Abstract
Deep brain stimulation (DBS) of Subthalamic nucleus (STN) is the most successful treatment for advanced Parkinson’s disease. Localization of the STN through Microelectrode recordings (MER) is a key step during the surgery. However, it is a complex task even for a skilled neurosurgeon. Different researchers have developed methodologies for processing and classification of MER signals to locate the STN. Previous works employ the classical paradigm of supervised classification, assuming independence between patients. The aim of this paper is to introduce a patient-dependent learning scenario, where the predictive ability for STN identification at the level of a particular patient, can be used to improve the accuracy for STN identification in other patients. Our inspiration is the multi-task learning framework, that has been receiving increasing interest within the machine learning community in the last few years. To this end, we employ the multi-task Gaussian processes framework that exhibits state of the art performance in multi-task learning problems. In our context, we assume that each patient undergoing DBS is a different task, and we refer to the method as multi-patient learning. We show that the multi-patient learning framework improves the accuracy in the identification of STN in a range from 4.1 to 7.7%, compared to the usual patient-independent setup, for two different datasets. Given that MER are non stationary and noisy signals. Traditional approaches in machine learning fail to recognize accurately the STN during DBS. By contrast in our proposed method, we properly exploit correlations between patients with similar diseases, obtaining an additional information. This information allows to improve the accuracy not only for locating STN for DBS but also for other biomedical signal classification problems.
Similar content being viewed by others
Notes
The interested reader can download this dataset from https://dl.dropboxusercontent.com/u/43310202/DB_UTP.rar.
The parametric classifiers, KNN1, KNN3 and the SVM are implemented using the PRTOOLS toolbox obtained from http://www.prtools.org/. GPR is implemented using the Gaussian Process Toolbox from http://staffwww.dcs.shef.ac.uk/people/N.Lawrence/gp/. GPC is implemented using the Gaussian Process Toolbox from http://www.gaussianprocess.org/gpml/code/matlab/doc/.
The latent functions \(u_q^i({\mathbf {x}})\) share the same covariance \(k_q({\mathbf {x}}, {\mathbf {x}}')\), irrespectively of the value of i
For simplicity, we assume all the task are evaluated at the same number of tests inputs, \(N_*\). This is mainly to avoid notation clutterness.
Strictly speaking, we also need the coefficients \(b_{d,d'}^q\) to lead to a positive semidefinite function for \(k_{d,d'}(\cdot , \cdot )\). This can be enforced by using \(b_{d,d'}^q = \sum _{i=1}^{R_q}a_{d,q}^ia_{d',q}^i\), and estimating \(a_{d,q}^i\) instead of \(b_{d,d'}^q\).
We implement MC and ML using the MULTIGP Toolbox retrieved from http://staffwww.dcs.shef.ac.uk/people/N.Lawrence/multigp/. We implement MI using software available at http://homepages.inf.ed.ac.uk/gsanguin/software.html.
References
Alvarez M, Lawrence N (2008) Sparse convolved Gaussian process for multi-output regression. In: Advances in Neural Information Processing Systems (NIPS), pp 57–64
Alvarez M, Rosasco L, Lawrence N (2012) Kernels for vector-valued functions: a review. Found Trends Mach Learn 4:195–266
Álvarez MA, Lawrence ND (2011) Computationally efficient convolved multiple output Gaussian processes. J Mach Learn Res 12:1459–1500
Argyriou A, Evgeniou T, Pontil M (2008) Convex multi-task feature learning. Mach Learn 73(3):243–272
Bakker B, Heskes T (2003) Task clustering and gating for Bayesian multitask learning. J Mach Learn Res 4:83–99
Benabid A (2003) Deep brain stimulation for Parkinson’s disease. Curr Opin Neurobiol 13:696–706
Bishop C (2006) Pattern recognition and machine learning. Springer, USA
Bonilla EV, Agakov FV, Williams CKI (2007) Kernel multi-task learning using task-specific features. In: Proceedings of the 11th International Conference on Artificial Intelligence and Statistics (AISTATS)
Caruana R (1997) Multitask learning. Mach Learn 28:41–75
Chai KM (2009) Generalization errors and learning curves for regression with multi-task Gaussian processes. In: Advances in Neural Information Processing Systems 22 (NIPS), pp 279–287
Chan H, Wu T, Lee S, Lin M, He S, Chao P, Tsai Y (2010) Unsupervised wavelet-based spike sorting with dynamic codebook searching and replenishment. Neurocomputing 73(7-9):1513–1527
Chuang WY, Young KY, Chao CP, Tsai ST, Chen SY (2012) Locating optimal electrodes placement via microelectrode recording in general anesthetic patients during deep brain stimulation. In: IEEE International Conference on Biomedical Engineering and Biotechnology, pp 1048–1051
Deslauriers G, Dubuc S (1989) Symmetric iterative interpolation processes. Constr Approx 5:49–68
Fu Q, Clements M, Mewes K (2005) Neural cell type recognition between globus pallidus externus and globus pallidus internus by gaussian mixture modelling. In: ICASSP, pp 53–56
Gemmar P, Gronz O, Henrichs T, Hertel F (2008) Advanced methods for target navigation using microelectrode recordings in stereotactic neurosurgery for deep brain stimulation. In: CBMS ?08: Proceedings of the 21st IEEE International Symposium on Computer-Based Medical Systems, IEEE Computer Society, vol 21, pp 99–104
Giraldo E, Castellanos G, Orozco A (2008) Feature extraction for MER signals using adaptive filter banks. In: Electronics, Robotics and Automotive Mechanics Conference, pp 582–585
Goovaerts P (1997) Geostatistics for natural resources evaluation. Oxford University Press, USA
Guarin D, Orozco A, Delgado E, Guijarro E (2010) On detecting determinism and nonlinearity in microelectrode recording signals: Approach based on non-stationary surrogate data methods. In: EMBC 2010, Proceedings of the 32nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp 4032–4035
Guillen P, Martinez F, Sanchez R, Argaez M, Velasquez L (2011) Characterization of subcortical structures during deep brain stimulation utilizing support vector machines. In: EMBC 2011, Proceedings of the 33rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp 7949–7952
Krack P, Batir A, Van Blercom N, Chabardes S, Fraix V, Ardouin C, Koudsie A, Limousin P, Benazzouz A, LeBas J, Benabid A, Pollak P (2003) Five-year follow-up of bilateral stimulation of the subthalamic nucleus in advanced Parkinson’s disease. N Engl J Med 349(20):1925–1934
Lees A, Hardy J, Revesz T (2009) Parkinson’s disease. Lancet 2009(373):2055–2066
Maks C, Butson C, Walter B, Vitek J, McIntyre C (2009) Deep brain stimulation activation volumes and their association with neurophysiological mapping and therapeutic outcomes. J Neurol Neurosurg Psychiatry 80(6):659–666
Miocinovic S, Lempka S, Russo G, Maks C, Butson C, Sakaie K, Vitek J, McIntyre C (2009) Experimental and theoretical characterization of the voltage distribution generated by deep brain stimulation. Exp Neurol 216(1):166–176
NINDS (2004) Parkinson’s disease: challenges, progress, and promise. National Institute of Health, pp 6–10
Niu H, Wang J (2014) Financial time series prediction by a random data-time effective rbf neural network. J Soft Comput 18:497–508
Novak P, Daniluk S, Elias S, Nazzaro J (2007) Detection of the subthalamic nucleus in microelectrographic recordings in Parkinson disease using the high frequency (500 hz) neuronal background. Neurosurgery 106:175–179
Orozco A, Álvarez M, Guijarro E, Castellanos G (2006) Identification of spike sources using proximity analysis through hidden Markov models. In: EMBC 2006, Proceedings of the 28th annual International Conference of the IEEE Engineering in Medicine and Biology Society, vol 28, pp 5555–5558
Pillonetto G, Dinuzzo F, De Nicolao G (2010) Bayesian online multitask learning of Gaussian processes. IEEE Trans Pattern Anal Mach Intell 32(2):193–205
Pinzon R, Garces M, Orozco A, Nazzar J (2009) Automatic identification of various nuclei in the basal ganglia for Parkinson’s disease neurosurgery. In: EMBC 2009, Proceedings of the 31st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, vol 31, pp 99–104
Pinzon R, Orozco A, Carmona H, Castellanos G (2010) Feature selection using an ensemble of optimal wavelet packet and learning machine: Application to MER signals. In: Communication Systems Networks and Digital Signal Processing (CSNDSP), vol 7, pp 25–30
Pizarro J, Guerrero E, Galindo P (2002) Multiple comparison procedures applied to model selection. Neurocomputing 48:155–173
Quiroga R, Ndasdy Z, Ben-Shaul Y (2004) Unsupervised spike detection and sorting with wavelets and superparamagnetic clustering. Neural Comput 16:1661–1687
Rasmussen CE, Williams CKI (2006) Gaussian processes for machine learning. MIT Press, Cambridge
Rodriguez A, Delgado E, Orozco A, Castellanos G, Guijarro E (2008) Nonlinear dynamics techniques for the detection of the brain areas using MER signals. IEEE Int Conf BioMed Eng Inf 2:198–202
Shoham S, Fellows M, Normann R (2003) Automatic spike sorting using mixtures of multivariate t-distributions. J Neurosci Methods 127:111–122
Skolidis G, Sanguinetti G (2011) Bayesian multi-task classification with Gaussian process priors. IEEE Trans Neural Netw 22:2011–2021
Tahgva A (2011) Hidden semi-Markov models in the computerized decoding of microelectrode recording data for deep brain stimulator placement. World Neurosurg 75:758–763
Taormina R et al (2015) Data-driven input variable selection for rainfall-run modeling using binary-coded particle swarm optimzation and extreme learning machines. J Hidrol 529:1617–1632
Vargas H, Orozco A, Alvarez M (2012) Multi-patient learning increases accuracy for Subthalamic nucleus identification in deep brain stimulation. In: EMBC 2012, Proceedings of the 34th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp 4341–4344
Wright D, Nakamura K, Maeda T, Kutsuzawa K, Miyawaki K, Nagata K (2008) Research and development of a portable device to quantify muscle tone in patients with Parkinsons disease. In: EMBC 2008, Proceedings of the 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, vol 30, pp 2825–2827
Wu C et al (2009) Methods to improve neural network performance in daily flows prediction. J Hydrol 372:80–93
Acknowledgements
We would like to thank the MD Hans Carmona Villada and the Institute of Epilepsy and Parkinson of Eje Cafetero, who helped in organizing the database DB-UTP. We also acknowledge Dr. Enrique Guijarro, for providing us with the DB-UPV database. This research has been developed under the project with code 111056934461 financed by Colciencias. H.D Vargas Cardona author is funded by Colciencias under the program: formación de alto nivel para la ciencia, la tecnología y la innovación—Convocatoria 617 de 2013.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Vargas Cardona, H.D., Álvarez, M.A. & Orozco, Á.A. Multi-task learning for subthalamic nucleus identification in deep brain stimulation. Int. J. Mach. Learn. & Cyber. 9, 1181–1192 (2018). https://doi.org/10.1007/s13042-017-0640-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13042-017-0640-5