FY2017 has seen significant progress towards realizing our goals and objectives. We have continued our efforts capturing and analyzing intracranial recordings while participants engage in cognitive tasks designed to probe memory encoding and retrieval. Patients with medically refractory epilepsy receiving intracranial electrodes and surgical treatment at the Clinical Center have been recruited for these studies. We have also continued our work analyzing local field potential and single unit spiking activity captured from the basal ganglia during deep brain stimulation surgery for patients with Parkinsons disease. Our previous efforts examining changes in the oscillatory power evident in these field potential signals have provided us with much insight into the changes in local population activity in the context of memory encoding and retrieval. Over the past year, however, we have largely focused our efforts on extending our analysis techniques in order to understand changes in the human brain across spatial scales, and in exploring the use of electrical stimulation to determine whether we can ascribe causality to these neural signals. At the larger scale, we have been exploring communication between brain regions. We have developed a metric of effective connectivity that is premised on the hypothesis that communication between brain regions occurs with consistent and stable timing. We have demonstrated that such effective connections indeed exist in the brain, and reflect stable pathways of communication through which information is propagated. At the smaller scale, we have extended our ability to capture neural activity to the level of individual neuronal units from micro-electrode arrays implanted on the cortical surface in patients treated for medically refractory epilepsy. We have successfully collected single unit spiking activity from these microelectrode arrays in seven patients while they have participated in our cognitive tasks. Finally, we have explored the use of single pulse electrical stimulation to identify stereotypical responses to stimulation in the human brain. We have developed an approach that allows us to predict the responses to stimulation with good fidelity, and are now exploring how these responses can be combined to generate patterns of neural activity. With these additional methodological advances, we are now currently focused on three main sets of studies. In the first set of studies, we have been interested in investigating whether neuronal activity is reinstated from memory encoding to memory retrieval, and in examining the precise spatiotemporal dynamics of such reinstatement. Using a paired associates episodic memory task, we have directly examined these questions. We have previously shown that, during successful recall, there is significantly greater reinstatement of patterns of oscillatory power distributed across multiple cortical locations and across frequencies. In this past year, we have now extended this analysis to examine the time course of activity during encoding and retrieval. We have found that when items are successfully retrieved from memory, the time course of neural activation is compressed compared to the activity that is present during encoding. This suggests that the act of retrieval involves a temporal compression of neural activity, and is consistent with previous evidence found in medial temporal lobe structures in rodents. We have described this work in a recent publication. Second, we have also examined the behavior of individual neurons during the paired associates task. We have found that while some neurons exhibit significant increases in firing during successful encoding while other exhibit significant decreases, when we examine the population activity of spiking neurons, we find that the same patterns of spiking activity that are present during encoding are also present during retrieval. This provides us with direct evidence that neural spiking activity in the human cortex is itself reinstated during successful retrieval. We have described this work in a recent publication. In a second set of studies, we have been interested in understanding how the fidelity of memory encoding is modulated by the state of the brain, and in particular the extent to which attention may mediate the formation of successful memories. We have collected data from intracranial electrodes and from our microelectrode arrays while participants engage in a behavioral task that specifically asks this question. Briefly, we present items to be encoded on a computer screen, and these items may or may not be preceded or followed by a visual cue indicating whether this item should be encoded in memory. We have found that attention improves verbal memory encoding by causing a suppression of neural activity and spiking responses in the anterior temporal lobe before items to be remembered are even presented. Moreover, resection of the anterior temporal lobe caused a significant impairment in the ability of attention to improve memory encoding. Together, these findings implicate this region, the anterior temporal lobe, with a specific role in attention-enhanced memorization, and suggest that the canonical model of attention in which fronto-parietal networks exert control over regions of perception needs to be extended to include the anterior temporal lobe when attention is used in the service of verbal memory. We have completed this project, and have submitted a manuscript describing these efforts for publication. Moving forward, we are currently integrating this accumulated knowledge with our methodological advances related to electrical stimulation. Ultimately, our goal is to use electrical stimulation to generate patterns of neural activity in the brain and to ask whether such patterns of activity play a causal role in facilitating memory encoding. We are currently developing experimental tasks and collecting data that explore how the use of electrical stimulation can manipulate the state of the brain, and subsequently lead to better memory formation. Finally, in a third set of studies, we have focused on understanding the interaction between the human memory and decision systems. It is becoming increasingly clear that the basal ganglia, with its widespread cortical connections, play a role in human decision processes. We have previously explored the role of the basal ganglia, and specifically the subthalamic nucleus, in simple sensorimotor decisions. One question, however, is whether similar circuit dynamics are present when making non-motor decisions related to memory. In our studies of attention and memory, participants actively choose to encode subsequent items based on a prospective cue. We have focused our recent efforts on exploring this process of deciding to encode, and have developed a task to explicitly test the hypothesis that the decision to attend to or encode a target stimulus co-opts the same neural mechanisms used to mediate motor decisions. We have found that oscillatory and spiking activity in the human subthalamic nucleus is modulated during non-motor decisions to encode items into memory or to ignore them. In addition, we have found that oscillatory communication between the subthalamic nucleus and the prefrontal cortex participates in this process. We have completed these efforts and have recently submitted a manuscript describing this work for publication. We are now currently interested in understanding how this interaction between decision and memory operates in the other direction, namely how does memory affect our ability to make decisions. We have developed a novel task in which participants must rely upon their memory for associations that they have formed to make decisions, and are interested in how this process is represented in the human brain.
Showing the most recent 10 out of 43 publications
