Feasibility of Engineered Neural Networks for Neuro-Restoration After Cortical Stroke

Over 8 million people in the US are living with chronic stroke. The resulting disabilities lead to significant public health costs and decreased quality of life. With no means to satisfactorily restore functions after stroke, new effective regenerative approaches are much needed.

Therefore, a new concept of “engineered neural networks” (ENN) is proposed. The ENN is envisioned to contain neurons derived from human adult stem cells and be structured to have precise inputs and feedback loops. These connections will enable the ENN’s integration with other brain and body areas and thereby learn behaviors.

Ultimately, this integration process can lead to ENNs replacing stroke-damaged brain and any associated functions. Before this goal can be achieved, the first step is to develop a benchtop testbed cultured neural network (CNN) and verify that it can be trained to perform an arbitrary sensory task and motor behavior. Successful completion of this study would demonstrate that CNNs can be trained to encode arbitrary behaviors and interact with other brain and body areas.

This would justify further research to pursue the development of ENNs and methods to implant such systems into the stroke-damaged brain. Ultimately, this may restore brain resources in a functionally meaningful manner, which may in turn lead to ground-breaking regenerative treatments for stroke rehabilitation. This study will also: (i) promote education and lifelong learning in biomedical engineering students and research trainees; (ii) broaden the participation of underrepresented groups in neuroengineering; (iii) increase scientific literacy.

This project will develop a novel neurorestorative concept of “engineered neural networks” (ENN) envisioned to contain neurons derived from human induced pluripotent stem cells (HIPSCs). The ENN will have a precise structure consisting of inputs and feedback loops, formed by excitatory and inhibitory neurons, to facilitate communication with other brain/body areas post-stroke.

By interfacing the ENN with microelectronics, consistent and precise patterns of excitatory and inhibitory inputs and error-driven feedback will be delivered to train the ENN to interface with other brain/body areas and learn arbitrary human behaviors. Ultimately, this can establish bidirectional communication of the ENN with the remaining post-stroke brain, thereby replacing the stroke-damaged cortex and subserved functions.

To gauge the feasibility of this concept, the first step is to develop a benchtop testbed cultured neural network (CNN) and verify that it can be trained to perform an arbitrary sensory task and motor behavior. Scientifically, the CNN platform can provide a means to study basic concepts of neural encoding and memory retrieval. From a neuroengineering perspective, this study may inform the optimal design of stem cell-derived neuronal tissue.

If successful, this study would demonstrate that CNNs can be trained to encode arbitrary behaviors and interact with other brain and body areas and would justify further research to pursue the development of ENNs and methods to implant such systems into the stroke-damaged brain. Ultimately, this technology may restore brain resources in a functionally meaningful manner, which may, in turn, lead to ground-breaking regenerative treatments for stroke rehabilitation.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.