Single neuron computations of cognition in the human brain
Understanding how information is encoded, processed, and decoded to produce behavior is a fundamental goal of neuroscience. In this dissertation, we aim to expand our understanding of our human decision-making processes at the single-neuronal level. We describe three studies exploring the neural substrate of decision-making in three separate brain regions. First, we describe a method for recording the activity of individual neurons in human subjects. The unique combination of behavioral and neurophysiological data will allow us to better understand the neural substrate of cognitive functions in humans. Second, we explored how decisions are represented in the brain. We recorded single neuronal responses in the human nucleus accumbens while subjects engaged in a financial decision-making task. We found that neurons in the nucleus accumbens predicted upcoming decisions well before the behavior was manifested. In addition, these neurons encoded a positive and negative prediction error signal, signaling the difference between expected and realized outcome. Third, we explored how the brain represents decision conflict and how it adapts to prime future decisions allowing tradeoff between speed and accuracy. We found that individual neurons in the human dorsal anterior cingulate cortex encode the level of decision conflict in a dose-dependent manner. In addition, these neurons encode historical conflict information, priming the neural circuit to future trials of the same or varying conflict levels. Following selective ablation of the dorsal anterior cingulate cortex, we found this signal was selectively abolished. Lastly, we explored how the brain represents decisions under conflict and if these decisions are malleable to external intervention. We found that neurons in the human subthalamic nucleus are selectively activated and encode the upcoming decision during situations of high decision conflict. Based on the physiological findings, we then applied intermittent stimulation through the implanted deep brain stimulation electrode during the same task, to demonstrate a causal interaction between the physiology and behavior. In conclusion, we describe a set of experiments that systematically explore human decision-making processes at the single-neuronal level.