Monitoring of Body Condition in Dairy Cows to Assess Disease Risk at the Individual and Herd Level.

A retrospective longitudinal study assessing the explanatory and predictive capacity of body condition score (BCS) in dairy cows on disease risk at the individual and herd level was carried out. Data from two commercial grazing herds from the Argentinean Pampa were gathered (Herd A = 2100 and herd B = 2600 milking cows per year) for 4 years. Logistic models were used to assess the association of BCS indicators with the odds for anestrus at the cow and herd level. Population attributable fraction (AFP) was estimated to assess the anestrus rate due to BCS indicators. We found that anestrus risk decreased in cows calving with BCS ≥ 3 and losing ≤ 0.5 (OR: 0.07-0.41), and that anestrus rate decreased in cohorts with a high frequency of cows with proper BCS (OR: 0.22-0.45). Despite aggregated data having a good explanatory power, their predictive capacity for anestrus rate at the herd level is poor (AUC: 0.574-0.679). The AFP varied along the study in both herds and tended to decrease every time the anestrous rate peaked. We conclude that threshold-based models with BCS indicators as predictors are useful to understand disease risk (e.g., anestrus), but conversely, they are useless to predict such multicausal disease events at the herd level.

Affective Neural Responses Sonified through Labeled Correlation Alignment.

Sound synthesis refers to the creation of original acoustic signals with broad applications in artistic innovation, such as music creation for games and videos. Nonetheless, machine learning architectures face numerous challenges when learning musical structures from arbitrary corpora. This issue involves adapting patterns borrowed from other contexts to a concrete composition objective. Using Labeled Correlation Alignment (LCA), we propose an approach to sonify neural responses to affective music-listening data, identifying the brain features that are most congruent with the simultaneously extracted auditory features. For dealing with inter/intra-subject variability, a combination of Phase Locking Value and Gaussian Functional Connectivity is employed. The proposed two-step LCA approach embraces a separate coupling stage of input features to a set of emotion label sets using Centered Kernel Alignment. This step is followed by canonical correlation analysis to select multimodal representations with higher relationships. LCA enables physiological explanation by adding a backward transformation to estimate the matching contribution of each extracted brain neural feature set. Correlation estimates and partition quality represent performance measures. The evaluation uses a Vector Quantized Variational AutoEncoder to create an acoustic envelope from the tested Affective Music-Listening database. Validation results demonstrate the ability of the developed LCA approach to generate low-level music based on neural activity elicited by emotions while maintaining the ability to distinguish between the acoustic outputs.

A Novel OpenBCI Framework for EEG-Based Neurophysiological Experiments.

An Open Brain-Computer Interface (OpenBCI) provides unparalleled freedom and flexibility through open-source hardware and firmware at a low-cost implementation. It exploits robust hardware platforms and powerful software development kits to create customized drivers with advanced capabilities. Still, several restrictions may significantly reduce the performance of OpenBCI. These limitations include the need for more effective communication between computers and peripheral devices and more flexibility for fast settings under specific protocols for neurophysiological data. This paper describes a flexible and scalable OpenBCI framework for electroencephalographic (EEG) data experiments using the Cyton acquisition board with updated drivers to maximize the hardware benefits of ADS1299 platforms. The framework handles distributed computing tasks and supports multiple sampling rates, communication protocols, free electrode placement, and single marker synchronization. As a result, the OpenBCI system delivers real-time feedback and controlled execution of EEG-based clinical protocols for implementing the steps of neural recording, decoding, stimulation, and real-time analysis. In addition, the system incorporates automatic background configuration and user-friendly widgets for stimuli delivery. Motor imagery tests the closed-loop BCI designed to enable real-time streaming within the required latency and jitter ranges. Therefore, the presented framework offers a promising solution for tailored neurophysiological data processing.

Association of bovine viral diarrhea virus, bovine herpesvirus 1, and Neospora caninum with late embryonic losses in highly supplemented grazing dairy cows.

The objectives of this study were: 1- to evaluate the association of Bovine Viral Diarrhea Virus (BVDV), Bovine Herpes Virus 1 (BoHV-1), and Neospora caninum (N. caninum) with the risk for Late Embryonic Loss (LEL) in grazing dairy cows, 2- to evaluate blood progesterone concentration at the time of LEL occurrence, and 3- to describe a novel ultrasound-guided technique for conceptus sampling. We run a prospective cohort study involving 92 cows (46 LEL and 46 NLEL). An LEL cow was that having an embryo with no heartbeat, detached membranes, or floating structures, including embryo remnants detected at pregnancy check by ultrasonography (US) 28-42 days post-AI, whereas an NLEL cow was that with embryo heartbeats detectable by US at pregnancy check 28-42 d post-IA. We took two blood samples from every cow at pregnancy check by US (the day of LEL detection) and 28 d later to perform serological diagnosis of BVDV, BoHV-1, and N. caninum; and to measure blood progesterone concentration at pregnancy check (28-42 d post-AI). We also sampled the conceptus from all the LEL cows. We performed PCR to detect BVDV, BoHV-1, and N. caninum in sampled conceptuses from LEL cows. Finally, we evaluated the associations of risk factors (serological titers, seroconversion, and progesterone) with LEL odds with logistic models. The risk for LEL was associated with serological titers to BVDV (P = 0.03) and tended to be associated with seroconversion to BVDV, given that 19.6% (9/46) in LEL and 6.5% (3/46) in NLEL cows seroconverted to BVDV (P = 0.09). In addition, BVDV was detected in conceptuses from LEL cows that seroconverted to BVDV but not in LEL cows that did not seroconvert. Conversely, the risk for LEL was not associated with the titers or seroconversion to BoHV-1 and N. caninum. BoHV-1 and N. caninum were not identified in any of the conceptuses. Finally, blood progesterone concentration was similar in LEL and NLEL cows, and it was not associated with the risk for LEL (P = 0.54). In conclusion, BVDV infection is a risk factor for LEL in dairy cows.

Subject-Dependent Artifact Removal for Enhancing Motor Imagery Classifier Performance under Poor Skills.

The Electroencephalography (EEG)-based motor imagery (MI) paradigm is one of the most studied technologies for Brain-Computer Interface (BCI) development. Still, the low Signal-to-Noise Ratio (SNR) poses a challenge when constructing EEG-based BCI systems. Moreover, the non-stationary and nonlinear signal issues, the low-spatial data resolution, and the inter- and intra-subject variability hamper the extraction of discriminant features. Indeed, subjects with poor motor skills have difficulties in practicing MI tasks against low SNR scenarios. Here, we propose a subject-dependent preprocessing approach that includes the well-known Surface Laplacian Filtering and Independent Component Analysis algorithms to remove signal artifacts based on the MI performance. In addition, power- and phase-based functional connectivity measures are studied to extract relevant and interpretable patterns and identify subjects of inefficency. As a result, our proposal, Subject-dependent Artifact Removal (SD-AR), improves the MI classification performance in subjects with poor motor skills. Consequently, electrooculography and volume-conduction EEG artifacts are mitigated within a functional connectivity feature-extraction strategy, which favors the classification performance of a straightforward linear classifier.

Deep and Wide Transfer Learning with Kernel Matching for Pooling Data from Electroencephalography and Psychological Questionnaires.

Motor imagery (MI) promotes motor learning and encourages brain-computer interface systems that entail electroencephalogram (EEG) decoding. However, a long period of training is required to master brain rhythms' self-regulation, resulting in users with MI inefficiency. We introduce a parameter-based approach of cross-subject transfer-learning to improve the performances of poor-performing individuals in MI-based BCI systems, pooling data from labeled EEG measurements and psychological questionnaires via kernel-embedding. To this end, a Deep and Wide neural network for MI classification is implemented to pre-train the network from the source domain. Then, the parameter layers are transferred to initialize the target network within a fine-tuning procedure to recompute the Multilayer Perceptron-based accuracy. To perform data-fusion combining categorical features with the real-valued features, we implement stepwise kernel-matching via Gaussian-embedding. Finally, the paired source-target sets are selected for evaluation purposes according to the inefficiency-based clustering by subjects to consider their influence on BCI motor skills, exploring two choosing strategies of the best-performing subjects (source space): single-subject and multiple-subjects. Validation results achieved for discriminant MI tasks demonstrate that the introduced Deep and Wide neural network presents competitive performance of accuracy even after the inclusion of questionnaire data.

Lameness in Early Lactation Is Associated with Lower Productive and Reproductive Performance in a Herd of Supplemented Grazing Dairy Cows.

The main aim of this study was to assess the associations between the timing of lameness clinical case occurrence in lactation with productive and reproductive performances in grazing Holstein cows. A cohort study was carried out on a dataset with records from a commercial dairy herd (Buenos Aires, Argentina) for cows that calved and were dried off from January 2010 through June 2017. The first recorded event of lameness per lactation was considered for the study. Criteria for lactation inclusion included not having uterine diseases, mastitis, or anovulatory cysts during the studied risk period (i.e., up to 200 DIM). Therefore, a total of 7156 out of 20,086 lactations were included in the statistical analysis. The association between lameness case occurrence in lactation (cows not lame (LG0) vs. lame cows between parturition and first service (LG1) vs. lame cows between first service and first pregnancy (LG2)) with productive (i.e., accumulated milk yield to 150 DIM (MILK150) and 300 DIM (MILK305)) and reproductive performances (hazard of insemination and pregnancy) was analyzed with linear regression models and proportional hazard regression models, respectively. Lame cows produced 161 and 183 kg less MILK150 and MILK305 than non-lame herd mates, respectively. Moreover, LG1 cows produced 216 kg less MILK150 and 200 kg less MILK305 than LG0 cows, and LG2 cows also produced 58 kg less MILK150 and 158 kg less MILK305 than LG0 cows. The LG1 cows had a lower hazard of service than LG0 cows (HR = 0.43, 95%CI = 0.39-0.47). Furthermore, LG1 cows had a lower hazard of pregnancy than LG0 cows (HR = 0.52, 95%CI = 0.46-0.59) and took longer to get pregnant than LG0 cows (median [95%CI], 139 [132-144] vs. 101 [99-103]). Moreover, LG2 cows had a much lower hazard of pregnancy than LG0 cows (HR = 0.08, 95%CI = 0.05-0.12) and much longer calving to first pregnancy interval than LG0 cows (188 [183-196] vs. 101 [99-103]). In conclusion, cows that become lame in early lactation produce less milk and have lower hazards of insemination and pregnancy than herd mates that are healthy or become lame later in lactation. In addition, cows that become lame immediately after the voluntarily waiting period have the poorest reproductive performance (i.e., they have the lowest hazard of pregnancy and the longest calving to pregnancy interval).

An Enhanced Joint Hilbert Embedding-Based Metric to Support Mocap Data Classification with Preserved Interpretability.

Motion capture (Mocap) data are widely used as time series to study human movement. Indeed, animation movies, video games, and biomechanical systems for rehabilitation are significant applications related to Mocap data. However, classifying multi-channel time series from Mocap requires coding the intrinsic dependencies (even nonlinear relationships) between human body joints. Furthermore, the same human action may have variations because the individual alters their movement and therefore the inter/intraclass variability. Here, we introduce an enhanced Hilbert embedding-based approach from a cross-covariance operator, termed EHECCO, to map the input Mocap time series to a tensor space built from both 3D skeletal joints and a principal component analysis-based projection. Obtained results demonstrate how EHECCO represents and discriminates joint probability distributions as kernel-based evaluation of input time series within a tensor reproducing kernel Hilbert space (RKHS). Our approach achieves competitive classification results for style/subject and action recognition tasks on well-known publicly available databases. Moreover, EHECCO favors the interpretation of relevant anthropometric variables correlated with players' expertise and acted movement on a Tennis-Mocap database (also publicly available with this work). Thereby, our EHECCO-based framework provides a unified representation (through the tensor RKHS) of the Mocap time series to compute linear correlations between a coded metric from joint distributions and player properties, i.e., age, body measurements, and sport movement (action class).

Single-Trial Kernel-Based Functional Connectivity for Enhanced Feature Extraction in Motor-Related Tasks.

Motor learning is associated with functional brain plasticity, involving specific functional connectivity changes in the neural networks. However, the degree of learning new motor skills varies among individuals, which is mainly due to the between-subject variability in brain structure and function captured by electroencephalographic (EEG) recordings. Here, we propose a kernel-based functional connectivity measure to deal with inter/intra-subject variability in motor-related tasks. To this end, from spatio-temporal-frequency patterns, we extract the functional connectivity between EEG channels through their Gaussian kernel cross-spectral distribution. Further, we optimize the spectral combination weights within a sparse-based ℓ2-norm feature selection framework matching the motor-related labels that perform the dimensionality reduction of the extracted connectivity features. From the validation results in three databases with motor imagery and motor execution tasks, we conclude that the single-trial Gaussian functional connectivity measure provides very competitive classifier performance values, being less affected by feature extraction parameters, like the sliding time window, and avoiding the use of prior linear spatial filtering. We also provide interpretability for the clustered functional connectivity patterns and hypothesize that the proposed kernel-based metric is promising for evaluating motor skills.

Deep Neural Regression Prediction of Motor Imagery Skills Using EEG Functional Connectivity Indicators.

Motor imaging (MI) induces recovery and neuroplasticity in neurophysical regulation. However, a non-negligible portion of users presents insufficient coordination skills of sensorimotor cortex control. Assessments of the relationship between wakefulness and tasks states are conducted to foster neurophysiological and mechanistic interpretation in MI-related applications. Thus, to understand the organization of information processing, measures of functional connectivity are used. Also, models of neural network regression prediction are becoming popular, These intend to reduce the need for extracting features manually. However, predicting MI practicing's neurophysiological inefficiency raises several problems, like enhancing network regression performance because of the overfitting risk. Here, to increase the prediction performance, we develop a deep network regression model that includes three procedures: leave-one-out cross-validation combined with Monte Carlo dropout layers, subject clustering of MI inefficiency, and transfer learning between neighboring runs. Validation is performed using functional connectivity predictors extracted from two electroencephalographic databases acquired in conditions close to real MI applications (150 users), resulting in a high prediction of pretraining desynchronization and initial training synchronization with adequate physiological interpretability.

Dynamic Modeling of Common Brain Neural Activity in Motor Imagery Tasks.

Evaluation of brain dynamics elicited by motor imagery (MI) tasks can contribute to clinical and learning applications. The multi-subject analysis is to make inferences on the group/population level about the properties of MI brain activity. However, intrinsic neurophysiological variability of neural dynamics poses a challenge for devising efficient MI systems. Here, we develop a time-frequency model for estimating the spatial relevance of common neural activity across subjects employing an introduced statistical thresholding rule. In deriving multi-subject spatial maps, we present a comparative analysis of three feature extraction methods: Common Spatial Patterns, Functional Connectivity, and Event-Related De/Synchronization. In terms of interpretability, we evaluate the effectiveness in gathering MI data from collective populations by introducing two assumptions: (i) Non-linear assessment of the similarity between multi-subject data originating the subject-level dynamics; (ii) Assessment of time-varying brain network responses according to the ranking of individual accuracy performed in distinguishing distinct motor imagery tasks (left-hand vs. right-hand). The obtained validation results indicate that the estimated collective dynamics differently reflect the flow of sensorimotor cortex activation, providing new insights into the evolution of MI responses.

Entropy-Based Estimation of Event-Related De/Synchronization in Motor Imagery Using Vector-Quantized Patterns.

Assessment of brain dynamics elicited by motor imagery (MI) tasks contributes to clinical and learning applications. In this regard, Event-Related Desynchronization/Synchronization (ERD/S) is computed from Electroencephalographic signals, which show considerable variations in complexity. We present an Entropy-based method, termed VQEnt, for estimation of ERD/S using quantized stochastic patterns as a symbolic space, aiming to improve their discriminability and physiological interpretability. The proposed method builds the probabilistic priors by assessing the Gaussian similarity between the input measured data and their reduced vector-quantized representation. The validating results of a bi-class imagine task database (left and right hand) prove that VQEnt holds symbols that encode several neighboring samples, providing similar or even better accuracy than the other baseline sample-based algorithms of Entropy estimation. Besides, the performed ERD/S time-series are close enough to the trajectories extracted by the variational percentage of EEG signal power and fulfill the physiological MI paradigm. In BCI literate individuals, the VQEnt estimator presents the most accurate outcomes at a lower amount of electrodes placed in the sensorimotor cortex so that reduced channel set directly involved with the MI paradigm is enough to discriminate between tasks, providing an accuracy similar to the performed by the whole electrode set.

Regression Networks for Neurophysiological Indicator Evaluation in Practicing Motor Imagery Tasks.

Motor Imagery (MI) promotes motor learning in activities, like developing professional motor skills, sports gestures, and patient rehabilitation. However, up to 30% of users may not develop enough coordination skills after training sessions because of inter and intra-subject variability. Here, we develop a data-driven estimator, termed Deep Regression Network (DRN), which jointly extracts and performs the regression analysis in order to assess the efficiency of the individual brain networks in practicing MI tasks. The proposed double-stage estimator initially learns a pool of deep patterns, extracted from the input data, in order to feed a neural regression model, allowing for infering the distinctiveness between subject assemblies having similar variability. The results, which were obtained on real-world MI data, prove that the DRN estimator fosters pre-training neural desynchronization and initial training synchronization to predict the bi-class accuracy response, thus providing a better understanding of the Brain-Computer Interface inefficiency of subjects.

Analysis of puerperal metritis treatment records in a grazing dairy farm in Argentina.

OBJECTIVE: To assess the efficacy of antibiotic usage for the treatment of puerperal metritis (PM) and its association with reproductive performance, a retrospective cohort study including a total of 9168 records of cows from a dairy farm in Argentina was run. MATERIAL AND METHODS: Cows having a PM3 (metricheck, scale 0-3) and treated with ceftiofur (ceftiofur crystalline free acid, 6.6 mg/kg) at 0-21 days postpartum (p. p.) (n = 2688), and cows having a PM 1-2 and not treated with an antibiotic at 0-21 days p. p. (n = 6480) were included in the study. All cows were reexamined with metricheck to assess the clinical cure (vaginal discharge [VD] score 0), partial cure (VD score similar or lower than previous), no cure (VD score higher than previous). Cows with a metricheck VD1-3 after 0-21 days p. p. were diagnosed as clinical endometritis (CE) 1-3. The occurrence of PM1-3, cure rate, calving to conception interval, the hazard of pregnancy, odds for non-pregnancy, and odds for CE were analyzed using SAS software. RESULTS: A total of 8876 PM1-3 records were included, 2435 records of PM3 treatments with ceftiofur (27.43 %), and 6441 records of PM1-2 (72.57 %) with no treatment. Cows having PM1 and PM2 became pregnant 14 and 12 days earlier than cows with PM3 (p < 0.001). The PM3 ceftiofur treated cows had a clinical cure of 24.85 % (PM0); 53.63 % had a partially cure; and 18.52 % no cure. Conversely, cows with PM1-2 had a 51.96 %, 20.70 %, and 24.53 % cure rate, respectively (p < 0.001). Cows having complete cure became pregnant 13 and 11 days earlier than cows having partial cure and no cure (p < 0.001). Cows that had PM3 during the first 21 days p. p. had twice the chances of developing CE compared to cows having PM1-2 (41.28 % vs. 24.14 %, p < 0.001). After 21 days p. p., less than 1 % of cows with clinical cure developed CE compared to 63.32 % that developed CE with partial cure, and 38.21 % with no cure (p < 0.001). CONCLUSION AND CLINICAL RELEVANCE: After ceftiofur treatment, 78 % of cows were cured when measured by disappearance of fetid VD but only 25 % of cows had clinical cure when measured by appearance of a clear VD. The cows that remained with clinical metritis had more chances of having CE after 21 days p. p. and had more days open than cows with clear normal VD.

Late embryonic losses in supplemented grazing lactating dairy cows: Risk factors and reproductive performance.

The main objective of this study was to evaluate the risk factors for late embryonic loss (LEL) in supplemented grazing dairy cows. Additional objectives were to assess the incidence of LEL and its association with the reproductive performance of cows. A data set containing productive, reproductive, and health records of 13,551 lactations was used. A retrospective case-control study involving 631 cows with LEL (cases) and 2,524 controls (4 controls per case within each study year) was run. A case of LEL was defined when the embryo had no heartbeat or there was evidence of detached membranes or floating structures including embryo remnants by ultrasonography (US) at 28 to 42 d post-artificial insemination (AI), whereas a non-case was defined as a cow diagnosed with positive pregnancy by US 28 to 42 d post-AI and reconfirmed as pregnant 90 ± 7 d post-AI. Four controls per case were randomly selected from the non-cases with a temporal matching criterion (±3 d around the date of the fecundating AI of the case). Multivariable logistic models were offered with the following predictors: year of LEL (2011 through 2015), season of LEL (summer vs. fall vs. winter vs. spring), parity (1 vs. 2 vs. ≥3), uterine disease (UD), non-uterine disease (NUD), body condition score at parturition, body condition score at 28 to 42 d post-AI (BCS-LEL), days in milk (DIM), and daily milk yield (MY). Statistical significance was set at P < 0.05 and a tendency was set at P ≤ 0.10. We found that 4.7, 22, and 23% of cows had LEL, UD, and NUD, respectively. Cases tended to have higher daily MY than controls (32.5 vs. 31.8 kg); also, cases had much longer calving to pregnancy interval (226 vs. 118 d), lower hazard of pregnancy [hazard ratio = 0.39, 95% confidence interval (CI) = 0.35-0.43], and higher odds for non-pregnancy [odds ratio (OR) = 2.89, 95% CI = 2.37-3.54] than controls. We found that the odds for LEL increased with parity number (OR = 2.48, 95% CI = 1.99-3.08 for parity ≥3) and with BCS-LEL <2.50 (OR = 1.81, 95% CI = 1.33-2.47). Conversely, the odds for LEL decreased with BCS-LEL >3.00 (OR = 0.70, 95% CI = 0.53-0.91). The odds for LEL increased with UD (OR = 1.23, 95% CI = 1.01-1.49), NUD (OR = 1.24, 95% CI = 1.01-1.54), DIM (OR = 1.03, 95% CI = 1.00-1.05), and daily MY (OR = 1.14, 95% CI = 1.04-1.25) in univariable models only. Finally, the odds for LEL were not associated with year, season, DIM, and body condition score at parturition. In conclusion, LEL is associated with extended calving to pregnancy interval, and among its risk factors are parity number and BCS-LEL.

Adaptive Bayesian label fusion using kernel-based similarity metrics in hippocampus segmentation.

The effectiveness of brain magnetic resonance imaging (MRI) as a useful evaluation tool strongly depends on the performed segmentation of associated tissues or anatomical structures. We introduce an enhanced brain segmentation approach of Bayesian label fusion that includes the construction of adaptive target-specific probabilistic priors using atlases ranked by kernel-based similarity metrics to deal with the anatomical variability of collected MRI data. In particular, the developed segmentation approach appraises patch-based voxel representation to enhance the voxel embedding in spaces with increased tissue discrimination, as well as the construction of a neighborhood-dependent model that addresses the label assignment of each region with a different patch complexity. To measure the similarity between the target and training atlases, we propose a tensor-based kernel metric that also includes the training labeling set. We evaluate the proposed approach, adaptive Bayesian label fusion using kernel-based similarity metrics, in the specific case of hippocampus segmentation of five benchmark MRI collections, including ADNI dataset, resulting in an increased performance (assessed through the Dice index) as compared to other recent works.

A Finite-Difference Solution for the EEG Forward Problem in Inhomogeneous Anisotropic Media.

Accurate source localization of electroencephalographic (EEG) signals requires detailed information about the geometry and physical properties of head tissues. Indeed, these strongly influence the propagation of neural activity from the brain to the sensors. Finite difference methods (FDMs) are head modelling approaches relying on volumetric data information, which can be directly obtained using magnetic resonance (MR) imaging. The specific goal of this study is to develop a computationally efficient FDM solution that can flexibly integrate voxel-wise conductivity and anisotropy information. Given the high computational complexity of FDMs, we pay particular attention to attain a very low numerical error, as evaluated using exact analytical solutions for spherical volume conductor models. We then demonstrate the computational efficiency of our FDM numerical solver, by comparing it with alternative solutions. Finally, we apply the developed head modelling tool to high-resolution MR images from a real experimental subject, to demonstrate the potential added value of incorporating detailed voxel-wise conductivity and anisotropy information. Our results clearly show that the developed FDM can contribute to a more precise head modelling, and therefore to a more reliable use of EEG as a brain imaging tool.

Kernel-Based Relevance Analysis with Enhanced Interpretability for Detection of Brain Activity Patterns.

We introduce Enhanced Kernel-based Relevance Analysis (EKRA) that aims to support the automatic identification of brain activity patterns using electroencephalographic recordings. EKRA is a data-driven strategy that incorporates two kernel functions to take advantage of the available joint information, associating neural responses to a given stimulus condition. Regarding this, a Centered Kernel Alignment functional is adjusted to learning the linear projection that best discriminates the input feature set, optimizing the required free parameters automatically. Our approach is carried out in two scenarios: (i) feature selection by computing a relevance vector from extracted neural features to facilitating the physiological interpretation of a given brain activity task, and (ii) enhanced feature selection to perform an additional transformation of relevant features aiming to improve the overall identification accuracy. Accordingly, we provide an alternative feature relevance analysis strategy that allows improving the system performance while favoring the data interpretability. For the validation purpose, EKRA is tested in two well-known tasks of brain activity: motor imagery discrimination and epileptic seizure detection. The obtained results show that the EKRA approach estimates a relevant representation space extracted from the provided supervised information, emphasizing the salient input features. As a result, our proposal outperforms the state-of-the-art methods regarding brain activity discrimination accuracy with the benefit of enhanced physiological interpretation about the task at hand.