ABSTRACT The hands constitute the primary means by which humans interface with the external world. Increasingly, these interactions involve objects such as tools that require significant manual dexterity to use properly. Impairment of fine motor control of the hand, such as commonly occurs after stroke, can profoundly impact object manipulation, thereby substantially reducing functionality of the hand, and of the entire upper limb in general. Clinical tools to support rehabilitation of fine manual dexterity are limited despite the high frequency of hand impairment in stroke survivors and the importance of the hand to overall function and daily living. Thus, this study will focus on the evaluation of a novel, multimodal treatment approach in which functional electrical stimulation (FES) is combined with use of a soft hand exoskeleton to promote active training of independent digit movement after stroke. These modalities will be used in a coordinated manner, with FES helping to excite appropriate motor units to guide neuroplasticity while the exoskeleton inhibits unwanted flexion of other digits to assist task performance. The FES will target nerves rather muscles to enable excitation of a greater range of muscles and to reduce fatigue. The soft exoskeleton will allow full active range of motion of all digits with minimal impedance when unactuated. Together they will be used in conjunction with the Actuated Virtual Keyboard (AVK) virtual reality platform. With this system, flexion of the user?s digits results in ?playing? displayed keys. The virtual nature of the keys facilitates repetitive practice with the proper level of challenge to maximize learning. For example, the amount of joint flexion required to play a key can be easily adjusted during a training session, as well as the level of FES and exoskeleton assistance provided to the user. As the user progresses, the exoskeleton can also be used to resist intended movement to provide further challenge. In these ways, task difficulty is matched to user ability. The system provide visual and audio feedback of success or failure of key strikes, as well as recording of finger movements producing errors, to inform the user and therapist. Stroke survivors in the subacute phase of recovery will complete 18 training sessions with this novel therapy paradigm over 6 weeks. A matched group of stroke survivors will participate in 18 training sessions of similar duration and intensity, but employing standard therapy techniques focused on the hand. Evaluation sessions will be conducted prior to initiation of training, midway through the training, after the completion of all training sessions, and one month later. The ability of participants to created independent movement and isometric force with each digit will be quantified; a high-density electrode array will be used to measure activation patterns of targeted hand muscles. It is anticipated that the group receiving the multimodal training with the AVK system will exhibit significantly greater improvement on clinical assessments of motor control of the hand, and in digit individuation and independent activation of muscles and muscle compartments. Encouraging outcomes for the novel therapeutic intervention introduced in this study would have direct ramifications for clinical practice.