Ensembles of change-point detectors: implications for real-time BMI applications
- 197 Downloads
Brain-machine interfaces (BMIs) have been widely used to study basic and translational neuroscience questions. In real-time closed-loop neuroscience experiments, many practical issues arise, such as trial-by-trial variability, and spike sorting noise or multi-unit activity. In this paper, we propose a new framework for change-point detection based on ensembles of independent detectors in the context of BMI application for detecting acute pain signals. Motivated from ensemble learning, our proposed “ensembles of change-point detectors” (ECPDs) integrate multiple decisions from independent detectors, which may be derived based on data recorded from different trials, data recorded from different brain regions, data of different modalities, or models derived from different learning methods. By integrating multiple sources of information, the ECPDs aim to improve detection accuracy (in terms of true positive and true negative rates) and achieve an optimal trade-off of sensitivity and specificity. We validate our method using computer simulations and experimental recordings from freely behaving rats. Our results have shown superior and robust performance of ECPDS in detecting the onset of acute pain signals based on neuronal population spike activity (or combined with local field potentials) recorded from single or multiple brain regions.
KeywordsBrain machine interface Change point detection Ensemble learning Population codes Acute pain Poisson linear dynamical system Support vector machine Event-related potential
The authors thank Eric J. Robinson for English proofreading. The work was supported by the NSF-CRCNS grant IIS-130764 (Z.C.), NSF-NCS grant #1835000 (Z.C., J.W.), NIH grants R01-NS100016 (Z.C., J.W.), R01-MH118928 (Z.C.) and R01-GM115384 (J.W.), as well as the China’s Natural Science Foundation #31627802 and Fundamental Research Funds for the Central Universities (Y.C.).
Compliance with Ethical Standards
Conflict of interest
The authors declare that they have no conflict of interest.
- Chen, Z. (Ed.). (2015). Advanced State Space Methods in Neural and Clinical Data. Cambridge: Cambridge University Press.Google Scholar
- Chen, Z., & Wang, J. (2016). Statistical analysis of neuronal population codes for encoding acute pain. In Proceedings of IEEE ICASSP (pp. 829–833).Google Scholar
- Chen, Z., Hu, S., Zhang, Q., Wang, J. (2017b). Quickest detection for abrupt changes in neuronal ensemble spiking activity using model-based and model-free approaches. In Proceedings of 8th International IEEE/EMBS Conference on Neural Engineering (NER’17) (pp. 481–484).Google Scholar
- Daou, I., Tuttle, A.H., Longo, G., Wieskopf, J.S., Bonin, R.P., Ase, A.R., Wood, J.N., De Koninck, Y., Ribeiro-da Silva, A., Mogil, J.S., Sgula, P. (2013). Remote optogenetic activation and sensitization of pain pathways in freely moving mice. The Journal of Neuroscience, 33, 18631–18640.CrossRefGoogle Scholar
- Dietterich, T.G., & Roli, F. (2000). Ensemble methods in machine learning. In Gayar, N.E., & Kittler, J. (Eds.) Multiple Classifier Systems: Springer.Google Scholar
- Goodman, I.N., & Johnson, D.H. (2008). Information theoretic bounds on neural prosthesis effectiveness: The importance of spike sorting. In Proceedings of ICASSP’08 (pp. 5204–5207).Google Scholar
- Hu, S., Zhang, Q., Wang, J., Chen, Z. (2017). A real-time rodent neural interface for deciphering acute pain signals from neuronal ensemble spike activity. In Proceedings of 51st Asilomar Conference of Signals, Systems, and Computers (pp. 93–97).Google Scholar
- Jaffe, A., Nadler, B., Kluger, Y. (2015). Estimating the accuracies of multiple classifiers without labeled data. In AISTAT’15.Google Scholar
- Kopepcke, L., Ashida, G., Kretzberg, J. (2016). Single and multiple change point detection in spike trains: comparison of different CUSUM, methods. Frontiers in Systems Neuroscience, 10, 51.Google Scholar
- Macke, J.H., Buesing, L., Sahani, M. (2015). Estimating state and parameters in state space models of spike trains. In Chen, Z. (Ed.) Advanced State Space Methods in Neural and Clinical Data. Cambridge: Cambridge University Press.Google Scholar
- Mallat, S. (2008). A Wavelet Tour of Signal Processing: the Sparse Way, 3rd edn. Cambridge: Academic Press.Google Scholar
- Raykar, V.C., Shipeng, Y., Zhao, L.H., Valdez, G.H., Florin, C., Bogoni, L., Moy, L. (2010). Learning with crowds. Journal of Machine Learning Research, 11, 1297–1322.Google Scholar
- Scholkopf, B., & Smola, A.J. (2001). Learning with kernels: Support vector machines regularization, Optimization, and Beyond. Cambridge: MIT Press.Google Scholar