Intracortical Brain-Computer Interfaces (iBCIs) aim to restore communication, mobility, and independence to Veterans and others with paralyzing disorders such as amyotrophic lateral sclerosis (ALS), stroke, or spinal cord injury. In the late stages of ALS, the progressive loss of mobility is accompanied by a loss of speech, resulting in tetraplegia and anarthria, or locked-in syndrome. Though assistive and augmentative communication (AAC) devices partially address this problem, such devices become less useful and eventually fail as motor power continues to decline. In contrast, iBCIs can record the neural activity associated with intended movement directly from cortex. In this renewal Merit Review application, we propose to expand upon the tremendous progress made in the development of the investigational BrainGate Neural Interface system toward providing Veterans with intuitive, always-available, wireless point-and-click control over a computer, tablet, or any other software- based communication system. In the proposed research, we will recruit two Veterans or other people with ALS at the Providence VA Medical Center to participate in the ongoing BrainGate pilot clinical trial. After placement of two 4x4x1.5 mm, 96-electrdode Blackrock (Utah) recording arrays in the dominant motor cortex, participants will engage in two or three recording sessions per week, in their place of residence. The research, which will also leverage other participants in the multi-site BrainGate trial, will focus over a year or more with each participant on the development of improved, robust neural decoders. As a first aim, we will extend the stability of neural control by developing a new class of relational decoders with improved flexibility, adaptability, and noise tolerance. This will be facilitated by the use of a discriminative rather than generative decoding approach that focuses on modeling the probability distribution of the (low- dimensional) volitional state outputs based on (high-dimensional) neural signals. This strategy does not rely upon an underlying assumption of cosine tuning to endpoint velocity, and allows for flexible, non- linear mapping across different intended movements and effectors with increased tolerance to noise. In the second aim, we will develop new strategies to rapidly calibrate and continuously update neural decoders. Our new methodology will allow us to transition directly to closed loop control and to calibrate functional neural decoders within ~1 minute of activating the system. We will also implement new strategies to maintain continuously both intended direction and click decoding by updating the decoder after every successful target selection, a useful step toward the design of embedded neuroprosthetic systems and practical, independent use of an iBCI. In both of these aims, decoders will be compared to the current state of the art approaches for BCI control. Finally, we will develop a cl...