscience, chap. Statistical analysis of neuronal data via integrate-and-fire models. Oxford: Oxford University Press.
Pillow, J., Ahmadian, Y., & Paninski, L. (2009). Model-based decoding, information estimation, and change-point detection in multi-neuron spike trains. Neural Computation (in press).
Press, W. H., Teukolsky, S. A., Vetterling, W. T., & Flannery, B. P. (1992). Numerical recipes in C. Cambridge: Cambridge University Press.
Rabiner, L. (1989). A tutorial on hidden Markov models and selected applications in speech recognition. Proceedings of the IEEE, 77, 257–286.
Rasmussen, C. E., & Williams, C. K. I. (2006). Gaussian processes for machine learning. Cambridge: MIT.
Roweis, S., & Ghahramani, Z. (1999). A unifying review of linear Gaussian models. Neural Computation, 11, 305–345.
Rybicki, G., & Hummer, D. (1991). An accelerated lambda iteration method for multilevel radiative transfer, appendix b: Fast solution for the diagonal elements of the inverse of a tridiagonal matrix. Astronomy and Astrophysics, 245, 171.
Rybicki, G. B., & Press, W. H. (1995). Class of fast methods for processing irregularly sampled or otherwise inhomogeneous one-dimensional data. Physical Review Letters, 74(7), 1060–1063. doi:10.1103/PhysRevLett.74.1060.
Salakhutdinov, R., Roweis, S. T., & Ghahramani, Z. (2003). Optimization with EM and expectation-conjugate-gradient. International Conference on Machine Learning, 20, 672–679.
Smith, A. C., & Brown, E. N. (2003). Estimating a state-space model from point process observations. Neural Computation, 15, 965–991.
Snyder, D. L. (1975). Random point processes. New York: Wiley.
Tierney, L., Kass, R. E., & Kadane, J. B. (1989). Fully exponential Laplace approximation to posterior expectations and variances. Journal of the American Statistical Association, 84, 710–716.
West, M., Harrison, J. P., & Migon, H. S. (1985). Dynamic generalized linear models and Bayesian forcasting. Journal of the American Statistical Association, 80, 73–83.
Ypma, A., & Heskes, T. (2005). Novel approximations for inference in nonlinear dynamical systems using expectation propagation. Neurocomputing, 69, 85–99. doi:10.1016/j.neucom.2005.02.020.
Yu, B. M., Shenoy, K. V., & Sahani, M. (2006). Expectation propagation for inference in non-linear dynamical models with Poisson observations. In Proceedings of the nonlinear statistical signal processing workshop. Piscataway: IEEE.
Acknowledgements
We thank Y. Ahmadian, R. Kass, M. Nikitchenko, K. Rahnama Rad, M. Vidne and J. Vogelstein for helpful conversations and comments. SK is supported by NIH grants R01 MH064537, R01 EB005847 and R01 NS050256. LP is supported by NIH grant R01 EY018003, an NSF CAREER award, and a McKnight Scholar award.
Author information
Authors and Affiliations
Corresponding author
Additional information
Action Editor: Nicolas Brunel
Appendices
Appendix A: Point process filter and smoother
A simple version of the point process filter approximates the filtered distribution Eq. (26) to a Gaussian centered by its mode (Brown et al. 1998). Let x i|i and V i|i be the (approximate) mode and covariance matrix for the filtered distribution Eq. (26), and x i|i − 1 and V i|i − 1 be the mode and covariance matrix for the predictive distribution Eq. (27) at time i. Let l(x i ) = log{ p(y i |x i ) p(x i |y 1:i − 1) }. The filtered distribution is then approximated to a Gaussian whose mean and covariance are \(x_{i|i} = \arg\max_{x_i}l(x_i)\) and \(V_{i|i} = -[\nabla\nabla_{x_i} l(x_{i|i})]^{-1}\), respectively. When the state-transition density is linear Gaussian, \(p(x_i|x_{i-1}) = \mathcal{N}(F_i x_{i-1}, Q_i)\), the predictive distribution Eq. (27) is also Gaussian, whose mean and covariance are computed as
Since the filtered and predictive distributions are Gaussian, the smoothing distribution Eq. (28) is also Gaussian, which can be computed by the standard Kalman smoother (Smith and Brown 2003). Let x i|N and V i|N be the mean and covariance the smoothing distribution at time i. The recursive smoothing equation corresponding to Eq. (28) is given by
There are now several versions of the point process filter depending on the choice of the mean and variance of the approximate filtered distribution. In Eden et al. (2004), the filtered distribution at each time step i is approximated to a Gaussian by expanding its logarithm in a Taylor series about x i|i − 1 up to the second-order term, which results in a simpler algorithm. Koyama et al. (unpublished manuscript) proposed a more sophisticated method by utilizing the fully exponential Laplace approximation (Tierney et al. 1989), which achieves second-order accuracy in approximating the posterior expectation.
For the leaky IF model with hard-threshold, the standard Taylor-series-based recursions (Brown et al. 1998) do not apply (due to the discontinuity of log p(y i |x i )), and therefore we have not included comparisons to the point-process smoother in Figs. 4–6. However, it is worth noting that in this case the filtered distribution Eq. (26) can be approximated recursively as a truncated Gaussian defined on ( − ∞ , x th ], and hence the approximate mean and variance can be obtained analytically; we found that this moment-matching method behaves similarly to the EP method (this is unsurprising, since EP is also based on a moment-matching procedure; data not shown).
Appendix B: Gaussian quadrature in EP algorithm
The expectation of a function of x i , f(x i ), with respect to p(x i |y 1:N ) in Eq. (38) is expressed as
where
is Gaussian since α i − 1(x i − 1) and p(x i |x i − 1) are also Gaussian. By introducing the Laplace approximation, \(p_L(x_i)\equiv \mathcal{N}(m, v) \approx p(y_i|x_i)\beta_i(x_i) g(x_i)\), as a proposal distribution, the expectation can be expressed as
After a linear change of variable, \(x_i=\sqrt{v}u+m\), we have the standard form of the Gauss-Hermite quadrature,
where the weights w l and evaluation points u l are chosen according to a quadrature rule. The advantages of this method is that it requires only an inner product once the weights and evaluation points are calculated. (These only have to be computed once.) The expectation of f(x i − 1) with respect to p(x i − 1|y 1:N ) in Eq. (39) can be computed in the same way.
For the leaky IF model with hard-threshold, the observation model is given by the step-function, and thus the integral in Eq. (73) becomes
As a result, the expectation Eq. (73) is reduced to the integral over a truncated Gaussian, which can be computed analytically.
Rights and permissions
About this article
Cite this article
Koyama, S., Paninski, L. Efficient computation of the maximum a posteriori path and parameter estimation in integrate-and-fire and more general state-space models. J Comput Neurosci 29, 89–105 (2010). https://doi.org/10.1007/s10827-009-0150-x
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10827-009-0150-x