Covariate shift estimation based adaptive ensemble learning for handling non-stationarity in motor imagery related EEG-based brain-computer interface

Abstract

The non-stationary nature of electroencephalography (EEG) signals makes an EEG-based brain-computer interface (BCI) a dynamic system, thus improving its performance is a challenging task. In addition, it is well-known that due to non-stationarity based covariate shifts, the input data distributions of EEG-based BCI systems change during inter-and intra-session transitions, which poses great difficulty for developments of online adaptive data-driven systems. Ensemble learning approaches have been used previously to tackle this challenge. However, passive scheme based implementation leads to poor efficiency while increasing high computational cost. This paper presents a novel integration of covariate shift estimation and unsupervised adaptive ensemble learning (CSE-UAEL) to tackle nonstationarity in motor-imagery (MI) related EEG classification. The proposed method first employs an exponentially weighted moving average model to detect the covariate shifts in the common spatial pattern features extracted from MI related brain responses. Then, a classifier ensemble was created and updated over time to account for changes in streaming input data distribution wherein new classifiers are added to the ensemble in accordance with estimated shifts. Furthermore, using two publicly available BCI-related EEG datasets, the proposed method was extensively compared with the state-of-the-art single-classifier based passive scheme, single-classifier based active scheme and ensemble based passive schemes. The experimental results show that the proposed active scheme based ensemble learning algorithm significantly enhances the BCI performance in MI classifications.

An EEG-EMG correlation-based brain-computer interface for hand orthosis supported neuro-rehabilitation

“…Using brain or muscle signals alone for the actuation and control of external devices poses their own challenges. The brain signal based controls often suffer from issues such as lower accuracy and need user-specific adaptation for ensuring the reliability [12,13,14]. The performances of EMG-based controls are also limited primarily because some of the patients suffering from neuro-muscular diseases may have little or no residual EMG activity, and may suffer from the spasticity and fatiguerelated alterations in EMG activity [15].…”

Recurrent Neural Network-based Acute Concussion Classifier using Raw Resting State EEG Data

“…EEG signals are well known to be strongly non-stationary on timescales greater than ∼0.25 s, exhibiting significant variations, and even shifts, in the statistical properties of the signal over time [61,62]. This behaviour poses a significant challenge for extracting stable features from the signal as well as for designing reliable brain-computer interface systems within the context of real environments [85][86][87][88]. In the present study, we leverage this behaviour to generate multiple, effectively independent, realizations of the raw90-rsEEG datasets from each of the participants.…”

“…Neither of these two attempts yielded classification accuracy above 65%, a finding that was not a complete surprise. It is well known that measures extracted from different segments of time duration ∆t, extracted from even a single non-stationary EEG time series collected at one sitting, can vary significantly [61,62,87]. However, prior to extracting the features, we had filtered the data [82][83][84] to remove signal drift, an approach that is commonly adopted to mitigate weak non-stationary behaviour.…”

Hierachical Resampling for Bagging in Multi-Study Prediction with Applications to Human Neurochemical Sensing

“…This is in contrast to much of the concept/hybrid drift literature that deal with streams of data (often time series) that change over time (in one of the manners described above) and are typically from one source, such as environmental or energy data (Yang et al, 2019;Raza et al, 2019;Karnick et al, 2008;Ditzler et al, 2010), (i.e., f (X t ) and/or f (Y t |X t ) change over time). As such, many concept drift methods employ approaches that are inherently tailored to the temporal nature of the data streams, such as moving averages ( [Raza, et al, 2008]), detection systems to identify the presence ( [Yang, et al, 2019]) or speed ( [Minku, et al, 2009]) of a shift at a given instance, and ensemble methods that add, remove, or reweight classifiers across time ( [Raza, et al, 2008]; [Minku, et al, 2009]). While virtual drift work often does not assume changes over time, its methods usually center on reweighting observations ( [Sugiyama, et al, 2008]; ( [Shimodaira, 2000]) and assume that f (Y|X) remains constant between training and test sets.…”