Brain dynamics that drive variability within and between patients are an important, but poorly understood, element of many cognitive disorders. The long-term goal of this research project is to develop technology that will identify brain activity patterns associated with successful performance on a given task, and use this pattern as a target for brain-computer interface (BCI) training. The overarching hypothesis is that using BCI training to more often have a brain state that is spontaneously correlated to good performance will, in turn, improve overall performance. This approach could be developed into a powerful tool for rehabilitation and therapy for many neurological and psychiatric disorders. Here we will investigate persistent developmental stuttering (PDS) as a model to study brain dynamics associated with successful vs. unsuccessful performance. PDS is a speech disorder where fluent speech is punctuated to various degrees by stuttering. Individuals with PDS are otherwise neurologically in the normal range, which avoids complicating factors in most patient populations. Stuttering is intermittent; thus on some occasions the brain is in a state conducive to fluent speech and at other times it is not. We propose to use EEG activity shortly before speaking to predict whether somebody with PDS will stutter or speak fluently. Preliminary data are given to show proof of concept with traditional EEG analysis methods. This approach will be expanded by first using advanced methods such as common spatial pattern analysis and machine learning over multiple subject sessions to identify EEG signals that distinguish fluent vs. dysfluent trials (Aim 1). PDS subjects will then be trained to produce and maintain their EEG pattern that is most strongly associated with fluent speech by using BCI methods. We hypothesize that individuals will learn to modulate EEG features to be more consistent with fluent trials, which in turn will significantly reduce stuttering rate. After successful completion of this project we envision a new BCI-based intervention that can be used to encourage neural states conducive to fluent speech in those who stutter. The BCI intervention would complement traditional speech therapy using behavioral methods. The ?two-step approach? of first identifying brain states associated with a patient?s best performance followed by BCI training to enter that state more often can be applied to rehabilitation in many other neurological and psychiatric disorders, such as Alzheimer?s disease, traumatic brain injury, and mood disorders, to name a few.
The goal of this project is to develop brain-computer interface technology to optimize brain function on an individual basis. This could have therapeutic applications to many neurological and psychiatric disorders, including stroke, Alzheimer?s disease, and traumatic brain injury.
