NEUROPROSTHETICS IN VETERINARY NEUROLOGY: FUNCTIONAL RECOVERY AND LIMB CONTROL IN PARAPLEGIC DOGS USING IMPLANTED BRAIN-COMPUTER INTERFACES

Abstract

Neuroprosthetics, powered by brain-computer interface (BCI) technology, hold great promise in restoring autonomy for individuals with severe motor impairments. This study presents a mixed-method experimental framework combining EEG-based signal decoding with user-centered assessments to evaluate the efficacy and adaptability of non-invasive BCI systems. EEG data collected from motor imagery tasks were processed using Common Spatial Pattern (CSP) and Power Spectral Density (PSD) techniques, followed by classification using Support Vector Machine (SVM) models. The system achieved high performance, with classification accuracies exceeding 90% in multiple participants and substantial gains in precision, recall, and F1-score. Additionally, a passive BCI layer leveraging Long Short-Term Memory (LSTM) networks was implemented to monitor user affective states, providing an adaptive layer that adjusted interaction parameters based on emotional fluctuations. A qualitative component, involving semi-structured interviews and System Usability Scale (SUS) evaluations, revealed high levels of user satisfaction, indicating the interface’s effectiveness in real-world neuroprosthetic scenarios. A total of 9 comprehensive tables and 12 figures were generated, capturing classification outcomes, affective monitoring, and usability metrics. Results underscore the critical role of machine learning in decoding neural intent and enhancing human-computer interaction. The integration of emotional awareness and volitional control within a unified BCI pipeline offers new avenues for user-adaptive neuroprosthetic solutions. This study advances the field by demonstrating a functional, scalable, and emotionally intelligent BCI system capable of restoring motor function while responding to the cognitive and emotional context of the user.