TRAKR - A reservoir-based tool for fast and accurate classification of neural time-series patterns

Abstract

AbstractNeuroscience has seen a dramatic increase in the types of recording modalities and complexity of neural time-series data collected from them. The brain is a highly recurrent system producing rich, complex dynamics that result in different behaviors. Correctly distinguishing such nonlinear neural time series in real-time, especially those with non-obvious links to behavior, could be useful for a wide variety of applications. These include detecting anomalous clinical events such as seizures in epilepsy, and identifying optimal control spaces for brain machine interfaces. It remains challenging to correctly distinguish nonlinear time-series patterns because of the high intrinsic dimensionality of such data, making accurate inference of state changes (for intervention or control) difficult. Simple distance metrics, which can be computed quickly do not yield accurate classifications. On the other end of the spectrum of classification methods, ensembles of classifiers or deep supervised tools offer higher accuracy but are slow, data-intensive, and computationally expensive to train and deploy. We introduce a reservoir-based tool, state tracker (TRAKR), which offers the high accuracy of ensembles or deep supervised methods while preserving the computational benefits of simple distance metrics. After one-shot training, TRAKR can accurately, and in real time, detect deviations in test patterns. By forcing the weighted dynamics of the reservoir to fit a desired pattern directly, we avoid many rounds of expensive optimization. Then, keeping the output weights frozen, we use the error signal generated by the reservoir in response to a particular test pattern as a classification boundary. We show that using this approach, TRAKR accurately detects changes in synthetic time series. We then compare our tool to several others, showing that it achieves classification performance on par with supervised deep networks on a benchmark dataset–sequential MNIST–, while outperforming all other approaches. When the samples are corrupted by noise, our approach maintains relatively high performance, while supervised deep networks show a sharp decline in performance. We also apply TRAKR to electrocorticography (ECoG) data from the macaque orbitofrontal cortex (OFC), a higher-order brain region involved in encoding the value of expected outcomes. We show that TRAKR can classify different behaviorally relevant epochs in the neural time series with high accuracy. Altogether, we show that TRAKR is a high performing tool for distinguishing patterns in complex nonlinear time-series data, such as neural recordings. With its high performance, robustness to noise, low train- and inference-time, and ease-of-use, it offers a viable alternative to more complex state-of-the art approaches, particularly for real-time applications.