Grant: Project Grant

Cortical stimulation could be a powerful way to treat mental illnesses and restore brain functions lost to stroke or brain injury. In theory, techniques like electrical, magnetic, or light stimulation could someday allow clinicians to directly change neural activity into some desired, therapeutic pattern. However, there is at least one substantial obstacle to the tremendous clinical promise of cortical stimulation: the technique is notoriously unreliable. Stimulation can produce unexpected effects, different effects at different times, or it can fail to produce any effect at all. We do not understand why this occurs. If we did, we could develop methods to increase the precision and reliability of stimulation and clear a path towards realizing the promise of cortical stimulation in the clinic. This proposal will first test a novel idea about how the state of the brain at the time of stimulation may alter how stimulation affects neurons and behavior. Next, it will determine whether we can improve stimulation by using knowledge about the state of the brain before stimulation. We will do this by measuring neural activity while we stimulate a part of the prefrontal cortex in which we can also measure the behavioral effects of stimulation. To make sense out of the data, we will use mathematical techniques, including tools from machine learning and a computational model of brain activity. One outcome will be tools for ensuring precise stimulation outcomes in this part of the prefrontal cortex. These tools could have applications in treating hemineglect (a common consequence of stroke and brain injury). If our central hypothesis is correct, this work also has the potential to identify a general set of tools that could be used to improve our ability to control populations of neurons across the cortex, with broad applications in the next generation of stimulation-based treatments for psychological and neurological conditions.