The Unlearning of Stimulus-Outcome Associations through Intracortical Microstimulation

University Of California Berkeley, Berkeley CA

Investigators

Linked publications & trials

Abstract

? DESCRIPTION (provided by applicant): Addiction is a disease of aberrant learning. Addictive substances, and the environmental cues that predict those substances, hijack normal reward circuitry in the brain and become overvalued relative to all else, leading to a catastrophic cycle of drug craving, seeking, and intake behaviors at the expense of healthy behaviors. These aberrant stimulus-outcome (SO) associations are manifested in frontolimbic (FL) networks that normally drive a wide range of behaviors vital for healthy function. In the addicted brain, these circuits become maladaptive, leading to compulsive behaviors, incapable of the flexibility the system originally possessed. Currently, two major avenues of treatment for addiction exist: behavioral and neuropsychiatric treatments. Behavioral treatments aim to manage maladaptive SO associations by breaking habitual drug taking behavior, though relapse rates are high. Neuropsychiatric treatments broadly target neurotransmitter systems as a whole without regard to specific SO associations, thereby impacting normal reward processing and learning. In the current grant, we aim to test an alternate approach to the treatment of addiction by the precise targeting and modification of SO associations at the systems level through a brain-machine interface (BMI) framework. We propose that by identifying the underlying neural representations of SO associations in FL networks and modifying them through electrical microstimulation, we can provide a means to unlearn addictive behaviors without unwanted side effects. Modifying SO associations for the treatment of addiction requires understanding how SO associations are formed in FL networks and how these formations can be manipulated through targeted electrical stimulation. To do so, large populations of neurons must be recorded and stimulated during SO learning. Recent advances in recording technology enable such experiments. For example, recent studies have manipulated sensory networks through electrical stimulation to affect learning, thus optimizing the control of BMI prostheses. We plan to adopt the same methods to manipulate value signals to alter SO associative learning in FL networks. Our experiments will focus on two major building blocks. First, we will determine optimal stimulation parameters and FL targets to modify SO associations. We will test this by determining the effects of a wide range of stimulation parameters and locations across the FL circuit on value-based learning and decision making (open loop). Second, we will use a BMI framework to decode value information about a given stimulus from large populations of FL neurons in real time as a control signal to preferentially apply targeted stimulation. This will enable closed-loop bidirectional control over SO associations. We will study how stimulation affects SO associations represented in the FL circuit across days in both open and closed-loop conditions.