Monte Carlo Dropout for Uncertainty Estimation and Motor Imagery Classification

Abstract

Motor Imagery (MI)-based Brain–Computer Interfaces (BCIs) have been widely used as an alternative communication channel to patients with severe motor disabilities, achieving high classification accuracy through machine learning techniques. Recently, deep learning techniques have spotlighted the state-of-the-art of MI-based BCIs. These techniques still lack strategies to quantify predictive uncertainty and may produce overconfident predictions. In this work, methods to enhance the performance of existing MI-based BCIs are proposed in order to obtain a more reliable system for real application scenarios. First, the Monte Carlo dropout (MCD) method is proposed on MI deep neural models to improve classification and provide uncertainty estimation. This approach was implemented using Shallow Convolutional Neural Network (SCNN-MCD) and with an ensemble model (E-SCNN-MCD). As another contribution, to discriminate MI task predictions of high uncertainty, a threshold approach is introduced and tested for both SCNN-MCD and E-SCNN-MCD approaches. The BCI Competition IV Databases 2a and 2b were used to evaluate the proposed methods for both subject-specific and non-subject-specific strategies, obtaining encouraging results for MI recognition.

Active Disturbance Rejection Control for Robot Manipulator

“…The ADRC is based on an extension of the system model with an additional and fictitious state variable, representing those elements of the system dynamics that the user does not include in the mathematical description of the plant. These virtual states (sum of internal and external disturbances, sometimes denoted as a total disturbance) are estimated online and used in the control loop to decouple the system from the actual perturbation acting on the plant (10) [34], [35], [36], [37], [38], [39] as…”

Measuring brain potentials of imagination linked to physiological needs and motivational states

“…Clear evidences have been provided of reliable electrophysiological markers for motor imagery (e.g., Milanés-Hermosilla et al, 2021 ; Mattioli et al, 2022 ), and, to a lesser extent, perceptual and cognitive imagery ( Cai et al, 2013 ; Leoni et al, 2021 , 2022 ; Proverbio et al, 2022 ), for communication with brain computer interface (BCI) systems ( Ash and Benson, 2018 ). In contrast, not much is known about the electrophysiological markers of motivational imagery (craves, wills, needs and desires), despite the fact that this particular type of mental content is valuable for interacting with patients with disorders of consciousness, such as coma or locked-in syndrome.…”

Robust Motor Imagery Tasks Classification Approach Using Bayesian Neural Network

“…In medical applications, uncertainty analysis has been widely used for disease diagnoses, such as COVID-19 [ 22 ], tuberculosis [ 23 ], ataxia [ 24 ], cancer [ 25 ], diabetic retinopathy [ 26 , 27 ], and epileptogenic brain malformations [ 28 ]. To the best of our knowledge, strategies for uncertainty analysis have been little explored in MI-based BCIs, excluding a previous study [ 29 ] that applied Monte Carlo dropout techniques with this purpose, considered as an approximate Bayesian inference in deep Gaussian processes [ 21 ]. The current paper follows a different method by applying BNN.…”