Metodologia para distinção de frequências de estimulação sonora ótimas para aplicações em ICMs baseadas no paradigma oddball auditivo

Abstract

Individuals with Amyotrophic Lateral Sclerosis face a continuous loss of functions, autonomy, and decreased possibilities for interaction and communication with the world, with the most severe cases evolving to complete paralysis. In this context, the Brain-Machine Interfaces are seen as candidates for alternative communication allowing communication and control without using the brain's natural output channels. Some common challenges include low signal-to-noise ratio, mainly in single tests. In this sense, among the elements that can influence both the morphology of the cortical signals and the system's performance, there are the stimulation parameters. When designing auditory interfaces, the following parameters stand out: inter-stimulus interval, probability of occurrence of the target stimulus, the intensity of the stimuli, and the frequency of tones. This work aims to evaluate the influence of the target sound frequency on the P300 potential to find the optimal frequency capable of causing significant evoked potentials for the control of Brain-Machine Interfaces. Data were collected from five volunteers and five different frequencies for the target stimulus presented to the right and left ears. Analyzes were performed on the amplitude and latency of the P300 potential, seeking to distinguish between the target and standard stimuli through a Support Vector Machine. The results indicated that different individuals have different cortical responses to the same stimulation pattern. The results have a direct impact on the design of Brain-Machine Interfaces. We argue that researchers should first evaluate the user's response to sound stimuli before any experiment to find out the best combination for the specific subject.