Subject-specific Classifiers Research Articles

Single-Trial Detection and Classification of Event-Related Optical Signals for a Brain-Computer Interface Application.

Event-related optical signals (EROS) measure fast modulations in the brain's optical properties related to neuronal activity. EROS offer a high spatial and temporal resolution and can be used for brain-computer interface (BCI) applications. However, the ability to classify single-trial EROS remains unexplored. This study evaluates the performance of neural network methods for single-trial classification of motor response-related EROS. EROS activity was obtained from a high-density recording montage covering the motor cortex during a two-choice reaction time task involving responses with the left or right hand. This study utilized a convolutional neural network (CNN) approach to extract spatiotemporal features from EROS data and perform classification of left and right motor responses. Subject-specific classifiers trained on EROS phase data outperformed those trained on intensity data, reaching an average single-trial classification accuracy of around 63%. Removing low-frequency noise from intensity data is critical for achieving discriminative classification results with this measure. Our results indicate that deep learning with high-spatial-resolution signals, such as EROS, can be successfully applied to single-trial classifications.

A generic error-related potential classifier based on simulated subjects.

Error-related potentials (ErrPs) are brain signals known to be generated as a reaction to erroneous events. Several works have shown that not only self-made errors but also mistakes generated by external agents can elicit such event-related potentials. The possibility of reliably measuring ErrPs through non-invasive techniques has increased the interest in the brain-computer interface (BCI) community in using such signals to improve performance, for example, by performing error correction. Extensive calibration sessions are typically necessary to gather sufficient trials for training subject-specific ErrP classifiers. This procedure is not only time-consuming but also boresome for participants. In this paper, we explore the effectiveness of ErrPs in closed-loop systems, emphasizing their dependency on precise single-trial classification. To guarantee the presence of an ErrPs signal in the data we employ and to ensure that the parameters defining ErrPs are systematically varied, we utilize the open-source toolbox SEREEGA for data simulation. We generated training instances and evaluated the performance of the generic classifier on both simulated and real-world datasets, proposing a promising alternative to conventional calibration techniques. Results show that a generic support vector machine classifier reaches balanced accuracies of 72.9%, 62.7%, 71.0%, and 70.8% on each validation dataset. While performing similarly to a leave-one-subject-out approach for error class detection, the proposed classifier shows promising generalization across different datasets and subjects without further adaptation. Moreover, by utilizing SEREEGA, we can systematically adjust parameters to accommodate the variability in the ErrP, facilitating the systematic validation of closed-loop setups. Furthermore, our objective is to develop a universal ErrP classifier that captures the signal's variability, enabling it to determine the presence or absence of an ErrP in real EEG data.

Use of subject-specific models to detect fatigue-related changes in running biomechanics: a random forest approach.

Running biomechanics are affected by fatiguing or prolonged runs. However, no evidence to date has conclusively linked this effect to running-related injury (RRI) development or performance implications. Previous investigations using subject-specific models in running have demonstrated higher accuracy than group-based models, however, this has been infrequently applied to fatigue. In this study, two experiments were conducted to determine whether subject-specific models outperformed group-based models to classify running biomechanics during non-fatigued and fatigued conditions. In the first experiment, 16 participants performed four treadmill runs at or around the maximal lactate steady state. In the second experiment, nine participants performed five prolonged runs using commercial wearable devices. For each experiment, two segments were extracted from each trial from early and late in the run. For each participant, a random forest model was applied with a leave-one-run-out cross-validation to classify between the early (non-fatigued) and late (fatigued) segments. Additionally, group-based classifiers with a leave-one-subject-out cross validation were constructed. For experiment 1, mean classification accuracies for the single-subject and group-based classifiers were 68.2 ± 8.2% and 57.0 ± 8.9%, respectively. For experiment 2, mean classification accuracies for the single-subject and group-based classifiers were 68.9 ± 17.1% and 61.5 ± 11.7%, respectively. Variable importance rankings were consistent within participants, but these rankings differed from each participant to those of the group. Although the classification accuracies were relatively low, these findings highlight the advantage of subject-specific classifiers to detect changes in running biomechanics with fatigue and indicate the potential of using big data and wearable technology approaches in future research to determine possible connections between biomechanics and RRI.

Possibilistic distribution distance metric: a robust domain adaptation learning method.

The affective Brain-Computer Interface (aBCI) systems, which achieve predictions for individual subjects through training on multiple subjects, often cannot achieve satisfactory results due to the differences in Electroencephalogram (EEG) patterns between subjects. One tried to use Subject-specific classifiers, but there was a lack of sufficient labeled data. To solve this problem, Domain Adaptation (DA) has recently received widespread attention in the field of EEG-based emotion recognition. Domain adaptation (DA) learning aims to solve the problem of inconsistent distributions between training and test datasets and has received extensive attention. Most existing methods use Maximum Mean Discrepancy (MMD) or its variants to minimize the problem of domain distribution inconsistency. However, noisy data in the domain can lead to significant drift in domain means, which can affect the adaptability performance of learning methods based on MMD and its variants to some extent. Therefore, we propose a robust domain adaptation learning method with possibilistic distribution distance measure. Firstly, the traditional MMD criterion is transformed into a novel possibilistic clustering model to weaken the influence of noisy data, thereby constructing a robust possibilistic distribution distance metric (P-DDM) criterion. Then the robust effectiveness of domain distribution alignment is further improved by a fuzzy entropy regularization term. The proposed P-DDM is in theory proved which be an upper bound of the traditional distribution distance measure method MMD criterion under certain conditions. Therefore, minimizing P-DDM can effectively optimize the MMD objective. Secondly, based on the P-DDM criterion, a robust domain adaptation classifier based on P-DDM (C-PDDM) is proposed, which adopts the Laplacian matrix to preserve the geometric consistency of instances in the source domain and target domain for improving the label propagation performance. At the same time, by maximizing the use of source domain discriminative information to minimize domain discrimination error, the generalization performance of the learning model is further improved. Finally, a large number of experiments and analyses on multiple EEG datasets (i.e., SEED and SEED-IV) show that the proposed method has superior or comparable robustness performance (i.e., has increased by around 10%) in most cases.

Evaluation of a Fast Test Based on Biometric Signals to Assess Mental Fatigue at the Workplace-A Pilot Study.

Non-pathological mental fatigue is a recurring, but undesirable condition among people in the fields of office work, industry, and education. This type of mental fatigue can often lead to negative outcomes, such as performance reduction and cognitive impairment in education; loss of focus and burnout syndrome in office work; and accidents leading to injuries or death in the transportation and manufacturing industries. Reliable mental fatigue assessment tools are promising in the improvement of performance, mental health and safety of students and workers, and at the same time, in the reduction of risks, accidents and the associated economic loss (e.g., medical fees and equipment reparations). The analysis of biometric (brain, cardiac, skin conductance) signals has proven to be effective in discerning different stages of mental fatigue; however, many of the reported studies in the literature involve the use of long fatigue-inducing tests and subject-specific models in their methodologies. Recent trends in the modeling of mental fatigue suggest the usage of non subject-specific (general) classifiers and a time reduction of calibration procedures and experimental setups. In this study, the evaluation of a fast and short-calibration mental fatigue assessment tool based on biometric signals and inter-subject modeling, using multiple linear regression, is presented. The proposed tool does not require fatigue-inducing tests, which allows fast setup and implementation. Electroencephalography, photopletismography, electrodermal activity, and skin temperature from 17 subjects were recorded, using an OpenBCI helmet and an Empatica E4 wristband. Correlations to self-reported mental fatigue levels (using the fatigue assessment scale) were calculated to find the best mental fatigue predictors. Three-class mental fatigue models were evaluated, and the best model obtained an accuracy of 88% using three features, (C3), and the (O2 and C3) ratios, from one minute of electroencephalography measurements. The results from this pilot study show the feasibility and potential of short-calibration procedures and inter-subject classifiers in mental fatigue modeling, and will contribute to the use of wearable devices for the development of tools oriented to the well-being of workers and students, and also in daily living activities.

Subject-Specific Cognitive Workload Classification Using EEG-Based Functional Connectivity and Deep Learning.

Cognitive workload is a crucial factor in tasks involving dynamic decision-making and other real-time and high-risk situations. Neuroimaging techniques have long been used for estimating cognitive workload. Given the portability, cost-effectiveness and high time-resolution of EEG as compared to fMRI and other neuroimaging modalities, an efficient method of estimating an individual’s workload using EEG is of paramount importance. Multiple cognitive, psychiatric and behavioral phenotypes have already been known to be linked with “functional connectivity”, i.e., correlations between different brain regions. In this work, we explored the possibility of using different model-free functional connectivity metrics along with deep learning in order to efficiently classify the cognitive workload of the participants. To this end, 64-channel EEG data of 19 participants were collected while they were doing the traditional n-back task. These data (after pre-processing) were used to extract the functional connectivity features, namely Phase Transfer Entropy (PTE), Mutual Information (MI) and Phase Locking Value (PLV). These three were chosen to do a comprehensive comparison of directed and non-directed model-free functional connectivity metrics (allows faster computations). Using these features, three deep learning classifiers, namely CNN, LSTM and Conv-LSTM were used for classifying the cognitive workload as low (1-back), medium (2-back) or high (3-back). With the high inter-subject variability in EEG and cognitive workload and recent research highlighting that EEG-based functional connectivity metrics are subject-specific, subject-specific classifiers were used. Results show the state-of-the-art multi-class classification accuracy with the combination of MI with CNN at 80.87%, followed by the combination of PLV with CNN (at 75.88%) and MI with LSTM (at 71.87%). The highest subject specific performance was achieved by the combinations of PLV with Conv-LSTM, and PLV with CNN with an accuracy of 97.92%, followed by the combination of MI with CNN (at 95.83%) and MI with Conv-LSTM (at 93.75%). The results highlight the efficacy of the combination of EEG-based model-free functional connectivity metrics and deep learning in order to classify cognitive workload. The work can further be extended to explore the possibility of classifying cognitive workload in real-time, dynamic and complex real-world scenarios.

Sensor Fusion of Vision, Kinetics, and Kinematics for Forward Prediction During Walking With a Transfemoral Prosthesis

Powered lower limb prostheses utilize activity specific controllers, where predicting desired activity and transitioning in a timely manner is essential for seamless locomotion without trips. Previously research considered EMG and mechanical sensors for these predictions; however, predictability must be improved to ensure safe usage. Previously we showed that combining features from mechanical sensors, EMG and Vision yielded greater repeatability, greater separability and lower variability. Here we compare performance of offline forward prediction systems combining these different sensor modalities. We trained and tested subject-specific classifiers for steady-state and transition steps, with data from 8 able-bodied subjects, 4 able-bodied subjects walking with a powered knee-ankle prosthesis using a bypass socket, and a single transfemoral amputee walking with a knee-ankle prosthesis. Fusing Mechanical, EMG, and Vision features produced the best classification for all subjects, with transition error rates in the range of 1% and steady-state error rates close to 0%. Though generalizability was good across able-bodied subjects, it was poor when training with able-bodied or bypass subjects and testing with our amputee subject, regardless of sensor modality, particularly for transition steps. Therefore, we believe a general classifier will require inclusion of amputee training data. Future work will test more subjects and continue development of a general control system.

Subject-specific mental workload classification using EEG and stochastic configuration network (SCN)

Mental workload assessment of the operators in some safety-critical human-machine systems is an important research topic. In this paper, an experiment was designed to obtain the electroencephalogram (EEG) data under three levels of mental workload. The EEG data of multiple subjects were used for the mental workload classification based on the stochastic configuration network (SCN). The subject-specific classifiers (SSCs) were built by the individual EEG data. The results showed that the range of SSC test accuracy was between 56.5 % and 90.2 % with an average of 75.9 %. The SSC accuracy had a positive correlation with the operating accuracy (r = 0.852, p < 0.01). For comparison, the subject-multiple classifiers (SMCs) were established with the EEG data of multiple subjects. The results showed that the SSCs had a lower time-consuming and higher prediction accuracy than the SMCs. But the SMCs might embody the trend of statistical performance for a large number of subjects. This study provided an effective modeling method for the classification of mental workload, and it would bring great convenience to the practical application in the future.

3D Visual passcode: Speech-driven 3D facial dynamics for behaviometrics

Face biometrics have achieved remarkable performance over the past decades, but unexpected spoofing of the static faces poses a threat to information security. There is an increasing demand for stable and discriminative biological modalities which are hard to be mimicked and deceived. Speech-driven 3D facial motion is a distinctive and measurable behavior-signature that is promising for biometrics. In this paper, we propose a novel 3D behaviometrics framework based on a “3D visual passcode” derived from speech-driven 3D facial dynamics. The 3D facial dynamics are jointly represented by 3D-keypoint-based measurements and 3D shape patch features, extracted from both static and speech-driven dynamic regions. An ensemble of subject-specific classifiers are then trained over selected discriminative features, which allows for a discriminant speech-driven 3D facial dynamics representation. We construct the first publicly available Speech-driven 3D Facial Motion dataset (S3DFM) that includes 2D-3D face video plus audio samples from 77 participants. The experimental results on the S3DFM show that the proposed pipeline achieves a face identification rate of 96.1%. Detailed discussions are presented, concerning anti-spoofing, head pose variation, video frame rate, and applicability cases. We also give comparison with other baselines on “deep” and “shallow” 2D face features.

Emotion Recognition for Brain Machine Interface: Non-linear Spectral Analysis of EEG Signals Using Empirical Mode Decomposition.

Emotions are a fundamental part of the human experience but currently there are no methods that can objectively detect and categorize them. This study utilizes the empirical mode decomposition (EMD) method to categorize emotions from encephalography (EEG) recordings. In the past, EMD has proven to be a very useful signal analysis tool because of its ability to decompose nonstationary signals, like those from an EEG, into component signals with varying frequency content called intrinsic mode functions (IMFs). The method in this paper utilizes three features extracted from the IMFs-the first difference of time, the first difference of phase, and the normalized energy-for data categorization using support vector machine (SVM) classifiers. Two classifiers were trained for each subject, one for valence and another for arousal. The mean accuracies yielded for valence and arousal were 75.86% and 75.31%, respectively. The results of this study verify previous findings by other researchers that these three features are useful in emotion recognition when applied to previously recorded EEG data, though we add the caveat that subject-specific classifiers are needed instead of generalized, global classifiers.

Estimation of joint angle based on surface electromyogram signals recorded at different load levels.

To control upper-limb exoskeletons and prostheses, surface electromyogram (sEMG) is widely used for estimation of joint angles. However, the variations in the load carried by the user can substantially change the recorded sEMG and consequently degrade the accuracy of joint angle estimation. In this paper, we aim to deal with this problem by training classification models using a pool of sEMG data recorded from all different loads. The classification models are trained as either subject-specific or subject-independent, and their results are compared with the performance of classification models that have information about the carried load. To evaluate the proposed system, the sEMG signals are recorded during elbow flexion and extension from three participants at four different loads (i.e. 1, 2, 4 and 6 Kg) and six different angles (i.e. 0, 30, 60, 90, 120, 150 degrees). The results show while the loads were assumed unknown and the applied training data was relatively small, the proposed joint angle estimation model performed significantly above the chance level in both the subject-specific and subject-independent models. However, transferring from known to unknown load in the subject-specific classifiers leads to 20% to 32% loss in the average accuracy.

Cross-Subject EEG Feature Selection for Emotion Recognition Using Transfer Recursive Feature Elimination.

Using machine-learning methodologies to analyze EEG signals becomes increasingly attractive for recognizing human emotions because of the objectivity of physiological data and the capability of the learning principles on modeling emotion classifiers from heterogeneous features. However, the conventional subject-specific classifiers may induce additional burdens to each subject for preparing multiple-session EEG data as training sets. To this end, we developed a new EEG feature selection approach, transfer recursive feature elimination (T-RFE), to determine a set of the most robust EEG indicators with stable geometrical distribution across a group of training subjects and a specific testing subject. A validating set is introduced to independently determine the optimal hyper-parameter and the feature ranking of the T-RFE model aiming at controlling the overfitting. The effectiveness of the T-RFE algorithm for such cross-subject emotion classification paradigm has been validated by DEAP database. With a linear least square support vector machine classifier implemented, the performance of the T-RFE is compared against several conventional feature selection schemes and the statistical significant improvement has been found. The classification rate and F-score achieve 0.7867, 0.7526, 0.7875, and 0.8077 for arousal and valence dimensions, respectively, and outperform several recent reported works on the same database. In the end, the T-RFE based classifier is compared against two subject-generic classifiers in the literature. The investigation of the computational time for all classifiers indicates the accuracy improvement of the T-RFE is at the cost of the longer training time.

Sleep apnea detection using time-delayed heart rate variability.

Sleep apnea is a sleep disorder distinguished by repetitive absence of breathing. Compared with the traditional expensive and cumbersome methods, sleep apnea diagnosis or screening with physiological information that can be easily acquired is needed. This paper describes algorithms using heart rate variability (HRV) to automatically detect sleep apneas as long as it can be easily acquired with unobtrusive sensors. Because the changes in cardiac activity are usually hysteretic than the presence of apneas with a few minutes, we propose to use the delayed HRV features to identify the episodes with sleep apneic events. This is expected to help improve the apnea detection performance. Experiments were conducted with a data set of 23 sleep apnea patients using support vector machine (SVM) classifiers and cross validations. Results show that using eleven HRV features with a time delay of 1.5 minutes rather than the features without time delay for SA detection, the overall accuracy increased from 74.9% to 76.2% and the Cohen's Kappa coefficient increased from 0.49 to 0.52. Further, an accuracy of 94.5% and a Kappa of 0.89 were achieved when applying subject-specific classifiers.

EEG-based workload estimation across affective contexts.

Workload estimation from electroencephalographic signals (EEG) offers a highly sensitive tool to adapt the human–computer interaction to the user state. To create systems that reliably work in the complexity of the real world, a robustness against contextual changes (e.g., mood), has to be achieved. To study the resilience of state-of-the-art EEG-based workload classification against stress we devise a novel experimental protocol, in which we manipulated the affective context (stressful/non-stressful) while the participant solved a task with two workload levels. We recorded self-ratings, behavior, and physiology from 24 participants to validate the protocol. We test the capability of different, subject-specific workload classifiers using either frequency-domain, time-domain, or both feature varieties to generalize across contexts. We show that the classifiers are able to transfer between affective contexts, though performance suffers independent of the used feature domain. However, cross-context training is a simple and powerful remedy allowing the extraction of features in all studied feature varieties that are more resilient to task-unrelated variations in signal characteristics. Especially for frequency-domain features, across-context training is leading to a performance comparable to within-context training and testing. We discuss the significance of the result for neurophysiology-based workload detection in particular and for the construction of reliable passive brain–computer interfaces in general.

FNIRS-based online deception decoding

Deception involves complex neural processes in the brain. Different techniques have been used to study and understand brain mechanisms during deception. Moreover, efforts have been made to develop schemes that can detect and differentiate deception and truth-telling. In this paper, a functional near-infrared spectroscopy (fNIRS)-based online brain deception decoding framework is developed. Deploying dual-wavelength fNIRS, we interrogate 16 locations in the forehead when eight able-bodied adults perform deception and truth-telling scenarios separately. By combining preprocessed oxy-hemoglobin and deoxy-hemoglobin signals, we develop subject-specific classifiers using the support vector machine. Deception and truth-telling states are classified correctly in seven out of eight subjects. A control experiment is also conducted to verify the deception-related hemodynamic response. The average classification accuracy is over 83.44% from these seven subjects. The obtained result suggests that the applicability of fNIRS as a brain imaging technique for online deception detection is very promising.

Improved patient specific seizure detection during pre-surgical evaluation

Objective There is considerable interest in improved off-line automated seizure detection methods that will decrease the workload of EEG monitoring units. Subject-specific approaches have been demonstrated to perform better than subject-independent ones. However, for pre-surgical diagnostics, the traditional method of obtaining a priori data to train subject-specific classifiers is not practical. We present an alternative method that works by adapting the threshold of a subject-independent to a specific subject based on feedback from the user. Methods A subject-independent quadratic discriminant classifier incorporating modified features based partially on the Gotman algorithm was first built. It was then used to derive subject-specific classifiers by determining subject-specific posterior probability thresholds via user interaction. The two schemes were tested on 529 h of intracranial EEG containing 63 seizures from 15 subjects undergoing pre-surgical evaluation. To provide comparison, the standard Gotman algorithm was implemented and optimised for this dataset by tuning the detection thresholds. Results Compared to the tuned Gotman algorithm, the subject-independent scheme reduced the false positive rate by 51% (0.23 to 0.11 h −1) while increasing sensitivity from 53% to 62%. The subject-specific scheme further improved sensitivity to 78%, but with a small increase in false positive rate to 0.18 h −1. Conclusions The results suggest that a subject-independent classifier scheme with modified features is useful for reducing false positive rate, while subject adaptation further enhances performance by improving sensitivity. The results also suggest that the proposed subject-adapted classifier scheme approximates the performance of the subject-specific Gotman algorithm. Significance The proposed method could potentially increase the productivity of offline EEG analysis. The approach could also be generalised to enhance the performance of other subject independent algorithms.