TRANSFORM DBS (Transdiagnostic Repair of Affective Networks by Systematic, Function-Oriented, Real-time Modeling and Deep Brain Stimulation) is a five-year program to develop next-generation treatments for severe psychiatric illness. Mental disorders, particularly post traumatic stress disorder (PTSD), depression, anxiety, and addiction, are the leading cause of disability and lost productivity world-wide. They are epidemic among civilians, members of the U.S. military, and veterans. Moreover, current treatments, whether medication or talk therapy, fail to help nearly half of patients even when delivered by experts. MGH has long been a leader in the use of neurosurgical procedures and implanted medical devices to help patients who have not gotten better with outpatient treatments. Unfortunately, even state-of-the-art approaches such as deep brain stimulation (DBS) are not fully effective. Even when we are able to place a device directly into the brain and change electrical activity at a very precise target, a substantial fraction of patients still do not get better. TRANSFORM will create a new generation of medical devices that actively sense a patient's brain activity, deliver electrical stimulation, then adjust that stimulation in real time based on the brain's electrical response. These "closed loop" systems offer tremendous potential for bringing new hope to patients living with severe mental illness

Our core approaches, as given in the TRANSFORM name, are:
Transdiagnostic: We believe that part of the problem is the current diagnostic system: that "depression" or "PTSD" can arise from multiple different causes in the brain, only some of which are treatable with brain stimulation at our known targets. We are developing algorithms and ideas based not on these categorical diagnoses, but by running patients through objective, standardized behavioral testing combined with high-resolution brain scanning.

Repairing Affective Networks: Our goal is to treat disorders of mood, anxiety, and self-regulation. We will do that by studying and improving the function of the brain's own networks. At MGH, we have long been leaders in human neuroscience, done by recording patients' brains as they undergo clinical neurosurgical procedures for epilepsy, movement disorders, or psychiatric illness. Our patients volunteer to let us record signals from the deep brain structures involved in feeling, motivation, and decision-making. By studying those emotional networks, we can understand how the human brain functions and how it sometimes goes wrong.

Systematic: We are creating "big data" – large and publicly available datasets of human and animal neuroscience, both invasive and non-invasive. We will share those data and our analyses with the world under the BRAIN Initiative, to ensure that every possible insight is gained.

Function-Oriented: We are building technologies that will allow patients to return to their homes and communities and to enjoy their daily lives. With our partner, Draper Laboratory, we are designing new devices that are flexible and reconfigurable to run innovative new algorithms, while simultaneously being small enough to be totally implantable and low-power enough to run for a full day without needing to be recharged or otherwise adjusted. Draper's long expertise in rugged, low-power systems is critical to our success.

Real-Time: Our devices will understand and respond to a patient's immediate clinical need, as expressed in their ongoing brain activity. We are constructing new algorithms to perform "state estimation" from neural activity – to measure, directly from the brain, a patient's current level of symptoms and whether their device is being effective at relieving those symptoms.

Model-Based: MGH, MIT, and Boston University jointly contain world experts in questions of analyzing, understanding, and simulating the brain's function. We are bringing together experts across all three institutions to build new mathematical representations of how the activity of different brain areas forms a processing network that gives rise to symptoms or to wellness. Those models will directly inform the control algorithms that will change stimulation to meet a patient's needs.